ProgrammableMarketing/promptless.ai (opens in a new tab)
PDF

Views

  • Plan
  • Roadmap
  • Research

On this page

  • The brief
  • Start here
  • Idea portfolio
  • Effort vs impact
  • This week · 7
  • This quarter · 12
  • Big bets · 3
  • Wildcards · 2
  • How we can help
Generated Nov 10, 2026 1 view
  1. Programmable
  2. Playbooks
  3. Promptless
Promptless logo

growth playbook · Nov 10, 2026

promptless.ai (opens in a new tab)

Promptless growth playbook

Diagnosis

Promptless is a five-person YC W25 team with a clear product, real public proof (Mautic merged 129 Promptless docs PRs in four months and cut code-to-docs time from 98 to 27 days; Vellum gets 54% of its docs PRs from Promptless) and a second product for governing coding-agent instructions. Its pitch is that docs must work for AI agents, and its free Agent Score tool grades exactly that.

Yet when an AI reader asks for the Blog, the Changelog, the Agent Score tool page or the Starport docs, the site returns 404. The growth gaps are mostly in the funnel around a strong product: a demo page with one email field and no proof, one customer story on the site, no price for Agent Instructions or Starport, no comparison pages while Mintlify, GitBook, Warp and DeepDocs move into the same job, and no ROI report for docs customers.

This playbook fixes agent access first, then turns Agent Score, the open-source PR record and the pilot benchmarks into pipeline, and bets on Starport, DoGBench and a self-serve instruction-debt diagnostic.

Growth Score

7/10

Biggest Opportunity

Promptless already has the rarest asset in this category, public and countable proof (merged PRs on Vitess, Helm, Bazel, Mautic and Runpod docs) plus a free tool that grades every prospect's docs. Turning both into the front door of the funnel, for people and for AI assistants, can carry a five-person team a long way.

Biggest Risk

Docs platforms are adding their own update agents (Mintlify's Autopilot proposes doc PRs on paid plans; GitBook flags stale content), and a Launch HN commenter argued AI coding tools could add the feature easily. If Promptless cannot show measured results quickly, it risks being treated as a feature of the platform it publishes to.

01

Start here

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

  1. 1

    Your Blog, Changelog and Agent Score pages return 404 to AI readers: fall back to the HTML page whenever no markdown version exists

    Request /blog, /changelog, /agent-score and /docs/for-docs/starport with a markdown Accept header, confirm the 404s, and find the rule that returns them.

    High impact First result in 1 week
  2. 2

    Turn the demo page into a booking page: a calendar, the logos and the Mautic numbers next to the email field

    Count how many /meet email submissions turned into booked calls last month.

    High impact First result in 2 weeks
  3. 3

    Publish Mautic's 98-to-27-day result as a Promptless customer story, and add a story page for each open-source project

    Ask Mautic's Education Team for permission to republish and quote its results.

    High impact First result in 1 month

02

Your idea portfolio

24 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.

  • 8 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 · 4Big bets · 8Fill-ins · 6Reconsider · 6
Effort →

Showing 24 of 24 ideas

This week 7

Quick wins with a clear first action.

  • 01

    Promptless serves AI readers a markdown version of its pages, which is exactly what it sells. But when we fetched the site with an AI-agent reader on November 9, 2026, /blog, /changelog and the Agent Score tool page /agent-score all returned HTTP 404, with and without a trailing slash and on repeat fetches, and the Starport docs page answered "This page does not exist in its markdown form." The same reader loaded /pricing, every individual post, every monthly changelog, /agent-score.md and public Agent Score reports without trouble. So the pages that show momentum (25 monthly changelogs since October 2024) and the free tool at the top of the funnel are the ones an assistant cannot open. These URLs are in the sitemap and the navigation, and the homepage invites visitors to ask AI assistants about Promptless. An assistant asked "what has Promptless shipped lately?" or "run Agent Score on my docs" hits an error and gets its answer somewhere else.

    High impactLow effort First result in 1 week Co founder & CTO

    First action

    Request /blog, /changelog, /agent-score and /docs/for-docs/starport with a markdown Accept header, confirm the 404s, and find the rule that returns them.

    How to measure it

    Agent-readable sitemap coverage

    Now
    –
    Target
    100% of 234 URLs
    By
    1 week

    Steps

    1. 1

      Reproduce

      Request each of the 234 sitemap URLs as a browser and as a markdown-requesting agent, and list every one that differs in status code.

