ProgrammableMarketing/heytenor.com (opens in a new tab)
PDF

Views

  • Plan
  • Roadmap
  • Research

On this page

  • The brief
  • Start here
  • Idea portfolio
  • Effort vs impact
  • This week · 6
  • This quarter · 12
  • Big bets · 3
  • Wildcards · 2
  • How we can help
Generated Nov 10, 2026
  1. Programmable
  2. Playbooks
  3. Tenor
Tenor logo

growth playbook · Nov 10, 2026

heytenor.com (opens in a new tab)

Tenor growth playbook

Diagnosis

Tenor is a three-person Y Combinator Summer 2026 company with a sharp thesis: AI should own jobs end to end and answer for outcomes, with a budget, a read-only start and human sign-off. The thesis is clearly written, but in the wrong places.

The fullest description lives in llms.txt, the only booking button sits on a Mission page the homepage doesn't link to, the homepage's "Contact" link opens a support page for existing Slack customers, and the one blog post in the sitemap is empty.

This playbook first gives buyers a way to book and a reason to: a homepage demo button, the llms.txt pitch on the homepage, the essay the sitemap promises, share previews and a qualifying booking form. Next it packages the read-only Audit as a paid first step and builds proof without logos: a sample Trial replay, a logistics page, a trust page, design partners and a weekly worker note. The big bets turn the outcome-and-budget thesis into the product and its pricing.

Growth Score

5/10

Biggest Opportunity

Tenor's read-only Audit and historical Trial replay are rare, low-risk ways for an operations leader to say yes. Put them on the homepage, price the Audit, and use replays of the buyer's own history as proof where rivals use logos.

Biggest Risk

Well-funded rivals selling "AI workers" to the same logistics buyers (HappyRobot with a Series C and DHL; Pallet with Lineage Logistics) lead with big proof and a demo button, while Tenor's site names no customers and doesn't ask for a call.

01

Start here

The three moves to make first. Each one is written up in full below.

  1. 1

    Put "Book a demo" on the homepage: today the only booking button is hidden on the Mission page

    Copy the Mission page's "Book a demo →" link into the homepage hero.

    High impact First result in 1 week
  2. 2

    Say on the homepage what llms.txt already says: AI workers that own recurring operational work, starting read-only

    Paste the four stages from llms.txt into a homepage draft.

    High impact First result in 2 weeks
  3. 3

    Publish "The Pilot Never Got Replaced": the sitemap already lists it, but the page is empty

    Take the empty URL out of the sitemap today, or publish a first draft in its place.

    Medium impact First result in 2 weeks

02

Your idea portfolio

23 ideas, grouped by when to run them. Open any idea for how to run it: the steps, who to approach, copy you can use, and how to measure it.

  • 7 ideas are marked Unexpected: plays you are unlikely to have tried.

Every idea placed by the effort it takes against the impact it could have. Quick wins sit top left; select a number to open that idea. The filters below narrow the list by status, timing, category and what Programmable can build.

Effort vs impactEach number is an idea. Select one to open it.
Impact →
Quick wins · 3Big bets · 11Fill-ins · 6Reconsider · 3
Effort →

Showing 23 of 23 ideas

This week 6

Quick wins with a clear first action.

  • 01

    Tenor's homepage has four links: Careers, Contact, Privacy and Terms. "Contact" opens a support page written for existing Slack customers (installs, permissions, delivery problems, deletion requests) that offers only an email address. The site's one "Book a demo" button, to a cal.com scheduler, sits at the bottom of /mission, and the homepage doesn't link to /mission. A buyer arriving from the YC launch, a referral or a shared link has no visible way to book a call, while HappyRobot and Pallet, selling AI workers to the same logistics buyers, put "Book a demo" first.

    High impactLow effort First result in 1 week CEO

    First action

    Copy the Mission page's "Book a demo →" link into the homepage hero.

    How to measure it

    Homepage-sourced demo bookings

    Now
    –
    Target
    3 a week
    By
    60 days

    Steps

    1. 1

      Add the button

      Hero and nav, linking to the existing cal.com team scheduler.

      CTO (co-founder) 1 hour

    2. 2

      Relabel Contact

      "Contact" becomes "Support"; add a Mission link.

      CTO (co-founder) 30 minutes

    3. 3

      Cover the side doors

      The same button at the end of /support/, /mission/ and /blog/.

      CTO (co-founder) 1 hour

    4. 4

      Tag the links

      A source parameter per page, so each booking shows where it came from.

      CTO (co-founder) 1 hour

    5. 5

      Count bookings

      Bookings by source page, every Monday.

      CEO (co-founder) weekly

    6. 6

      Test the label

      "Book a demo" against "Start a read-only audit".

      CEO (co-founder) 4 weeks

    First result
    1 week
    Cost
    $0
    Owner
    CEO (co founder)

    Smallest test

    Add the hero button today and count bookings for two weeks against the previous two.

    Why it works

    Use the scheduler that already exists. Make it the one button in the homepage hero and the nav ("Book a demo", or test "Start a read-only audit"). Rename "Contact" to "Support" so prospects don't land in install troubleshooting, and add a Mission link. Put the same button at the end of /support/ and /blog/ for prospects who arrive there. Tag each button with a source parameter so the founders can see which page books calls.

    At three people, every booked call is pipeline the founders would otherwise never see. The trust signals are already on the page (YC, a team from Stanford, the IMO, Tesla and Waymo) and the scheduler already exists; the fix connects them. It costs an hour and doesn't change the minimal design.

    Ready to use

    ChecklistHomepage booking fix

    [ ] "Book a demo" button in the hero, linking to the cal.com team scheduler
    [ ] Same button in the nav
    [ ] "Contact" renamed to "Support"
    [ ] Mission link added to the nav or footer
    [ ] Button at the end of /support/, /mission/ and /blog/
    [ ] Source parameter on each button
    [ ] Weekly count of bookings by source

    Risks and how to handle them

    • More unqualified calls

      Add one required qualifying question to the booking form (idea 5).

    • A button feels salesy on a minimal page

      Keep one quiet text button in the hero; no pop-ups or banners.

    Evidence

    • Your websiteHomepage (checked November 10, 2026): H1 "The scale of a company can now be engineered." and one line of product copy, "Our AI infrastructure makes expertise scale like software, allowing what makes a company exceptional to make it enormous."; "We work with people who know their industries too well to leave them as they are."; team backgrounds from Stanford, UC Berkeley, Technical University of Munich, International Mathematical Olympiad, Amazon, Tesla and Waymo. The only links are Careers, Contact (which opens /support/), Privacy and Terms: no "Book a demo" button, calendar link, product, customer or pricing section.Source
    • Your websiteMission page: "The infrastructure for companies built around AI." with four principles ("From capable models to capable companies.", "Built around real expertise.", "Learn from outcomes.", "Build for the long term."). It holds the site's only "Book a demo →" button, linking to a cal.com team scheduler; the homepage does not link to this page.Source
    • Your websiteSupport page (the homepage "Contact" link): "Tenor support." covers "Slack installation, permissions, message delivery, security reports, and data requests"; a Slack workspace admin must approve the Tenor app and its permissions; the only contact is an email address; no scheduling link.Source
    • Competitor researchRivals' own homepages: HappyRobot sells AI agents it calls "AI workers" that "automate communication-heavy work at enterprise scale" in logistics and seven other industries, citing a Series C, "15M+" interactions a month and DHL as a customer; Pallet sells AI systems to "the companies that run the physical world", citing Lineage Logistics: "Automated 2 million tasks at >99% accuracy"; both lead with "Book a demo". Lindy sells an "AI employee" with plans that "start at $29.99 per user per month" and "$50 in free credits".Source
  • 02

    Tenor's clearest pitch is in llms.txt, a file written for agents and crawlers: a worker "owns a responsibility end to end, not a task"; it starts with a read-only Audit and a Trial that "replays historical work" before anything changes; there are "No migrations"; consequential actions need human sign-off; and the first buyers are freight, logistics and supply chain operators. The homepage says only that "AI infrastructure makes expertise scale like software", and the YC one-liner says something else again. Crawlers get a better pitch than people do.

    High impactLow effort First result in 2 weeks CEO

    First action

    Paste the four stages from llms.txt into a homepage draft.

    How to measure it

    Homepage visitor to demo booking

    Now
    –
    Target
    2%
    By
    60 days

    Steps

    1. 1

      Pull the copy

      The worker definition, four stages and buyer line from llms.txt.

      CEO (co-founder) 2 hours

    2. 2

      Write the worker line

      One sentence: what a worker owns, where it runs, what it never does without sign-off.

      CPO (co-founder) 1 hour

    3. 3

      Lay out the stages

      Audit, Design, Trial, Deploy as a four-step strip.

      CPO (co-founder) 1 day

    4. 4

      Name the buyer

      Freight, logistics and supply chain operators first, as llms.txt says.

      CEO (co-founder) 1 hour

    5. 5

      Ship

      Keep the current H1 and design; add the sections below it.

      CTO (co-founder) 1 day

    6. 6

      Keep in sync

      One source for the homepage, llms.txt and the YC profile.

      CPO (co-founder) monthly

    First result
    2 weeks
    Cost
    $0
    Owner
    CEO (co founder)

    Smallest test

    Add only the four-stage strip and compare bookings for two weeks.

    Why it works

    Add three short homepage sections under the H1, using copy that already exists in llms.txt: what a Tenor worker is (one sentence: it owns one recurring responsibility, inside your tools, reporting in Slack); the four stages, with "read-only first", "replay before anything changes" and "no migrations" spelled out; and who it is for (teams with high volumes of recurring operational work, freight and logistics first). End with the demo button from idea 1. Then keep llms.txt, the homepage and the YC profile saying the same thing.

    Operations leaders buy reduced risk. "Read-only pilot, replay before change, human sign-off" answers their first three objections, and none of it is on the page they see.

    Risks and how to handle them

    • Naming logistics puts off other buyers

      Say "freight and logistics first" and keep the general line above it.

    Builds on

    Evidence

    • Your websiteHomepage (checked November 10, 2026): H1 "The scale of a company can now be engineered." and one line of product copy, "Our AI infrastructure makes expertise scale like software, allowing what makes a company exceptional to make it enormous."; "We work with people who know their industries too well to leave them as they are."; team backgrounds from Stanford, UC Berkeley, Technical University of Munich, International Mathematical Olympiad, Amazon, Tesla and Waymo. The only links are Careers, Contact (which opens /support/), Privacy and Terms: no "Book a demo" button, calendar link, product, customer or pricing section.Source
    • Your websitellms.txt, the site's context file for AI agents and crawlers, has the fullest product description: "the operating layer for a company's AI workforce", "persistent AI that learns how your business operates and owns recurring work inside the tools you already use"; "A worker owns a responsibility end to end, not a task"; four stages: Audit ("A structured, read-only pass over how work moves across teams, systems, and decision points"), Design, Trial (the worker "replays historical work so its decisions can be reviewed before anything changes") and Deploy ("No migrations"); "Consequential actions require human sign-off until authority is explicitly expanded" and actions are logged; aimed at companies "that run high volumes of recurring operational work", including "large freight, logistics, and supply chain operators". It names no customers, deployment counts or metrics, and says "Backed by Y Combinator."Source
    • Y CombinatorYC company profile: Summer 2026 batch, founded 2026, team of 3 (CEO, CPO and CTO co-founders), San Francisco, tagged B2B, Automation and AI. One-liner: "Infrastructure for building and scaling AI-native businesses." Launch post, "Tenor: AI that owns jobs end-to-end and is accountable for business outcomes": companies can measure what AI tokens cost but not what they produce, so Tenor ties each piece of AI work to a defined outcome and budget, tracks spend against results, and uses the results to decide where future AI resources go.Source
  • 03

    The sitemap sends search engines and AI crawlers to /blog/the-pilot-never-got-replaced/, but the page shows only "Read the latest notes from Tenor." and /blog/ says "New writing is on the way." The title is the best pitch Tenor owns: it names the buyer's real frustration, AI pilots that stall, for a company whose path is Audit, Trial, Deploy.

    Medium impactLow effort First result in 2 weeks CEO

    First action

    Take the empty URL out of the sitemap today, or publish a first draft in its place.

    How to measure it

    Bookings attributed to the essay

    Now
    –
    Target
    5
    By
    60 days

    Steps

    1. 1

      Fix the sitemap

      Remove the empty URL until the post is live.

      CTO (co-founder) 15 minutes

    2. 2

      Outline

      Why pilots stall: no owner, no outcome, no budget.

      CEO (co-founder) 1 day

    3. 3

      Draft

      1,000-1,500 words with one worked example.

      CEO (co-founder) 3 days

    4. 4

      Add proof

      An anonymised Trial replay or a synthetic one (idea 9).

      CPO (co-founder) 1 day

    5. 5

      Publish with the button

      Date, author role and a share image.

      CTO (co-founder) 1 hour

    6. 6

      Distribute

      Founder LinkedIn and X, the YC community, and outbound emails.

      CEO (co-founder) 1 week

    First result
    2 weeks
    Cost
    Founder time
    Owner
    CEO (co founder)

    Smallest test

    Publish a 600-word version this week.

    Why it works

    Write the essay: why AI pilots stall (no owner, no outcome, no budget, no path from read-only to trusted) and how a worker that owns one responsibility, replays history before changing anything and is measured on spend against results gets past the pilot. Until it is ready, take the URL out of the sitemap. End the post with the demo button and share it from the founders' accounts and in the YC community.

    Founder essays are the main channel for a three-person company, and this one gives every outbound email a link worth opening.

    Risks and how to handle them

    • Generic thought leadership

      Use concrete, anonymised examples from Audits.

    Evidence

    • Your websiteSitemap (checked November 10, 2026): 9 URLs, none with a lastmod date: home, mission, blog, one post, privacy, terms, support, support/zoom and llms.txt. The listed post, /blog/the-pilot-never-got-replaced/, shows only "Notes | Tenor" and "Read the latest notes from Tenor." with no article, and /blog/ says "New writing is on the way." There are no product, use-case, security, pricing or customer pages.Source
    • Your websitellms.txt, the site's context file for AI agents and crawlers, has the fullest product description: "the operating layer for a company's AI workforce", "persistent AI that learns how your business operates and owns recurring work inside the tools you already use"; "A worker owns a responsibility end to end, not a task"; four stages: Audit ("A structured, read-only pass over how work moves across teams, systems, and decision points"), Design, Trial (the worker "replays historical work so its decisions can be reviewed before anything changes") and Deploy ("No migrations"); "Consequential actions require human sign-off until authority is explicitly expanded" and actions are logged; aimed at companies "that run high volumes of recurring operational work", including "large freight, logistics, and supply chain operators". It names no customers, deployment counts or metrics, and says "Backed by Y Combinator."Source
    • Y CombinatorYC company profile: Summer 2026 batch, founded 2026, team of 3 (CEO, CPO and CTO co-founders), San Francisco, tagged B2B, Automation and AI. One-liner: "Infrastructure for building and scaling AI-native businesses." Launch post, "Tenor: AI that owns jobs end-to-end and is accountable for business outcomes": companies can measure what AI tokens cost but not what they produce, so Tenor ties each piece of AI work to a defined outcome and budget, tracks spend against results, and uses the results to decide where future AI resources go.Source
  • 04

    The homepage has no social share image. Tenor sells a Slack app to teams that live in Slack, and its founders share links in Slack, LinkedIn and YC threads, where a link with no image and a one-line tagline looks like a placeholder.

    Medium impactLow effort First result in 1 week CTO

    First action

    Design the share card from the hero image.

    How to measure it

    Clicks per shared link (tracked)

    Now
    –
    Target
    +20%
    By
    30 days

    Steps

    1. 1

      Design the card

      1200x630, H1 plus one line on what a worker does.

      CPO (co-founder) 2 hours

    2. 2

      Add the tags

      og:image, og:title, og:description, twitter:card on all nine pages.

      CTO (co-founder) 1 hour

    3. 3

      Write descriptions

      One per page, under 155 characters.

      CEO (co-founder) 1 hour

    4. 4

      Test unfurls

      Slack, LinkedIn Post Inspector and X.

      CTO (co-founder) 30 minutes

    5. 5

      Card per post

      Each blog post gets its own image.

      CTO (co-founder) 1 hour per post

    6. 6

      Measure

      Clicks on tracked shared links.

      CEO (co-founder) monthly

    First result
    1 week
    Cost
    $0
    Owner
    CTO (co founder)

    Smallest test

    Add the homepage og:image and post the link in one Slack channel.

    Why it works

    Add og:image (a 1200x630 card with the H1 and a one-line description of the worker), og:title, og:description and twitter:card to all nine pages, plus matching meta descriptions. Check how the links unfurl in Slack, LinkedIn and X before the next launch post.

    Link previews are the first impression most visitors get of a seed-stage B2B company, and Slack, where buyers will meet the product, is where they see them most.

    Risks and how to handle them

    • Old previews stay cached

      Refresh with each platform's inspector tool.

    Evidence

    • PageSpeedGoogle PageSpeed (November 10, 2026): mobile score 84, desktop 100. In the lab phone test the main content (the hero image) appears at 4.6s against a 2.5s target; the homepage downloads 2.4 MB, about 1.3 MB of it video from Cloudflare Stream, with 602 KB of possible image savings. Real Chrome visitors pass Core Web Vitals (75th percentile LCP 1.09s, INP 61ms, CLS 0). The homepage has no social share image.Source
    • Your websiteSupport page (the homepage "Contact" link): "Tenor support." covers "Slack installation, permissions, message delivery, security reports, and data requests"; a Slack workspace admin must approve the Tenor app and its permissions; the only contact is an email address; no scheduling link.Source
  • 05

    The only booking path is a cal.com team scheduler. Tenor's Audit starts by mapping "how work moves across teams, systems, and decision points", and each worker is built around one responsibility. Asking for that responsibility before the call lets the founders arrive with a first sketch of the worker instead of spending the call on discovery.

    Medium impactLow effort First result in 2 weeks CEO

    First action

    Add the three questions to the cal.com event.

    How to measure it

    Call-to-Audit rate

    Now
    –
    Target
    40%
    By
    90 days

    Steps

    1. 1

      Add the questions

      Job, tools and frequency on the booking form.

      CEO (co-founder) 30 minutes

    2. 2

      Require one

      Only the recurring job is mandatory.

      CEO (co-founder) 10 minutes

    3. 3

      Log the answers

      One sheet of jobs, tools and volume.

      CPO (co-founder) weekly

    4. 4

      Send a pre-read

      A short note on how a worker would own that job.

      CEO (co-founder) per call

    5. 5

      Spot patterns

      Which jobs and tools come up again and again.

      CPO (co-founder) monthly

    6. 6

      Feed the roadmap

      Integrations and worker templates (idea 21).

      CTO (co-founder) monthly

    First result
    2 weeks
    Cost
    $0
    Owner
    CEO (co founder)

    Smallest test

    Add one optional question for two weeks and read the answers.

    Why it works

    Add to the booking form: "Which recurring job would you hand to a worker first?" (required), "Which tools does it run through?" and "Roughly how often does it happen each week?". Before each call, send a one-paragraph note on how a worker could own that job. Log every answer in a sheet: after a few months it is a map of which jobs and tools buyers bring.

    It shortens discovery, qualifies callers, and turns booking data into the roadmap for integrations and templates.

    Risks and how to handle them

    • Fewer bookings

      Keep one required question and two optional ones.

    Builds on

    Evidence

    • Your websiteMission page: "The infrastructure for companies built around AI." with four principles ("From capable models to capable companies.", "Built around real expertise.", "Learn from outcomes.", "Build for the long term."). It holds the site's only "Book a demo →" button, linking to a cal.com team scheduler; the homepage does not link to this page.Source
    • Your websitellms.txt, the site's context file for AI agents and crawlers, has the fullest product description: "the operating layer for a company's AI workforce", "persistent AI that learns how your business operates and owns recurring work inside the tools you already use"; "A worker owns a responsibility end to end, not a task"; four stages: Audit ("A structured, read-only pass over how work moves across teams, systems, and decision points"), Design, Trial (the worker "replays historical work so its decisions can be reviewed before anything changes") and Deploy ("No migrations"); "Consequential actions require human sign-off until authority is explicitly expanded" and actions are logged; aimed at companies "that run high volumes of recurring operational work", including "large freight, logistics, and supply chain operators". It names no customers, deployment counts or metrics, and says "Backed by Y Combinator."Source
  • 06

    This is not urgent: real Chrome visitors pass Core Web Vitals, with a 75th-percentile LCP of 1.09s. But Google's lab phone test shows the hero image at 4.6s against a 2.5s target, and the 2.4 MB homepage includes about 1.3 MB of Cloudflare Stream video and 602 KB of avoidable image weight. That matters for an operations buyer opening a shared link on mobile data, in a warehouse or at a trade show.