      Founding Engineer 2 hours

    2. 2

      Find the rule

      Locate the content-negotiation or rewrite rule that returns 404 when no .md file exists.

      Co-founder & CTO 2 hours

    3. 3

      Add the fallback

      Serve the HTML page with a 200 whenever the markdown version is missing, instead of an error.

      Founding Engineer half a day

    4. 4

      Generate the missing markdown

      Build blog and changelog index pages from the existing llms.txt lists, map /agent-score to agent-score.md, and export the Starport overview.

      Founding Engineer 1 day

    5. 5

      Guard it in the build

      Add a deploy check that requests every sitemap URL as an agent and fails on any 404, alongside the afdocs checks Agent Score already runs.

      Co-founder & CTO half a day

    6. 6

      Fix parity

      Close the two markdown and HTML parity failures Agent Score reports on the governance and triggers docs.

      Forward-deployed technical writer 1 day

    7. 7

      Score the whole site

      Run Agent Score on promptless.ai as well as /docs and publish the report.

      Co-founder & CEO 1 hour

    8. 8

      Watch the referrals

      Track visits from AI assistants to the blog, changelog and Agent Score in the analytics tool, weekly.

      Co-founder & CEO weekly

    First result
    1 week
    Cost
    About two engineer-days
    Owner
    Co founder & CTO

    Smallest test

    Fix /agent-score alone today and confirm an AI reader can open it.

    Why it works

    Change the rule that serves markdown so that it never returns 404 when only the markdown version is missing: return the normal HTML page with a 200, or a markdown page generated from what already exists. Blog and changelog listings can be built from /blog/llms.txt and /changelog/llms.txt, /agent-score can return the content of /agent-score.md, and Starport needs its overview page exported. Then add a check to the deploy pipeline that requests every sitemap URL the way an agent does and fails the build on any 404. Finally, run Agent Score on promptless.ai itself (not only /docs) and publish the A+ report as proof that the company follows its own advice.

    Promptless's own blog and Agent Score argue that agents are now a primary reader of docs and product sites. Every page that answers an agent with an error is a missed citation in an assistant's answer, and it contradicts the product's own pitch in front of the most technical buyers. The fix is small, can be tested in an afternoon, and protects every future page.

    Ready to use

    ChecklistAgent-readable site check

    [ ] /blog returns 200 to a markdown-requesting agent
    [ ] /changelog returns 200 to a markdown-requesting agent
    [ ] /agent-score returns the tool description (agent-score.md)
    [ ] /docs/for-docs/starport has a markdown version
    [ ] No sitemap URL returns 404 to agents but 200 to browsers
    [ ] Deploy check fails on any agent-only 404
    [ ] Markdown and HTML parity fixed on the 2 flagged docs pages
    [ ] Agent Score run on promptless.ai, not only /docs

    Risks and how to handle them

    • Serving HTML to agents breaks the markdown promise

      Treat HTML as the fallback only; generate markdown for the listing pages so agents get the clean version.

    • Assistants cache the old 404s

      Resubmit the sitemap and link the fixed pages from llms.txt so crawlers fetch them again.

    Evidence

    • Your websiteAI-reader check (November 9, 2026, each URL fetched at least twice, with and without a trailing slash): /blog, /changelog and the Agent Score tool page /agent-score returned HTTP 404, and /docs/for-docs/starport returned "404 – Page Not Found" with "This page does not exist in its markdown form." The same reader got full content from /pricing, /pricing.md, /agent-score.md, individual blog posts, each monthly changelog and public Agent Score reports. /blog/, /changelog/ and /agent-score/ are in the sitemap, Blog and Changelog are in the site navigation, and llms.txt lists Starport and Agent Score as pages.Source
    • Your websiteSitemap (234 URLs, checked November 9, 2026): 58 blog posts, of which one is a customer story (Vellum); 25 monthly changelog entries from October 2024 to October 2026; 135 docs pages; one use-case page (/use-cases/technical-writer-power-tool/); free tools Agent Score and a Broken Link Report; event pages /lavacon/, /bingo/ and /bingo-leaderboard/; no comparison or alternatives pages.Source
    • Your websiteAgent Score: a free tool, "no signup", that scores from 0 to 100 how well AI agents can read a docs site, using 28 checks in 7 categories from the open-source afdocs project; every report is public, with fixes, score history and an SVG badge. The leaderboard lists 23 scored sites. Promptless's own docs score 100 (A+, scan of October 19, 2026) but fail the markdown and HTML parity check on 2 of 50 pages. Agent Score "checks infrastructure, not content quality", and the page points to a content strategy audit with a free 14-day trial.Source
    • Your websiteHomepage: headline "Continuously improve your AI workforce and docs" with two products, Promptless for Docs ("Write the docs. Skip the busywork.") and Promptless for Agent Instructions; 14 customer logos including Megaport, Mezmo, Vitess, Helm, Aptible, Runpod, Flatfile, Latitude, Mautic and Vellum; testimonials from a Senior Technical Writer at Runpod ("My team literally calls me a 10x tech writer"), the CTO of Latitude.sh ("basically pays for itself right away") and the Head of Developer Docs at Prove ("a solo tech writer's godsend"); buttons that open AI assistants to ask about Promptless; "Every plan includes a 14-day free trial"; Agent Instructions claims typical 30-day results of token spend down 18%, first-attempt completion up 15%, wall-clock time down 32% and human interruptions down 42%, with "white-glove onboarding with a forward-deployed AI engineer and technical writer".Source
  • 02