    Low impactLow effort First result in 1 day CTO

    First action

    Compress the hero image and add mobile-sized versions.

    How to measure it

    Mobile LCP (PageSpeed lab)

    Now
    –
    Target
    Under 2.5s
    By
    2 weeks

    Steps

    1. 1

      Compress the hero

      AVIF/WebP with fetchpriority=high.

      CTO (co-founder) 1 hour

    2. 2

      Size for phones

      srcset with mobile widths.

      CTO (co-founder) 1 hour

    3. 3

      Defer the video

      Start it after load, or skip it on small screens.

      CTO (co-founder) 2 hours

    4. 4

      Set caching

      Long cache lifetimes for static files.

      CTO (co-founder) 30 minutes

    5. 5

      Re-test

      PageSpeed mobile and desktop.

      CTO (co-founder) 15 minutes

    6. 6

      Watch field data

      Keep Core Web Vitals passing as pages are added.

      CTO (co-founder) monthly

    First result
    1 day
    Cost
    $0
    Owner
    CTO (co founder)

    Smallest test

    Re-run PageSpeed after compressing only the hero image.

    Why it works

    Serve the hero image as a compressed AVIF/WebP, sized for mobile, with fetchpriority=high; start the video only after the page has painted (or only on large screens), with the still image as its poster; set long cache lifetimes for static assets. Re-test in PageSpeed.

    It is cheap insurance that keeps the first impression fast and light on any connection.

    Risks and how to handle them

    • The hero loses its motion

      Use the first video frame as the poster image.

    Evidence

    • PageSpeedGoogle PageSpeed (November 10, 2026): mobile score 84, desktop 100. In the lab phone test the main content (the hero image) appears at 4.6s against a 2.5s target; the homepage downloads 2.4 MB, about 1.3 MB of it video from Cloudflare Stream, with 602 KB of possible image savings. Real Chrome visitors pass Core Web Vitals (75th percentile LCP 1.09s, INP 61ms, CLS 0). The homepage has no social share image.Source

This quarter 12

The core plays for the next three months.

  • 07

    The Audit is the first of Tenor's four stages and it is read-only, which makes it the easiest thing for an operations leader to approve. Today it is one line in llms.txt; there are no public prices, and the Terms put fees in "your order, subscription, or other agreement". Rivals ask for a demo call; none of their homepages offers a bounded first step.

    High impactMedium effort First result in 1 month CEO

    First action

    Write a one-page Audit scope from the last three sales calls.

    How to measure it

    Paid Audits

    Now
    –
    Target
    4
    By
    1 quarter

    Steps

    1. 1

      Define the scope

      One team, one responsibility, two weeks, read-only.

      CEO (co-founder) 2 days

    2. 2

      Set the price

      A fixed fee, credited against deployment.

      CEO (co-founder) 1 day

    3. 3

      Template the report

      Findings, evidence and the proposed worker.

      CPO (co-founder) 1 week

    4. 4

      Build the page

      /audit, linked from the homepage button.

      CTO (co-founder) 2 days

    5. 5

      Offer it on every call

      As the default next step.

      CEO (co-founder) ongoing

    6. 6

      Track conversion

      Audit to Trial to Deploy.

      CEO (co-founder) monthly

    First result
    1 month
    Cost
    Founder time
    Owner
    CEO (co founder)

    Smallest test

    Offer a paid Audit on the next five calls.

    Why it works

    Package the Audit: one team, one recurring responsibility, two weeks, read-only access, and a findings report with evidence plus an estimate of the hours and errors a worker would remove. Publish /audit with the deliverable, the access needed, the timeline and a price or price range, credited against the first worker if the customer deploys.

    A paid, bounded first step qualifies buyers, pays for a three-person team's time, and makes the Trial and Deploy a natural next step rather than a new sale.

    Risks and how to handle them

    • A price deters early design partners

      Waive it for design partners (idea 16) but show the list price.

    • Audits turn into consulting

      Cap them at one responsibility and two weeks.

    Evidence

    • Your websitellms.txt, the site's context file for AI agents and crawlers, has the fullest product description: "the operating layer for a company's AI workforce", "persistent AI that learns how your business operates and owns recurring work inside the tools you already use"; "A worker owns a responsibility end to end, not a task"; four stages: Audit ("A structured, read-only pass over how work moves across teams, systems, and decision points"), Design, Trial (the worker "replays historical work so its decisions can be reviewed before anything changes") and Deploy ("No migrations"); "Consequential actions require human sign-off until authority is explicitly expanded" and actions are logged; aimed at companies "that run high volumes of recurring operational work", including "large freight, logistics, and supply chain operators". It names no customers, deployment counts or metrics, and says "Backed by Y Combinator."Source
    • Your websiteTerms (last updated July 28, 2026): Tenor Technologies; they cover "Tenor's websites, connected services, and Slack app" and say Tenor delivers results through Slack and other connected services; fees "are described in your order, subscription, or other agreement with Tenor"; no public prices or free trial.Source
    • Competitor researchRivals' own homepages: HappyRobot sells AI agents it calls "AI workers" that "automate communication-heavy work at enterprise scale" in logistics and seven other industries, citing a Series C, "15M+" interactions a month and DHL as a customer; Pallet sells AI systems to "the companies that run the physical world", citing Lineage Logistics: "Automated 2 million tasks at >99% accuracy"; both lead with "Book a demo". Lindy sells an "AI employee" with plans that "start at $29.99 per user per month" and "$50 in free credits".Source
  • 08

    llms.txt names freight, logistics and supply chain operators as Tenor's buyers, but no page on the site says so. In that market HappyRobot ("AI workers", DHL, a Series C) and Pallet (Lineage Logistics, "2 million tasks") lead with big proof. Tenor's difference is in its own words: a worker owns a back-office responsibility end to end in the tools the team already uses, starts read-only, replays history before changing anything, and needs no migration.

    High impactMedium effort First result in 1 month CEO

    First action

    List five recurring logistics jobs from sales calls.

    How to measure it

    Logistics-page bookings

    Now
    –
    Target
    6
    By
    1 quarter

    Steps

    1. 1

      Pick the jobs

      Five recurring jobs heard in calls and Audits.

      CEO (co-founder) 2 days

    2. 2

      Write the page

      Plain language, one worked example.

      CEO (co-founder) 1 week

    3. 3

      Fit the stages

      Audit, Trial and Deploy for a freight team.

      CPO (co-founder) 2 days

    4. 4

      Add the security block

      Link to the trust page (idea 11).

      CTO (co-founder) 1 day

    5. 5

      Publish

      Linked from the homepage buyer line.

      CTO (co-founder) 1 day

    6. 6

      Use in outbound

      The page as the link in every logistics email.

      CEO (co-founder) ongoing

    First result
    1 month
    Cost
    Founder time
    Owner
    CEO (co founder)

    Smallest test

    Publish a one-screen version and link it in ten outbound emails.

    Why it works

    Build /logistics: the recurring back-office responsibilities Tenor takes on (confirmed from booking answers and Audits), how a worker runs next to the team in Slack, what Audit, Trial and Deploy look like for a freight team, the access and approval model, and the booking button. Describe the difference in approach (a worker that owns one responsibility versus agents that handle communication volume) without claims about rivals beyond what their own homepages say.

    Specialist buyers convert on pages written for their work, and "no migrations, read-only first" is most believable when it is set against their own systems.

    Risks and how to handle them

    • A thin page with no customers

      Add a design-partner quote as soon as one exists (idea 16).

    Builds on

    Evidence

    • Your websitellms.txt, the site's context file for AI agents and crawlers, has the fullest product description: "the operating layer for a company's AI workforce", "persistent AI that learns how your business operates and owns recurring work inside the tools you already use"; "A worker owns a responsibility end to end, not a task"; four stages: Audit ("A structured, read-only pass over how work moves across teams, systems, and decision points"), Design, Trial (the worker "replays historical work so its decisions can be reviewed before anything changes") and Deploy ("No migrations"); "Consequential actions require human sign-off until authority is explicitly expanded" and actions are logged; aimed at companies "that run high volumes of recurring operational work", including "large freight, logistics, and supply chain operators". It names no customers, deployment counts or metrics, and says "Backed by Y Combinator."Source
    • Competitor researchRivals' own homepages: HappyRobot sells AI agents it calls "AI workers" that "automate communication-heavy work at enterprise scale" in logistics and seven other industries, citing a Series C, "15M+" interactions a month and DHL as a customer; Pallet sells AI systems to "the companies that run the physical world", citing Lineage Logistics: "Automated 2 million tasks at >99% accuracy"; both lead with "Book a demo". Lindy sells an "AI employee" with plans that "start at $29.99 per user per month" and "$50 in free credits".Source
    • Your websiteSitemap (checked November 10, 2026): 9 URLs, none with a lastmod date: home, mission, blog, one post, privacy, terms, support, support/zoom and llms.txt. The listed post, /blog/the-pilot-never-got-replaced/, shows only "Notes | Tenor" and "Read the latest notes from Tenor." with no article, and /blog/ says "New writing is on the way." There are no product, use-case, security, pricing or customer pages.Source
  • 09

    In Tenor's Trial the worker "replays historical work so its decisions can be reviewed before anything changes". Tenor names no customers or metrics, while rivals publish DHL, Lineage Logistics and "2 million tasks at >99% accuracy". A replay report needs no logo: it shows the decisions the worker would have taken against what people actually did, and where it would have stopped to ask for sign-off.

    High impactMedium effort First result in 1 month CPO

    First action

    Mock up the report layout with synthetic data.

    How to measure it

    Replay requests

    Now
    –
    Target
    8
    By
    1 quarter

    Steps

    1. 1

      Choose the job

      One common recurring responsibility.

      CPO (co-founder) 1 day

    2. 2

      Build the data

      A realistic synthetic history with known answers.

      Member of Technical Staff, Evals & Task Engineering 1 week

    3. 3

      Run the replay

      The worker's decisions plus its sign-off requests.

      Founding Engineer, Agent and Data Systems 1 week

    4. 4

      Design the report

      Agreement rate, escalations, time saved, errors caught, spend.

      CPO (co-founder) 3 days

    5. 5

      Publish

      A web page, clearly labelled as sample data.

      CTO (co-founder) 1 day

    6. 6

      Offer it on calls

      "We can run this on your last 30 days."

      CEO (co-founder) ongoing

    First result
    1 month
    Cost
    About two weeks of engineering time
    Owner
    CPO (co founder)

    Smallest test

    Show the mock report on three calls.

    Why it works

    Build a sample replay report on synthetic or anonymised data for one recurring job: agreement with human decisions, cases it escalated for sign-off, time saved, errors it would have caught, and spend against budget. Publish it as a web page with the demo button, and offer every prospect "a replay on your own last 30 days" as the Trial.

    It is proof without logos, it shows the outcome-and-budget thesis from the YC launch post in practice, and a buyer's own history is more persuasive than any rival's customer list.

    Risks and how to handle them

    • Synthetic data looks staged

      Label it clearly and replace it with an approved, anonymised real replay.

    Evidence

    • Your websitellms.txt, the site's context file for AI agents and crawlers, has the fullest product description: "the operating layer for a company's AI workforce", "persistent AI that learns how your business operates and owns recurring work inside the tools you already use"; "A worker owns a responsibility end to end, not a task"; four stages: Audit ("A structured, read-only pass over how work moves across teams, systems, and decision points"), Design, Trial (the worker "replays historical work so its decisions can be reviewed before anything changes") and Deploy ("No migrations"); "Consequential actions require human sign-off until authority is explicitly expanded" and actions are logged; aimed at companies "that run high volumes of recurring operational work", including "large freight, logistics, and supply chain operators". It names no customers, deployment counts or metrics, and says "Backed by Y Combinator."Source
    • Y CombinatorYC company profile: Summer 2026 batch, founded 2026, team of 3 (CEO, CPO and CTO co-founders), San Francisco, tagged B2B, Automation and AI. One-liner: "Infrastructure for building and scaling AI-native businesses." Launch post, "Tenor: AI that owns jobs end-to-end and is accountable for business outcomes": companies can measure what AI tokens cost but not what they produce, so Tenor ties each piece of AI work to a defined outcome and budget, tracks spend against results, and uses the results to decide where future AI resources go.Source
    • Competitor researchRivals' own homepages: HappyRobot sells AI agents it calls "AI workers" that "automate communication-heavy work at enterprise scale" in logistics and seven other industries, citing a Series C, "15M+" interactions a month and DHL as a customer; Pallet sells AI systems to "the companies that run the physical world", citing Lineage Logistics: "Automated 2 million tasks at >99% accuracy"; both lead with "Book a demo". Lindy sells an "AI employee" with plans that "start at $29.99 per user per month" and "$50 in free credits".Source
  • 10

    The YC launch post says each piece of AI work starts with an outcome and a budget and Tenor tracks spend against results. The Terms put fees in private orders, and the site has no pricing page. Lindy publishes plans from $29.99 per user per month, which sets the buyer's frame unless Tenor offers its own.

    Medium impactLow effort First result in 1 month CEO

    First action

    Write down how the last three proposals were priced.

    How to measure it

    Pricing-page visitors who book

    Now
    –
    Target
    10%
    By
    90 days

    Steps

    1. 1

      Agree the model

      Audit fee, Trial terms, per-worker pricing with a budget.

      CEO (co-founder) 2 days

    2. 2

      Write the page

      How it works, not a rate card.

      CEO (co-founder) 2 days

    3. 3

      Add an example

      One worker, one budget, one outcome.

      CPO (co-founder) 1 day

    4. 4

      Publish

      Linked from the nav and /audit.

      CTO (co-founder) 1 day

    5. 5

      Use it on calls

      The same frame in every proposal.

      CEO (co-founder) ongoing

    6. 6

      Review

      Adjust after each quarter of deals.

      CEO (co-founder) quarterly

    First result
    1 month
    Cost
    $0
    Owner
    CEO (co founder)

    Smallest test

    Add a three-question pricing FAQ to the homepage.

    Why it works

    Publish /pricing as an explanation of the model rather than a price list: the Audit (a fixed fee, idea 7), the Trial, then each worker priced on the responsibility it owns, with a monthly budget cap and the outcome it is measured on, and a contact route for volume. Show one worked example: one worker, one budget, one outcome. Contrast "per responsibility" with "per seat".

    Buyers shortlist vendors that explain how they charge, and an outcome frame keeps Tenor out of a per-seat comparison with cheaper general agents.

    Risks and how to handle them

    • Fixing prices too early

      Publish the model and ranges, not fixed numbers.

    Builds on

    Evidence

    • Y CombinatorYC company profile: Summer 2026 batch, founded 2026, team of 3 (CEO, CPO and CTO co-founders), San Francisco, tagged B2B, Automation and AI. One-liner: "Infrastructure for building and scaling AI-native businesses." Launch post, "Tenor: AI that owns jobs end-to-end and is accountable for business outcomes": companies can measure what AI tokens cost but not what they produce, so Tenor ties each piece of AI work to a defined outcome and budget, tracks spend against results, and uses the results to decide where future AI resources go.Source
    • Your websiteTerms (last updated July 28, 2026): Tenor Technologies; they cover "Tenor's websites, connected services, and Slack app" and say Tenor delivers results through Slack and other connected services; fees "are described in your order, subscription, or other agreement with Tenor"; no public prices or free trial.Source
    • Competitor researchRivals' own homepages: HappyRobot sells AI agents it calls "AI workers" that "automate communication-heavy work at enterprise scale" in logistics and seven other industries, citing a Series C, "15M+" interactions a month and DHL as a customer; Pallet sells AI systems to "the companies that run the physical world", citing Lineage Logistics: "Automated 2 million tasks at >99% accuracy"; both lead with "Book a demo". Lindy sells an "AI employee" with plans that "start at $29.99 per user per month" and "$50 in free credits".Source
  • 11

    Tenor's strongest reassurances are scattered across pages written for other jobs. Zoom access is read-only and Tenor "does not create, modify, or delete Zoom users, meetings, or recordings"; the Slack app needs admin approval; pilots start read-only; consequential actions need human sign-off; actions are logged; deletion is on request. They sit in support docs and llms.txt, not where an IT reviewer looks, and the sitemap has no security page.

    High impactLow effort First result in 1 month CTO

    First action

    Collect every access and data statement from the support, Zoom and Terms pages and llms.txt into one draft.

    How to measure it

    Time to access

    Now
    –
    Target
    Under 7 days
    By
    1 quarter

    Steps

    1. 1

      Gather

      Every access and data statement already published.

      CTO (co-founder) 1 day

    2. 2

      Access table

      Slack and Zoom: read, write, approver.

      CTO (co-founder) 2 days

    3. 3

      Retention and deletion

      How long data is kept and how to remove it.

      CTO (co-founder) 1 day

    4. 4

      Sub-processors

      Confirm with counsel what to list.

      CEO (co-founder) 1 day

    5. 5

      Publish and link

      Footer, /audit, /logistics and the booking confirmation.

      CTO (co-founder) 1 day

    6. 6

      Answer pack

      Standard questionnaire answers, ready to send.

      CTO (co-founder) 1 week

    First result
    1 month
    Cost
    $0
    Owner
    CTO (co founder)

    Smallest test

    Publish only the per-integration access table.

    Why it works

    Publish /security: an access table per integration (what is read, what is written, who approves), approval boundaries, logging, data retention and deletion, sub-processors, and how to report a security issue. Add a short answer pack for common procurement questionnaires. Link it from the footer, /audit, the logistics page and the booking confirmation.

    An operations champion needs IT to sign off, and that step is often where AI pilots stall. A page they can forward shortens it.

    Risks and how to handle them

    • Overclaiming

      State only the controls in place today and mark plans as plans.

    Evidence

    • Your websiteSupport page (the homepage "Contact" link): "Tenor support." covers "Slack installation, permissions, message delivery, security reports, and data requests"; a Slack workspace admin must approve the Tenor app and its permissions; the only contact is an email address; no scheduling link.Source
    • Your websiteZoom integration page: Tenor reads completed Zoom cloud-recording transcripts so a workspace can use meeting context, with read-only access approved by a Zoom account admin; "Tenor does not create, modify, or delete Zoom users, meetings, or recordings." Customers can disconnect it themselves; data deletion is by email request.Source
    • Your websitellms.txt, the site's context file for AI agents and crawlers, has the fullest product description: "the operating layer for a company's AI workforce", "persistent AI that learns how your business operates and owns recurring work inside the tools you already use"; "A worker owns a responsibility end to end, not a task"; four stages: Audit ("A structured, read-only pass over how work moves across teams, systems, and decision points"), Design, Trial (the worker "replays historical work so its decisions can be reviewed before anything changes") and Deploy ("No migrations"); "Consequential actions require human sign-off until authority is explicitly expanded" and actions are logged; aimed at companies "that run high volumes of recurring operational work", including "large freight, logistics, and supply chain operators". It names no customers, deployment counts or metrics, and says "Backed by Y Combinator."Source
  • 12

    Tenor delivers its work through a Slack app that a workspace admin must approve, and its Terms and support page are written around that app. A web search found no public Slack Marketplace listing for Tenor. A listing gives admins a familiar, reviewed place to check the app and gives Tenor a page that people searching for AI workers in Slack can find.

    Medium impactMedium effort First result in 1 quarter CTO

    First action

    Check the app against Slack's Marketplace submission guidelines.

    How to measure it

    Marketplace-sourced Audits

    Now
    –
    Target
    3
    By
    2 quarters

    Steps

    1. 1

      Check the guidelines

      Scopes, privacy and review requirements.

      CTO (co-founder) 1 day

    2. 2

      Close the gaps

      Scopes, onboarding and policy links.

      Member of Technical Staff, Systems Engineering 1 week

    3. 3

      Write the listing

      One worker example, read-only start, screenshots.

      CPO (co-founder) 2 days

    4. 4

      Submit

      Answer reviewer questions quickly.

      CTO (co-founder) 1 day

    5. 5

      Link it

      From /support/ and the Audit onboarding email.

      CTO (co-founder) 1 hour

    6. 6

      Track

      Listing views, installs and Audits booked.

      CEO (co-founder) monthly

    First result
    1 quarter
    Cost
    About a week of engineering time
    Owner
    CTO (co founder)

    Smallest test

    Submit the app as it is and read the reviewer feedback.