    Every "Get a demo" button leads to /meet, titled "Book a 15-minute demo", but the page has only a work-email field and a "Send" button: no calendar, no logos, no testimonials and no results. A buyer who has decided to book still has to wait for an email reply, and the page gives them nothing to share with their manager.

    High impactLow effort First result in 2 weeks Co founder & CEO

    First action

    Count how many /meet email submissions turned into booked calls last month.

    How to measure it

    Booked calls from /meet

    Now
    –
    Target
    +40%
    By
    60 days

    Steps

    1. 1

      Measure today

      Visits, email submissions and booked calls from /meet over the last 60 days.

      Co-founder & CEO 1 hour

    2. 2

      Add the scheduler

      Show a 15-minute calendar after the email is entered, routed by product.

      Founding Engineer 1 day

    3. 3

      Add proof

      Logos, the Mautic result and one testimonial beside the form.

      Co-founder & CEO 2 hours

    4. 4

      Keep the fallback

      Email-only option for visitors who are not ready to book.

      Founding Engineer 1 hour

    5. 5

      Confirm by email

      Send a calendar invite and a two-line prep note with each booking.

      Co-founder & CEO 2 hours

    6. 6

      Compare

      Booked-call rate before and after, monthly.

      Co-founder & CEO monthly

    First result
    2 weeks
    Cost
    Scheduler subscription and one engineer-day
    Owner
    Co founder & CEO

    Smallest test

    Add a scheduling link to the confirmation message today.

    Why it works

    Embed a scheduler that opens after the email field (passing the product choice from ?content= so the right founder takes the call), and put three things beside it: the customer logos, one line of Mautic's result (code-to-docs time cut from 98 to 27 days) and one testimonial. Keep "Not ready to book? Leave your email" as the fallback.

    Developer-tool buyers who reach a demo page are already convinced enough to act, and losing them to an email round-trip is the most common leak in founder-led sales. Proof next to the form also reassures the second person in the buying decision.

    Risks and how to handle them

    • Calendar fills with poor-fit calls

      Ask two short questions (docs pages, docs tool) before the calendar shows.

    Evidence

    • Your websiteDemo page /meet: titled "Book a 15-minute demo"; its only element is a "Work email" field with a "Send" button and the line "Not ready to book? Leave your email and we'll follow up"; no calendar, logos, testimonials or results on the page.Source
    • PressMautic's own write-up (September 3, 2026): 129 Promptless docs PRs merged in about four months, against 392 human PRs over four-plus years; in a sample, Promptless updates took 27 days from code merge to docs merge versus 98 days by hand (about 3.5 times faster); the docs, stuck at version 5.2 while the app was at 7.0, began catching up. Human review stays: "a fast first draft, with a person making the final call."Source
    • Your websiteHomepage: headline "Continuously improve your AI workforce and docs" with two products, Promptless for Docs ("Write the docs. Skip the busywork.") and Promptless for Agent Instructions; 14 customer logos including Megaport, Mezmo, Vitess, Helm, Aptible, Runpod, Flatfile, Latitude, Mautic and Vellum; testimonials from a Senior Technical Writer at Runpod ("My team literally calls me a 10x tech writer"), the CTO of Latitude.sh ("basically pays for itself right away") and the Head of Developer Docs at Prove ("a solo tech writer's godsend"); buttons that open AI assistants to ask about Promptless; "Every plan includes a 14-day free trial"; Agent Instructions claims typical 30-day results of token spend down 18%, first-attempt completion up 15%, wall-clock time down 32% and human interruptions down 42%, with "white-glove onboarding with a forward-deployed AI engineer and technical writer".Source
  • 03