    Why it works

    Check the app against Slack's Marketplace submission guidelines, close any gaps, and submit. Write the listing around one concrete worker example and the read-only start, and make the main call to action "Book an Audit" rather than an open self-serve install. Link the listing from /support/ and use it in the Audit onboarding email, so the admin approval step happens on a page they trust.

    Admins approve apps they can look up in the Marketplace more readily, and the listing is a free, high-intent discovery channel.

    Risks and how to handle them

    • Review takes weeks

      Submit early and keep the direct install working.

    • Self-serve installs Tenor can't support

      Route installs to booking an Audit.

    Evidence

    • Your websiteSupport page (the homepage "Contact" link): "Tenor support." covers "Slack installation, permissions, message delivery, security reports, and data requests"; a Slack workspace admin must approve the Tenor app and its permissions; the only contact is an email address; no scheduling link.Source
    • Your websiteTerms (last updated July 28, 2026): Tenor Technologies; they cover "Tenor's websites, connected services, and Slack app" and say Tenor delivers results through Slack and other connected services; fees "are described in your order, subscription, or other agreement with Tenor"; no public prices or free trial.Source
    • ResearchPublic footprint (checked November 10, 2026): domain registered May 3, 2026; the site links to no social profiles; no Hacker News mentions and no news coverage in the last 30 days; a web search found no public Slack Marketplace listing for Tenor.Source
  • 13

    Tenor already reads completed Zoom cloud-recording transcripts, read-only and approved by an admin. The Audit is a read-only pass over how work moves across teams. Weekly operations stand-ups and handover calls describe the exceptions, hand-offs and recurring chores in the team's own words, so they are the fastest map of the work a worker should own.

    High impactMedium effort First result in 1 month CPO

    First action

    Run the analysis on Tenor's own meeting transcripts.

    How to measure it

    Time to first finding

    Now
    –
    Target
    3 days
    By
    1 quarter

    Steps

    1. 1

      Dogfood

      Run it on Tenor's own meetings.

      Member of Technical Staff, Context & Memory 3 days

    2. 2

      Define the output

      Recurring task, who handles it, frequency, tools.

      CPO (co-founder) 2 days

    3. 3

      Write the consent

      Which meetings, what is kept, how to delete.

      CEO (co-founder) 1 day

    4. 4

      Pilot

      One prospect's Audit.

      CEO (co-founder) 2 weeks

    5. 5

      Fold into the Audit

      The map opens every Audit report.

      CPO (co-founder) 1 week

    6. 6

      Measure

      Days to first finding, before and after.

      CPO (co-founder) monthly

    First result
    1 month
    Cost
    Engineering time
    Owner
    CPO (co founder)

    Smallest test

    Run it on Tenor's own meetings this week.

    Why it works

    With the admin's choice of meetings, run the Audit's first pass on two weeks of operations transcripts: the recurring tasks mentioned, who handles them, how often they come up and which tools are named. Open the Audit report with that map. Dogfood it on Tenor's own meetings first, and use it to tune the Context & Memory work.

    Audits get faster, the customer sees its work described in its own words, and the engineering team gets real context data to build on.

    Risks and how to handle them

    • Concerns about meeting privacy

      The admin picks the meetings; access stays read-only, with deletion on request as the Zoom page already says.

    Builds on

    Evidence

    • Your websiteZoom integration page: Tenor reads completed Zoom cloud-recording transcripts so a workspace can use meeting context, with read-only access approved by a Zoom account admin; "Tenor does not create, modify, or delete Zoom users, meetings, or recordings." Customers can disconnect it themselves; data deletion is by email request.Source
    • Your websitellms.txt, the site's context file for AI agents and crawlers, has the fullest product description: "the operating layer for a company's AI workforce", "persistent AI that learns how your business operates and owns recurring work inside the tools you already use"; "A worker owns a responsibility end to end, not a task"; four stages: Audit ("A structured, read-only pass over how work moves across teams, systems, and decision points"), Design, Trial (the worker "replays historical work so its decisions can be reviewed before anything changes") and Deploy ("No migrations"); "Consequential actions require human sign-off until authority is explicitly expanded" and actions are logged; aimed at companies "that run high volumes of recurring operational work", including "large freight, logistics, and supply chain operators". It names no customers, deployment counts or metrics, and says "Backed by Y Combinator."Source
    • HiringCareers page: three open full-time roles, San Francisco first: Member of Technical Staff, Systems Engineering ("runtime, tools, and verification"); Context & Memory ("context, retrieval, and memory"); Evals & Task Engineering ("tasks, evaluation, and quality"); plus "Build Your Own Role". Applicants send one artifact and no cover letter; "This is not a normal coding job", and the page values verifying changes with logs, tests and user behaviour. The YC profile also lists a Founding Engineer, Agent and Data Systems ($150K-$200K, 0.50%-1.50% equity).Source
  • 14

    Tenor is hiring for "runtime, tools, and verification", "context, retrieval, and memory" and "tasks, evaluation, and quality", asks applicants for an artifact, and values verifying changes with logs, tests and user behaviour. The blog is empty. Operations buyers worry about exactly this: how do you know the worker is right, and how do you prove what it did?

    Medium impactMedium effort First result in 1 month CTO

    First action

    Pick the first topic: how a Trial replay is scored.

    How to measure it

    Applicants and bookings from posts

    Now
    –
    Target
    10 applicants and 2 bookings a month
    By
    1 quarter

    Steps

    1. 1

      Pick topics

      One per area for the first three months.

      CTO (co-founder) 1 day

    2. 2

      Write the first post

      How a Trial replay is scored.

      Member of Technical Staff, Evals & Task Engineering 1 week

    3. 3

      Review

      No customer data, internal endpoints or credentials.

      CTO (co-founder) 1 day

    4. 4

      Publish

      With a share image and both buttons.

      CTO (co-founder) 1 hour

    5. 5

      Share

      Hacker News, X, LinkedIn and the YC community.

      CEO (co-founder) 1 day

    6. 6

      Repeat

      Rotate authors each month.

      CTO (co-founder) monthly

    First result
    1 month
    Cost
    About two days of engineering time a month
    Owner
    CTO (co founder)

    Smallest test

    Turn one internal design note into a post.

    Why it works

    Publish one technical post a month, rotating across the three areas: how a worker's decisions are evaluated before deployment, how it remembers past decisions, how actions are verified and logged. End each post with both "Book a demo" and "Careers". Check every post before publishing: no customer data, internal endpoints or credentials.

    One piece of writing does recruiting and selling, and technical depth is what Hacker News and engineering-minded buyers respond to (idea 15).

    Risks and how to handle them

    • Distracts a small team

      One post a month, rotating authors.

    Evidence

    • HiringCareers page: three open full-time roles, San Francisco first: Member of Technical Staff, Systems Engineering ("runtime, tools, and verification"); Context & Memory ("context, retrieval, and memory"); Evals & Task Engineering ("tasks, evaluation, and quality"); plus "Build Your Own Role". Applicants send one artifact and no cover letter; "This is not a normal coding job", and the page values verifying changes with logs, tests and user behaviour. The YC profile also lists a Founding Engineer, Agent and Data Systems ($150K-$200K, 0.50%-1.50% equity).Source
    • Your websiteSitemap (checked November 10, 2026): 9 URLs, none with a lastmod date: home, mission, blog, one post, privacy, terms, support, support/zoom and llms.txt. The listed post, /blog/the-pilot-never-got-replaced/, shows only "Notes | Tenor" and "Read the latest notes from Tenor." with no article, and /blog/ says "New writing is on the way." There are no product, use-case, security, pricing or customer pages.Source
  • 15

    Tenor has a YC launch post, but we found no Hacker News mentions or news coverage, and the site links to no social profiles. The domain is six months old, so outside YC almost nobody has heard the story.

    Medium impactLow effort First result in 2 weeks CEO

    First action

    Add LinkedIn and X profile links to the site footer.

    How to measure it

    Launch-week bookings

    Now
    –
    Target
    10
    By
    2 weeks after launch

    Steps

    1. 1

      Add social links

      LinkedIn and X in the footer.

      CTO (co-founder) 30 minutes

    2. 2

      Prepare the demo

      The sample replay report (idea 9).

      CPO (co-founder) 1 week

    3. 3

      Write the post

      What it does, how it's verified, what it can't do yet.

      CEO (co-founder) 2 days

    4. 4

      Ask the network

      Introductions to alumni with operations-heavy teams.

      CEO (co-founder) 1 day

    5. 5

      Launch

      Founders answer comments all day.

      CEO (co-founder) 1 day

    6. 6

      Follow up

      Reply to every booking and serious comment.

      CEO (co-founder) 1 week

    First result
    2 weeks
    Cost
    $0
    Owner
    CEO (co founder)

    Smallest test

    Post one YC network request for three introductions.

    Why it works

    Once the essay (idea 3) and the sample replay (idea 9) are live, post a Launch HN with a concrete demo and honest limits. Ask the YC network for introductions to alumni companies with operations-heavy teams. Add LinkedIn and X links to the site footer so the launch has somewhere to send followers.

    YC alumni are early adopters and warm introductions, and Hacker News brings links, candidates and a few technical buyers in one day.

    Risks and how to handle them

    • Harsh Hacker News feedback

      Lead with engineering detail and limits, not claims.

    Builds on

    Evidence

    • Y CombinatorYC company profile: Summer 2026 batch, founded 2026, team of 3 (CEO, CPO and CTO co-founders), San Francisco, tagged B2B, Automation and AI. One-liner: "Infrastructure for building and scaling AI-native businesses." Launch post, "Tenor: AI that owns jobs end-to-end and is accountable for business outcomes": companies can measure what AI tokens cost but not what they produce, so Tenor ties each piece of AI work to a defined outcome and budget, tracks spend against results, and uses the results to decide where future AI resources go.Source
    • ResearchPublic footprint (checked November 10, 2026): domain registered May 3, 2026; the site links to no social profiles; no Hacker News mentions and no news coverage in the last 30 days; a web search found no public Slack Marketplace listing for Tenor.Source
  • 16

    llms.txt says Tenor states no customer names, deployment counts or metrics. Its logistics rivals publish DHL, Lineage Logistics and "2 million tasks at >99% accuracy". At this stage, a few named design partners with measured results close most of that gap.

    High impactHigh effort First result in 1 quarter CEO

    First action

    List 20 operations-heavy companies from the founders' and YC networks.

    How to measure it

    Case studies

    Now
    –
    Target
    2
    By
    2 quarters

    Steps

    1. 1

      List targets

      20 companies from the founders' and YC networks.

      CEO (co-founder) 2 days

    2. 2

      Make the offer

      Free Audit, discounted Trial, publishing rights.

      CEO (co-founder) 2 weeks

    3. 3

      Sign three

      Three different recurring jobs.

      CEO (co-founder) 1 month

    4. 4

      Run the work

      Audit, Trial and Deploy.

      Founding Engineer, Agent and Data Systems 6 weeks

    5. 5

      Measure

      Baseline before, outcome and spend after.

      CPO (co-founder) ongoing

    6. 6

      Publish

      Case study, quote and the replay numbers.

      CEO (co-founder) 2 weeks

    First result
    1 quarter
    Cost
    Discounted work for three partners
    Owner
    CEO (co founder)

    Smallest test

    Make the offer to five operators this month.

    Why it works

    Offer three operations-heavy logistics companies a free Audit and a discounted Trial in exchange for permission to publish results and take reference calls. Pick three different recurring jobs so each partner teaches something new. Measure a baseline before the worker starts, then outcome and spend after, and publish each story (named, or described by industry and size if the partner prefers).

    Design partners turn the thesis into proof and supply the templates for the big bets.

    Risks and how to handle them

    • Partners won't be named

      Agree up front to describe them by industry and size.

    Builds on

    Evidence

    • Your websitellms.txt, the site's context file for AI agents and crawlers, has the fullest product description: "the operating layer for a company's AI workforce", "persistent AI that learns how your business operates and owns recurring work inside the tools you already use"; "A worker owns a responsibility end to end, not a task"; four stages: Audit ("A structured, read-only pass over how work moves across teams, systems, and decision points"), Design, Trial (the worker "replays historical work so its decisions can be reviewed before anything changes") and Deploy ("No migrations"); "Consequential actions require human sign-off until authority is explicitly expanded" and actions are logged; aimed at companies "that run high volumes of recurring operational work", including "large freight, logistics, and supply chain operators". It names no customers, deployment counts or metrics, and says "Backed by Y Combinator."Source
    • Competitor researchRivals' own homepages: HappyRobot sells AI agents it calls "AI workers" that "automate communication-heavy work at enterprise scale" in logistics and seven other industries, citing a Series C, "15M+" interactions a month and DHL as a customer; Pallet sells AI systems to "the companies that run the physical world", citing Lineage Logistics: "Automated 2 million tasks at >99% accuracy"; both lead with "Book a demo". Lindy sells an "AI employee" with plans that "start at $29.99 per user per month" and "$50 in free credits".Source
  • 17

    Tenor's pilots start read-only and "surface findings with evidence before anything changes". The Slack app needs admin approval, and the support page exists partly because installs fail and workers can't see channels. The gap between install and first finding is when an internal champion decides whether to keep pushing.

    High impactMedium effort First result in 1 month CPO

    First action

    Measure install-to-first-finding time for the current pilots.

    How to measure it

    Median days to first finding

    Now
    –
    Target
    7
    By
    1 quarter

    Steps

    1. 1

      Measure today

      Install to first finding for each pilot.

      CPO (co-founder) 1 day

    2. 2

      Write the plan

      Day 1, day 3, day 7 deliverables.

      CPO (co-founder) 2 days

    3. 3

      Install check

      Detect missing scopes and channels automatically.

      Member of Technical Staff, Systems Engineering 1 week

    4. 4

      Findings template

      Finding, evidence link, proposed worker.

      CPO (co-founder) 2 days

    5. 5

      Stall alert

      Notify the founders when a step is late.

      CTO (co-founder) 2 days

    6. 6

      Review

      Every pilot's timeline, weekly.

      CEO (co-founder) weekly

    First result
    1 month
    Cost
    Engineering time
    Owner
    CPO (co founder)

    Smallest test

    Send a day-7 findings note to the current pilot.

    Why it works

    Set a seven-day plan for every new workspace: day 1, an automated check that the right scopes and channels are visible; day 3, a first map of the work; day 7, a findings note in Slack with evidence links and a proposed worker. Alert the founders if an install stalls at any step.

    Early, concrete value keeps champions selling internally and makes the move from Audit to Trial an easy decision.

    Risks and how to handle them

    • Rushed, shallow findings

      Ship one solid finding rather than many weak ones.

    Evidence

    • Your websitellms.txt, the site's context file for AI agents and crawlers, has the fullest product description: "the operating layer for a company's AI workforce", "persistent AI that learns how your business operates and owns recurring work inside the tools you already use"; "A worker owns a responsibility end to end, not a task"; four stages: Audit ("A structured, read-only pass over how work moves across teams, systems, and decision points"), Design, Trial (the worker "replays historical work so its decisions can be reviewed before anything changes") and Deploy ("No migrations"); "Consequential actions require human sign-off until authority is explicitly expanded" and actions are logged; aimed at companies "that run high volumes of recurring operational work", including "large freight, logistics, and supply chain operators". It names no customers, deployment counts or metrics, and says "Backed by Y Combinator."Source
    • Your websiteSupport page (the homepage "Contact" link): "Tenor support." covers "Slack installation, permissions, message delivery, security reports, and data requests"; a Slack workspace admin must approve the Tenor app and its permissions; the only contact is an email address; no scheduling link.Source
  • 18

    The YC launch post says Tenor ties each piece of AI work to an outcome and a budget and tracks spend against results; llms.txt says actions are logged and consequential steps need sign-off. That is exactly the data for a weekly report a buyer can forward to their own boss, and it is the moment they decide to give the worker more authority or a second job.

    High impactMedium effort First result in 1 month CPO

    First action

    Write the note for one worker by hand this Monday.

    How to measure it

    Average workers per customer

    Now
    –
    Target
    2
    By
    2 quarters

    Steps

    1. 1

      Write one by hand

      For one worker, this Monday.

      CPO (co-founder) 1 hour

    2. 2

      Define the fields

      Work, escalations, approvals, spend, outcome.

      CPO (co-founder) 2 days

    3. 3

      Automate

      Generated from the action log.

      Founding Engineer, Agent and Data Systems 1 week

    4. 4

      Approve in the note

      Approve or reject waiting actions from Slack.

      Member of Technical Staff, Systems Engineering 1 week

    5. 5

      Monthly roll-up

      For the economic buyer, with a suggested next job.

      CPO (co-founder) 1 week

    6. 6

      Track expansion

      Authority added and new workers.

      CEO (co-founder) monthly

    First result
    1 month
    Cost
    Engineering time
    Owner
    CPO (co founder)

    Smallest test

    Send one hand-written note this Monday and ask the manager if it helped.

    Why it works

    Every Monday, each worker posts a five-line note to its human manager in Slack: responsibilities handled, exceptions escalated, approvals waiting (approvable from the note), spend against budget, and outcome against target. Once a month, roll the notes up for the economic buyer, with a suggested next responsibility.

    Value that is visible every week keeps renewals safe and makes expansion (more authority, more workers) the obvious next step.

    Risks and how to handle them

    • Becomes noise

      Five lines with links; nothing on quiet weeks.

    Evidence

    • Y CombinatorYC company profile: Summer 2026 batch, founded 2026, team of 3 (CEO, CPO and CTO co-founders), San Francisco, tagged B2B, Automation and AI. One-liner: "Infrastructure for building and scaling AI-native businesses." Launch post, "Tenor: AI that owns jobs end-to-end and is accountable for business outcomes": companies can measure what AI tokens cost but not what they produce, so Tenor ties each piece of AI work to a defined outcome and budget, tracks spend against results, and uses the results to decide where future AI resources go.Source
    • Your websitellms.txt, the site's context file for AI agents and crawlers, has the fullest product description: "the operating layer for a company's AI workforce", "persistent AI that learns how your business operates and owns recurring work inside the tools you already use"; "A worker owns a responsibility end to end, not a task"; four stages: Audit ("A structured, read-only pass over how work moves across teams, systems, and decision points"), Design, Trial (the worker "replays historical work so its decisions can be reviewed before anything changes") and Deploy ("No migrations"); "Consequential actions require human sign-off until authority is explicitly expanded" and actions are logged; aimed at companies "that run high volumes of recurring operational work", including "large freight, logistics, and supply chain operators". It names no customers, deployment counts or metrics, and says "Backed by Y Combinator."Source
    • Your websiteTerms (last updated July 28, 2026): Tenor Technologies; they cover "Tenor's websites, connected services, and Slack app" and say Tenor delivers results through Slack and other connected services; fees "are described in your order, subscription, or other agreement with Tenor"; no public prices or free trial.Source

Big bets 3

Larger plays that can change the trajectory.

  • 19

    Tenor's YC launch argues that companies can measure what AI tokens cost but not what they produce, and says Tenor tracks spend against results. Rivals sell "AI workers" and an "AI employee" and lead with volume or accuracy numbers; none of their homepages leads with per-job accountability. A ledger buyers can open and finance can export turns the thesis into something they can see every day.

    High impactHigh effort First result in 1 quarter CTO

    First action

    Define the ledger fields using the first pilot's data.

    How to measure it

    Monthly active ledger viewers per customer

    Now
    –
    Target
    3
    By
    2 quarters

    Steps

    1. 1

      Define the fields

      Outcome, budget, actions, approvals, cost per outcome.

      CTO (co-founder) 1 week

    2. 2

      Instrument

      Capture every action, cost and result.

      Founding Engineer, Agent and Data Systems 1 month

    3. 3

      Slack view

      A ledger summary on demand in Slack.

      Member of Technical Staff, Systems Engineering 2 weeks

    4. 4

      Finance export

      Monthly CSV for the buyer's finance team.

      Founding Engineer, Agent and Data Systems 1 week

    5. 5

      Pilot

      With one customer for a month.

      CEO (co-founder) 1 month

    6. 6

      Publish the method

      How cost per outcome is calculated.

      CPO (co-founder) 2 weeks

    First result
    1 quarter
    Cost
    A quarter of engineering focus
    Owner
    CTO (co founder)

    Smallest test

    Build the ledger as a spreadsheet for one customer.

    Why it works

    Build a ledger view in Slack and on the web: for each worker, the outcome target, the budget, actions taken, approvals requested and given, exceptions, errors caught and cost per outcome, with an export for finance. Publish how cost per outcome is calculated. Use the ledger in renewals, in pricing (idea 20) and as the headline product screen.