    Mautic published its own account of using Promptless in September 2026: 129 Promptless docs PRs merged in about four months against 392 human PRs in four-plus years, and docs that land 27 days after a code merge instead of 98. That is the strongest third-party proof Promptless has, yet promptless.ai has exactly one customer story (Vellum) and links only to raw PR lists for Vitess, Helm, Runpod, Mautic and Bazel.

    High impactLow effort First result in 1 month Forward deployed technical wr…

    First action

    Ask Mautic's Education Team for permission to republish and quote its results.

    How to measure it

    Demos from story pages

    Now
    –
    Target
    15 a month
    By
    90 days

    Steps

    1. 1

      Get permission

      Confirm Mautic is happy to be quoted and linked.

      Co-founder & CEO 2 days

    2. 2

      Write the Mautic story

      Problem, workflow, the 129 PRs and 98-to-27 days, and how human review works.

      Forward-deployed technical writer 2 days

    3. 3

      Pull the OSS numbers

      Merged Promptless PRs and time to merge for Vitess, Helm, Runpod and Bazel docs.

      Founding Engineer 1 day

    4. 4

      Write short stories

      One page per project with three example PRs.

      Forward-deployed technical writer 1 week

    5. 5

      Place the proof

      Numbers on the homepage, pricing and demo pages.

      Founding Engineer 1 day

    6. 6

      Share

      Send each story to the project's maintainers to share with their community.

      Co-founder & CEO 1 week

    First result
    1 month
    Cost
    Writing time only
    Owner
    Forward deployed technical writer

    Smallest test

    Publish the Mautic story alone and track its traffic for a month.

    Why it works

    With Mautic's permission, publish a Promptless story that summarises its write-up and links to it, leading with the two numbers. Then write a short story page for each open-source project in llms.txt, built from its public merged-PR history (count, median time to merge, examples), and add the Vellum and Mautic numbers to the homepage and pricing page.

    Third-party, countable results beat testimonials with technical buyers, and the open-source PR record is public, so anyone can check the numbers. Each story page is also a page AI assistants can cite when someone asks whether doc agents work.

    Risks and how to handle them

    • Projects dislike being used in marketing

      Ask first and let maintainers review every page.

    Evidence

    • PressMautic's own write-up (September 3, 2026): 129 Promptless docs PRs merged in about four months, against 392 human PRs over four-plus years; in a sample, Promptless updates took 27 days from code merge to docs merge versus 98 days by hand (about 3.5 times faster); the docs, stuck at version 5.2 while the app was at 7.0, began catching up. Human review stays: "a fast first draft, with a person making the final call."Source
    • Your websiteSitemap (234 URLs, checked November 9, 2026): 58 blog posts, of which one is a customer story (Vellum); 25 monthly changelog entries from October 2024 to October 2026; 135 docs pages; one use-case page (/use-cases/technical-writer-power-tool/); free tools Agent Score and a Broken Link Report; event pages /lavacon/, /bingo/ and /bingo-leaderboard/; no comparison or alternatives pages.Source
    • Your websiteOpen-source program: free for eligible non-commercial projects (CNCF and Linux Foundation given as examples), applied for by email, with no listed criteria; llms.txt links public lists of Promptless pull requests on Vitess, Helm, Runpod, Mautic and Bazel docs repositories.Source
  • 04

    The docs tiers have real prices, but all three Agent Instructions tiers say "Contact us", Starport (a full docs site with hosting) is not on the pricing page at all, and the page has no FAQ. On Launch HN, pricing without a sales call was the most pressed question. Technical buyers who can't see a number often assume it is too high and leave.

    Medium impactLow effort First result in 1 month Co founder & CEO

    First action

    Write down the lowest price you would accept for an Agent Instructions Startup deal.

    How to measure it

    Pricing-page conversion

    Now
    –
    Target
    +20%
    By
    60 days

    Steps

    1. 1

      Set floors

      Starting prices for the three Agent Instructions tiers and Starport.