    It makes Tenor the system of record for AI work, which is far harder to replace than any single agent, and answers the CFO's question directly.

    Risks and how to handle them

    • Disputes over what counts as an outcome

      Agree the outcome definition in the Audit report.

    Builds on

    Evidence

    • Y CombinatorYC company profile: Summer 2026 batch, founded 2026, team of 3 (CEO, CPO and CTO co-founders), San Francisco, tagged B2B, Automation and AI. One-liner: "Infrastructure for building and scaling AI-native businesses." Launch post, "Tenor: AI that owns jobs end-to-end and is accountable for business outcomes": companies can measure what AI tokens cost but not what they produce, so Tenor ties each piece of AI work to a defined outcome and budget, tracks spend against results, and uses the results to decide where future AI resources go.Source
    • Your websitellms.txt, the site's context file for AI agents and crawlers, has the fullest product description: "the operating layer for a company's AI workforce", "persistent AI that learns how your business operates and owns recurring work inside the tools you already use"; "A worker owns a responsibility end to end, not a task"; four stages: Audit ("A structured, read-only pass over how work moves across teams, systems, and decision points"), Design, Trial (the worker "replays historical work so its decisions can be reviewed before anything changes") and Deploy ("No migrations"); "Consequential actions require human sign-off until authority is explicitly expanded" and actions are logged; aimed at companies "that run high volumes of recurring operational work", including "large freight, logistics, and supply chain operators". It names no customers, deployment counts or metrics, and says "Backed by Y Combinator."Source
    • Competitor researchRivals' own homepages: HappyRobot sells AI agents it calls "AI workers" that "automate communication-heavy work at enterprise scale" in logistics and seven other industries, citing a Series C, "15M+" interactions a month and DHL as a customer; Pallet sells AI systems to "the companies that run the physical world", citing Lineage Logistics: "Automated 2 million tasks at >99% accuracy"; both lead with "Book a demo". Lindy sells an "AI employee" with plans that "start at $29.99 per user per month" and "$50 in free credits".Source
  • 20

    Tenor's Trial replays historical work before anything changes, and its launch post ties every piece of AI work to an outcome and a budget. Together they make a measurable guarantee possible: the buyer sees on its own history whether the worker would have met the bar before paying to deploy it. Fees today are set privately in each order.

    High impactMedium effort First result in 1 quarter CEO

    First action

    Set the pass bar for one type of worker.

    How to measure it

    Trials that become paid deployments

    Now
    –
    Target
    60%
    By
    2 quarters

    Steps

    1. 1

      Define the bar

      Metrics and the history window, fixed before the replay.

      CPO (co-founder) 1 week

    2. 2

      Contract terms

      Order-form wording, reviewed by counsel.

      CEO (co-founder) 2 weeks

    3. 3

      Replay scoring

      Automated, repeatable scoring against the bar.

      Member of Technical Staff, Evals & Task Engineering 1 month

    4. 4

      Offer to three

      The next three Trials.

      CEO (co-founder) 1 month

    5. 5

      Review

      Pass rates, margins and disputes.

      CEO (co-founder) quarterly

    6. 6

      Publish

      On the pricing page (idea 10).

      CEO (co-founder) 1 week

    First result
    1 quarter
    Cost
    Possible unpaid Trials
    Owner
    CEO (co founder)

    Smallest test

    Offer it to the next prospect who hesitates on price.

    Why it works

    Write a pass bar into each Audit report (for example, agreement with human decisions, hours saved and errors caught on a fixed window of history). If the Trial replay misses the bar, there is no deployment fee; if it passes, the worker is priced on the outcome it owns, within an agreed budget. Keep the Audit paid (idea 7) so the guarantee doesn't strain cash.

    It takes the risk off the buyer, sets Tenor apart from per-seat agent pricing, and forces the team to deploy only workers that work.

    Risks and how to handle them

    • Disputes over whether the bar was met

      Fix the bar and the data window in writing before the replay.

    • Cash flow

      Keep the Audit paid.

    Builds on

    Evidence

    • Your websitellms.txt, the site's context file for AI agents and crawlers, has the fullest product description: "the operating layer for a company's AI workforce", "persistent AI that learns how your business operates and owns recurring work inside the tools you already use"; "A worker owns a responsibility end to end, not a task"; four stages: Audit ("A structured, read-only pass over how work moves across teams, systems, and decision points"), Design, Trial (the worker "replays historical work so its decisions can be reviewed before anything changes") and Deploy ("No migrations"); "Consequential actions require human sign-off until authority is explicitly expanded" and actions are logged; aimed at companies "that run high volumes of recurring operational work", including "large freight, logistics, and supply chain operators". It names no customers, deployment counts or metrics, and says "Backed by Y Combinator."Source
    • Y CombinatorYC company profile: Summer 2026 batch, founded 2026, team of 3 (CEO, CPO and CTO co-founders), San Francisco, tagged B2B, Automation and AI. One-liner: "Infrastructure for building and scaling AI-native businesses." Launch post, "Tenor: AI that owns jobs end-to-end and is accountable for business outcomes": companies can measure what AI tokens cost but not what they produce, so Tenor ties each piece of AI work to a defined outcome and budget, tracks spend against results, and uses the results to decide where future AI resources go.Source
    • Your websiteTerms (last updated July 28, 2026): Tenor Technologies; they cover "Tenor's websites, connected services, and Slack app" and say Tenor delivers results through Slack and other connected services; fees "are described in your order, subscription, or other agreement with Tenor"; no public prices or free trial.Source
  • 21

    Tenor targets freight, logistics and supply chain operators with high volumes of recurring work, and today each worker is built in a Design stage. A three-person team hiring for runtime, memory and evals can't design every worker from scratch, and rivals with bigger teams already sell to the same buyers.

    High impactHigh effort First result in 1 quarter CPO

    First action

    Rank the jobs from booking answers and Audits by how often they come up.

    How to measure it

    Median days, Audit to Deploy

    Now
    –
    Target
    Under 21
    By
    2 quarters

    Steps

    1. 1

      Rank the jobs

      From booking answers, Audits and design partners.

      CPO (co-founder) 1 week

    2. 2

      Pick three

      Frequency times value.

      CEO (co-founder) 1 day

    3. 3

      Build the first template

      Plan, boundaries and tool connections.

      Founding Engineer, Agent and Data Systems 1 month

    4. 4

      Evals per template

      Replay datasets and pass bars.

      Member of Technical Staff, Evals & Task Engineering 1 month

    5. 5

      Memory defaults

      What the worker remembers between decisions.

      Member of Technical Staff, Context & Memory 2 weeks

    6. 6

      Market

      One page per template, linked from /logistics.

      CEO (co-founder) 2 weeks

    First result
    1 quarter
    Cost
    Two quarters of engineering focus
    Owner
    CPO (co founder)

    Smallest test

    Reuse one worker's plan for a second customer and time the Design stage.

    Why it works

    Rank the recurring jobs that come up in booking answers (idea 5), Audits and design partners (idea 16), and pick the three most common. Build a template for each: a default operating plan, approval boundaries, evals, memory defaults and a replay dataset. Design then covers only the last 20% for each customer. Give each template its own page linked from /logistics.

    Deployments get faster, margins improve, and marketing becomes concrete: three named jobs instead of "AI infrastructure".

    Risks and how to handle them

    • Templates don't fit real variation

      Keep Design for the last 20% of each deployment.

    Builds on

    Evidence

    • Your websitellms.txt, the site's context file for AI agents and crawlers, has the fullest product description: "the operating layer for a company's AI workforce", "persistent AI that learns how your business operates and owns recurring work inside the tools you already use"; "A worker owns a responsibility end to end, not a task"; four stages: Audit ("A structured, read-only pass over how work moves across teams, systems, and decision points"), Design, Trial (the worker "replays historical work so its decisions can be reviewed before anything changes") and Deploy ("No migrations"); "Consequential actions require human sign-off until authority is explicitly expanded" and actions are logged; aimed at companies "that run high volumes of recurring operational work", including "large freight, logistics, and supply chain operators". It names no customers, deployment counts or metrics, and says "Backed by Y Combinator."Source
    • HiringCareers page: three open full-time roles, San Francisco first: Member of Technical Staff, Systems Engineering ("runtime, tools, and verification"); Context & Memory ("context, retrieval, and memory"); Evals & Task Engineering ("tasks, evaluation, and quality"); plus "Build Your Own Role". Applicants send one artifact and no cover letter; "This is not a normal coding job", and the page values verifying changes with logs, tests and user behaviour. The YC profile also lists a Founding Engineer, Agent and Data Systems ($150K-$200K, 0.50%-1.50% equity).Source
    • Competitor researchRivals' own homepages: HappyRobot sells AI agents it calls "AI workers" that "automate communication-heavy work at enterprise scale" in logistics and seven other industries, citing a Series C, "15M+" interactions a month and DHL as a customer; Pallet sells AI systems to "the companies that run the physical world", citing Lineage Logistics: "Automated 2 million tasks at >99% accuracy"; both lead with "Book a demo". Lindy sells an "AI employee" with plans that "start at $29.99 per user per month" and "$50 in free credits".Source

Wildcards 2

Unexpected experiments, each with a cheap way to test it.

  • 22

    The one post title in Tenor's sitemap, "The Pilot Never Got Replaced", names a common buyer frustration, and Tenor's model (an owner, an outcome, a budget, a read-only path to trust) is a direct answer to it. Tenor has had no news coverage, and original data is the easiest way for a new company to get quoted.

    Medium impactMedium effort First result in 1 quarter CEO

    First action

    Write ten survey questions.

    How to measure it

    Respondent bookings

    Now
    –
    Target
    10
    By
    2 quarters

    Steps

    1. 1

      Write the questions

      Ten questions, five minutes.

      CEO (co-founder) 2 days

    2. 2

      Recruit

      YC network, LinkedIn and logistics operator groups.

      CEO (co-founder) 3 weeks

    3. 3

      Analyse

      Stall reasons by company size and industry.

      CPO (co-founder) 1 week

    4. 4

      Write

      Findings plus the CEO's essay.

      CEO (co-founder) 1 week

    5. 5

      Publish

      A web page with charts and a share image.

      CTO (co-founder) 1 day

    6. 6

      Pitch

      Trade press and operations newsletters.