      Co-founder & CEO 1 day

    2. 2

      Write the FAQ

      Page counting, trial end, OSS, data processing, supported tools.

      Forward-deployed technical writer 1 day

    3. 3

      Add Starport

      A line with what is included and the starting price.

      Founding Engineer half a day

    4. 4

      Mirror in markdown

      Update pricing.md so assistants quote the same numbers.

      Founding Engineer 1 hour

    5. 5

      Brief calls

      Use the published floor as the anchor on demos.

      Co-founder & CEO 1 hour

    6. 6

      Measure

      Pricing-page conversion and deal size, monthly.

      Co-founder & CEO monthly

    First result
    1 month
    Cost
    $0
    Owner
    Co founder & CEO

    Smallest test

    Add the FAQ this week and the prices next week.

    Why it works

    Add a "from" price to each Agent Instructions tier (it is already banded by instruction count), add Starport as a line or add-on with its starting price, and add a short FAQ: how pages are counted, what happens after the trial, the open-source program, where data is processed, and which docs tools are supported.

    A starting price qualifies buyers before the call and shortens it, and a five-person team can't afford calls that end at the price.

    Risks and how to handle them

    • A public floor caps enterprise deals

      Publish "from" prices and keep Enterprise custom.

    Evidence

    • Your websitePricing page: docs Startup "$500/mo" for up to 200 pages with Slack and GitHub, unlimited contributors and a "Sign up" button; Growth priced by page bundle from $500-$1,000/mo (200-500 pages) to $2,000-$4,000/mo (2,000-5,000 pages), adding Linear, Jira, Confluence and Notion, translations and Promptless Capture screenshots, behind "Book demo"; Enterprise custom. All three Agent Instructions tiers (under 50, 50-200 and over 200 instructions) say "Contact us". "All plans include a 14-day free trial." The page has no FAQ and does not mention Starport; it links an OSS program for non-commercial projects.Source
    • PressLaunch HN (February 18, 2025, 107 points, 61 comments): the most pressed theme was pricing without a sales call ("let me sign up and give it a spin"); commenters also asked about free use for open-source maintainers, accuracy checks and keeping private discussions out of public docs, and one argued AI IDEs could add the feature easily.Source
    • Your websiteStarport: a docs site built on the open-source Astro Starlight framework, shipping llms.txt files, a read-only MCP server for AI clients, copy-as-Markdown page actions, agent-readiness scoring, broken-link checks, redirects, site search, PDF export, Doc Detective testing and hosting on Vercel. No price is shown, and the docs include migration guides (choose a platform, migrate, preserve URLs and redirects).Source
  • 05

    Agent Score is free, needs no signup and publishes every report, and it says itself that it "checks infrastructure, not content quality". The person reading a report has just learned their docs are hard for agents to read, and the natural next question, whether the content is still correct, is exactly what Promptless answers. The pilot already starts by running on the last 30 days of pull requests.

    High impactLow effort First result in 1 month Founding Engineer

    First action

    Count Agent Score scans in the last 30 days and how many came from new domains.

    How to measure it

    Report-to-trial rate

    Now
    –
    Target
    5% of scans
    By
    60 days

    Steps

    1. 1

      Measure scans

      Scans, unique domains and repeat visitors.

      Founding Engineer 1 hour

    2. 2

      Write the block

      What the score covers, what it can't see, and the 30-day offer.

      Forward-deployed technical writer half a day

    3. 3

      Build it

      Add the block to reports and their .md versions.

      Founding Engineer 1 day

    4. 4

      Prefill sign-up

      Carry the scanned domain into the trial.

      Founding Engineer 1 day

    5. 5

      Follow up

      Founder email to every report owner who starts a trial.

      Co-founder & CEO ongoing

    6. 6

      Track

      Scan to trial to paid, monthly.

      Co-founder & CEO monthly

    First result
    1 month
    Cost
    Two engineer-days
    Owner
    Founding Engineer

    Smallest test

    Add a plain-text link to the trial at the end of reports for two weeks.

    Why it works

    Add a closing block to every report: "Your score covers structure. It can't tell whether your docs match your product." Offer to run Promptless on the site's last 30 days of PRs and show the suggestions it would have made, with a one-click trial and a demo link. Pass the scanned domain into the sign-up so the trial starts on the right repository.

    It turns a free tool into a funnel at the moment of highest intent, using a hand-off the tool already admits it needs, and it uses the pilot's own opening move as the offer.

    Risks and how to handle them

    • Reports feel like ads

      Keep the block short and below the results.

    Builds on

    Evidence

    • Your websiteAgent Score: a free tool, "no signup", that scores from 0 to 100 how well AI agents can read a docs site, using 28 checks in 7 categories from the open-source afdocs project; every report is public, with fixes, score history and an SVG badge. The leaderboard lists 23 scored sites. Promptless's own docs score 100 (A+, scan of October 19, 2026) but fail the markdown and HTML parity check on 2 of 50 pages. Agent Score "checks infrastructure, not content quality", and the page points to a content strategy audit with a free 14-day trial.Source
    • Your websiteRun a pilot: 14 days for Growth and Enterprise, with the clock starting at the first suggestion; a 90-minute onboarding workshop after Promptless runs on historical events such as the last 30 days of pull requests; weekly 15-30 minute check-ins. Benchmarks: 20-40 suggestions a month for one writer, 30-100 for teams of 2-5 writers, under 20 a month suggests the team may be too early; quality target at least 8/10.Source
  • 06

    Every Agent Score report already has an SVG badge, and the leaderboard has 23 sites, several scoring A or better. Docs teams are proud of a good score, and a badge in a docs footer or README is a permanent link from exactly the kind of site Promptless sells to.

    Medium impactLow effort First result in 1 month Co founder & CEO

    First action

    List leaderboard sites scoring A or above.

    How to measure it

    Sites displaying the badge

    Now
    –
    Target
    25
    By
    90 days

    Steps

    1. 1

      List high scorers

      Every scanned site at A or above.

      Founding Engineer 1 hour

    2. 2

      Make the snippet

      HTML and markdown badge linking to the report.

      Founding Engineer 2 hours

    3. 3

      Add the button

      "Get your badge" on A-grade reports.

      Founding Engineer half a day

    4. 4

      Email teams

      Short founder note with the snippet.

      Co-founder & CEO 1 day

    5. 5

      Rescan monthly

      Keep badges accurate.

      Founding Engineer monthly

    6. 6

      Count

      Badges live and scans referred from them.

      Co-founder & CEO monthly

    First result
    1 month
    Cost
    $0
    Owner
    Co founder & CEO

    Smallest test

    Email the top five sites this week.

    Why it works

    Email each team that scores A or above with their badge snippet (linking to their public report) and a one-line "how we score agents" note. Add a "Get your badge" button to high-scoring reports, and rescan monthly so the badge stays current.

    Badges spread because they flatter the publisher; each one is a backlink and an invitation for that site's readers to scan their own docs.

    Risks and how to handle them

    • Scores drop and badges embarrass sites

      Show the latest score and send a fix list when it falls.

    Evidence

    • Your websiteAgent Score: a free tool, "no signup", that scores from 0 to 100 how well AI agents can read a docs site, using 28 checks in 7 categories from the open-source afdocs project; every report is public, with fixes, score history and an SVG badge. The leaderboard lists 23 scored sites. Promptless's own docs score 100 (A+, scan of October 19, 2026) but fail the markdown and HTML parity check on 2 of 50 pages. Agent Score "checks infrastructure, not content quality", and the page points to a content strategy audit with a free 14-day trial.Source
  • 17

    On desktop the homepage scores 100, but on mobile it scores 78: in Google's lab test the headline takes 3.8 seconds to appear, with about 3.4 seconds of render-blocking requests and 344 KB of fonts. Real visitors load the main content in 2.4 seconds, which passes, but interaction responsiveness (INP) is 282 ms, so the page fails Core Web Vitals on INP. The analytics script alone is 247 KB with 288 ms of blocking time.

    Medium impactLow effort First result in 1 month Founding Engineer

    First action

    Defer the analytics script and re-test.

    How to measure it

    Field INP (p75)

    Now
    –
    Target
    under 200 ms
    By
    60 days

    Steps

    1. 1

      Defer analytics

      Load after interaction or when idle.

      Founding Engineer 2 hours

    2. 2

      Fix fonts

      Subset and preload, or a system font for the headline.

      Founding Engineer half a day

    3. 3

      Inline critical CSS

      Remove render-blocking stylesheets.

      Founding Engineer half a day

    4. 4

      Lazy-load media

      Video and images below the fold.

      Founding Engineer 2 hours

    5. 5

      Check the tabs

      Make product-tab switching respond quickly on phones.