      CEO (co-founder) 2 weeks

    First result
    1 quarter
    Cost
    Founder time; optional small incentive for respondents
    Owner
    CEO (co founder)

    Smallest test

    Ask ten operations leaders three questions and publish what they say.

    Why it works

    Survey 50-100 operations leaders: how many AI pilots they have run, how many reached production, and why the rest stalled (no owner, no measurement, IT access, accuracy, cost). Publish the findings with simple charts, an essay from the CEO and a booking button, and offer every respondent a free Audit call.

    Data gets cited by newsletters, investors and buyers, and every respondent is a qualified prospect who has just thought hard about the problem Tenor solves.

    Risks and how to handle them

    • Too few answers

      Publish as an interview series if fewer than 50 respond.

    Builds on

    Evidence

    • Your websiteSitemap (checked November 10, 2026): 9 URLs, none with a lastmod date: home, mission, blog, one post, privacy, terms, support, support/zoom and llms.txt. The listed post, /blog/the-pilot-never-got-replaced/, shows only "Notes | Tenor" and "Read the latest notes from Tenor." with no article, and /blog/ says "New writing is on the way." There are no product, use-case, security, pricing or customer pages.Source
    • Y CombinatorYC company profile: Summer 2026 batch, founded 2026, team of 3 (CEO, CPO and CTO co-founders), San Francisco, tagged B2B, Automation and AI. One-liner: "Infrastructure for building and scaling AI-native businesses." Launch post, "Tenor: AI that owns jobs end-to-end and is accountable for business outcomes": companies can measure what AI tokens cost but not what they produce, so Tenor ties each piece of AI work to a defined outcome and budget, tracks spend against results, and uses the results to decide where future AI resources go.Source
    • ResearchPublic footprint (checked November 10, 2026): domain registered May 3, 2026; the site links to no social profiles; no Hacker News mentions and no news coverage in the last 30 days; a web search found no public Slack Marketplace listing for Tenor.Source
  • 23

    Tenor's homepage says "We work with people who know their industries too well to leave them as they are", the Mission page says "Built around real expertise", and the careers page invites candidates to "Build Your Own Role". The team's listed backgrounds are engineering and research (Stanford, UC Berkeley, TUM, the IMO, Amazon, Tesla, Waymo), so deep logistics operations know-how has to come from outside.

    High impactMedium effort First result in 1 quarter CEO

    First action

    Write a one-paragraph "Operator in residence" role and post it on the careers page.

    How to measure it

    Operator-sourced Audits

    Now
    –
    Target
    3
    By
    2 quarters

    Steps

    1. 1

      Define the role

      Time, equity or revenue share, expectations.

      CEO (co-founder) 2 days

    2. 2

      Post it

      On the careers page and in the YC network.

      CEO (co-founder) 1 day

    3. 3

      Recruit

      One or two former operations leaders.

      CEO (co-founder) 1 month

    4. 4

      Onboard

      Sit in on two Audits.

      CPO (co-founder) 1 week

    5. 5

      Co-design

      Shape the worker templates (idea 21).

      CPO (co-founder) ongoing

    6. 6

      Co-author

      Logistics page and the stalled-pilot report.

      CEO (co-founder) quarterly

    First result
    1 quarter
    Cost
    Advisor-level equity or a share of revenue
    Owner
    CEO (co founder)

    Smallest test

    Ask one former operations leader for a paid month of advice.

    Why it works

    Recruit one or two former logistics operations leaders as operators in residence: part-time, paid in equity or a share of the deals they source. They sit in on Audits, shape the worker templates (idea 21), open doors to their networks and co-author content (ideas 8 and 22). Post the role under "Build Your Own Role" so the homepage line becomes a programme.

    Operators bring credibility, introductions and judgement about which jobs matter, and they turn the homepage's promise into something visible.

    Risks and how to handle them

    • Equity cost with little return

      Tie grants to sourced Audits and time spent.

    Evidence

    • Your websiteHomepage (checked November 10, 2026): H1 "The scale of a company can now be engineered." and one line of product copy, "Our AI infrastructure makes expertise scale like software, allowing what makes a company exceptional to make it enormous."; "We work with people who know their industries too well to leave them as they are."; team backgrounds from Stanford, UC Berkeley, Technical University of Munich, International Mathematical Olympiad, Amazon, Tesla and Waymo. The only links are Careers, Contact (which opens /support/), Privacy and Terms: no "Book a demo" button, calendar link, product, customer or pricing section.Source
    • Your websiteMission page: "The infrastructure for companies built around AI." with four principles ("From capable models to capable companies.", "Built around real expertise.", "Learn from outcomes.", "Build for the long term."). It holds the site's only "Book a demo →" button, linking to a cal.com team scheduler; the homepage does not link to this page.Source
    • HiringCareers page: three open full-time roles, San Francisco first: Member of Technical Staff, Systems Engineering ("runtime, tools, and verification"); Context & Memory ("context, retrieval, and memory"); Evals & Task Engineering ("tasks, evaluation, and quality"); plus "Build Your Own Role". Applicants send one artifact and no cover letter; "This is not a normal coding job", and the page values verifying changes with logs, tests and user behaviour. The YC profile also lists a Founding Engineer, Agent and Data Systems ($150K-$200K, 0.50%-1.50% equity).Source

03

How Programmable can help

Where an idea in this playbook is something Programmable builds or runs, it's listed here with what we'd do. Everything else is yours to run.

ProgrammableDigital

Show it, don't just describe it

These ideas need something to show: a demo, a video, an animation, graphics or a deck. Digital turns screenshots or the site into all of them, in the company's own brand.

Make a demo (opens programmable.digital in a new tab)
More example playbooks and reviews

More growth playbooks

Other companies we have researched, each read from what they are hiring for.

  • Sona growth strategyAn AI-native workforce management platform for frontline organizations: demand forecasting, auto-scheduling, time and attendance, HR…20 growth ideas
  • SavvyMoney growth strategyEmbedded credit score and financial wellness software for banks, credit unions and fintechs: credit score and monitoring, personalized…22 growth ideas
  • GLG growth strategyThe expert network: expert calls (including AI-moderated calls), surveys, an Expert Content Library, events, advisory and board…21 growth ideas
  • Ashna Rana (a.r.) growth strategyFreelance conversion rate optimization and creative strategy: on-site testing programs, landing pages, UX research and design, data…21 growth ideas
  • flowmondo growth strategyA UK AI and automation consultancy (Compound Digital Technologies Ltd) that builds integrations, process and document automation, custom…22 growth ideas
  • Micro-Biz Consulting & Services (MBCS) growth strategyA Tampa consultancy that builds and runs a small business's digital system: Joomla websites, SEO, social media, website maintenance and…23 growth ideas
  • Fabcon Precast growth strategyA precast concrete manufacturer and installer (since 1971) of insulated wall panels, total precast structures and concrete repair (W2E)…20 growth ideas
  • Pax Historia growth strategyA browser-based alternate-history sandbox game where AI runs the world: players pick a country and a moment in history, type their…21 growth ideas

See every playbook and website review

Independent research. Programmable wrote this playbook from public sources, as they stood on November 10, 2026, and cites them throughout. Tenor did not commission, review or endorse it, and Programmable is not affiliated with Tenor. Company and product names are trademarks of their owners, used only to identify them. The ideas are suggestions and opinion, not professional, legal or financial advice, and not statements about how Tenor runs its business.

Something wrong, or do you represent Tenor? Email [email protected] and we will correct it or take it down.

30/60/90 roadmap

What to focus on in each month, and the ideas that carry it.

  1. Days 1 to 307 ideas

    Make the site book calls: the homepage demo button, the llms.txt pitch on the homepage, the empty essay, share previews, booking questions and lighter hero media; submit the Slack Marketplace listing.

  2. Days 31 to 6011 ideas

    Package and prove: the paid Audit, the logistics page, the sample replay report, the pricing model, the trust page, Zoom-led Audits, engineering posts, the HN and YC launch, design partners, seven-day findings and the weekly worker note.

  3. Days 61 to 905 ideas

    Commit to the big bets (the outcome ledger, replay-based pricing, logistics worker templates) and the wildcards (the stalled-pilot report and operators in residence).

The research behind the plan

3 research areas, each summarised in a few lines. Open any of them for the full detail.

Customers

What customers say

Customers value A read-only start: findings with evidence before anything changes. The most cited problem: AI pilots that never reach production.

  • Works inside existing tools, with no migrations
  • Knowing what AI costs but not what it produces
  • In their words: “owns recurring work”

Market

Competitive landscape

3 competitors compared: HappyRobot, Pallet and Lindy.

  • Gap to exploit: A read-only, replay-first way to prove an AI worker on the buyer's own history
  • HappyRobot: Built around communication; less about owning a back-office responsibility end to end
  • Pallet: Enterprise sales motion; no public low-risk first step

Hiring signals

Tenor's hiring is all engineering, and it maps to the product's hardest promises: Systems Engineering for "runtime, tools, and verification" (ideas 12, 17, 18, 19), Context & Memory for "context, retrieval, and memory"…

  • Member of Technical Staff: Systems Engineering: Runtime, tools and verification. Ideas 12, 17, 18.
  • Member of Technical Staff: Context & Memory: Context, retrieval and memory. Ideas 13, 21.
  • Member of Technical Staff: Evals & Task Engineering: Tasks, evaluation and quality. Ideas 9, 14, 20.

From the team that wrote this playbook

All products
  • ProgrammableAgencyWorking AI systems in days, not quarters (opens programmable.agency in a new tab)
  • ProgrammableInfoAn agent that finds, researches and asks firstNew accounts get 100,000 free enrichment credits with code BETA100K (opens programmable.info in a new tab)
  • ProgrammableDigitalScreenshots in, an animated product demo out (opens programmable.digital in a new tab)
  • ProgrammableWatchKnow the day a supplier falls out of compliance (opens programmable.watch in a new tab)
  • ProgrammableTalkGet registrants for your events, and know who’s coming (opens programmable.talk in a new tab)
  • ProgrammableMediaAn AI note taker, interpreter and webinar host for your calls (opens programmable.media in a new tab)
  • ProgrammablePageText what you need, your customer gets a page that works (opens programmable.page in a new tab)
  • ProgrammableCommunityForums with labelled AI members, never empty (opens programmable.community in a new tab)
ProgrammableMarketing
ExamplesLeaderboardsProductsTermsPrivacyCookiesDPALegalContact

© 2026 AI-Powered Growth Intelligence