      Founding Engineer half a day

    6. 6

      Re-test

      PageSpeed now and field data after 28 days.

      Co-founder & CTO 1 month

    First result
    1 month
    Cost
    About two engineer-days
    Owner
    Founding Engineer

    Smallest test

    Defer the analytics script on the homepage only.

    Why it works

    Inline the critical CSS, preload or subset the fonts (or use a system font for the headline), and load the analytics script after the first interaction or when idle. Check that the explainer video and screenshots are lazy-loaded below the fold. Re-run PageSpeed and watch real-visitor INP over 28 days.

    Developer tools are often first opened on a phone from a social or community link; a fast first screen and responsive tabs keep that visitor reading.

    Risks and how to handle them

    • Deferred analytics misses early events

      Queue events until the script loads.

    Evidence

    • PageSpeedPageSpeed (homepage, November 9, 2026): mobile performance 78, desktop 100; mobile lab LCP 3.8s with about 3.4s of render-blocking requests and 344 KB of fonts in a 1,453 KB page; real visitors (field data) see LCP 2.4s and CLS 0 but INP 282 ms, so the page fails Core Web Vitals on INP; PostHog loads 247 KB with 288 ms of blocking time.Source

This quarter 12

The core plays for the next three months.

  • 07

    The sitemap has no comparison or alternatives pages, while rivals now describe the same job on their own homepages: Mintlify's "Always-on agent for self-updating docs", GitBook's stale-content detection, Warp's Documentation Agent and DeepDocs. A Launch HN commenter argued coding tools could add the feature easily. Buyers are searching for exactly these comparisons, and today someone else writes them.

    High impactMedium effort First result in 2 months Forward deployed technical wr…

    First action

    Ask the last ten prospects which alternative they considered.

    How to measure it

    Comparison-page demos

    Now
    –
    Target
    10 a month
    By
    90 days

    Steps

    1. 1

      Pick five

      The alternatives prospects mention most.

      Co-founder & CEO 1 day

    2. 2

      Check facts

      Each rival's current features and price from its own pages.

      Forward-deployed technical writer 2 days

    3. 3

      Write pages

      Fair tables, where each wins, and switching help.

      Forward-deployed technical writer 2 weeks

    4. 4

      Add the DIY page

      What a home-made action needs to match review, style and citations.

      Forward-deployed AI engineer 3 days

    5. 5

      Publish with markdown

      HTML and .md versions, listed in llms.txt.

      Founding Engineer 1 day

    6. 6

      Refresh

      Re-check every quarter and update the date.

      Forward-deployed technical writer quarterly

    First result
    2 months
    Cost
    Writing time only
    Owner
    Forward deployed technical writer

    Smallest test

    Publish the DIY GitHub Action page first.

    Why it works

    Write one honest page per alternative: what triggers updates, which context sources it reads (Slack, tickets, Jira, Linear), whether it works across docs platforms, screenshots, style learning, citations, review workflow and price. Include the DIY page (a coding agent in a GitHub Action) with what it takes to maintain. Date each page and link to the rival's own page.

    Comparison searches are the highest-intent searches in a new category, and platform neutrality (Promptless publishes to Mintlify, GitBook, ReadMe and Docusaurus) is a real difference that only a comparison makes visible.

    Risks and how to handle them

    • Partners such as docs platforms feel attacked

      Frame platform pages as "use both" where Promptless publishes to them.

    Evidence

    • Competitor researchRivals' own pages: Mintlify lists "Autopilot: Always-on agent for self-updating docs" and says that on paid plans "the agent proposes documentation updates through pull requests", with a free Starter tier; GitBook offers stale-content detection that "flags it for review" and gap tracking; Warp's Documentation Agent watches merged PRs and opens doc PRs for Mintlify or GitBook; DeepDocs is a GitHub agent, freemium with paid plans up to $25 a month (AlternativeTo).Source
    • Your websiteSitemap (234 URLs, checked November 9, 2026): 58 blog posts, of which one is a customer story (Vellum); 25 monthly changelog entries from October 2024 to October 2026; 135 docs pages; one use-case page (/use-cases/technical-writer-power-tool/); free tools Agent Score and a Broken Link Report; event pages /lavacon/, /bingo/ and /bingo-leaderboard/; no comparison or alternatives pages.Source
    • PressLaunch HN (February 18, 2025, 107 points, 61 comments): the most pressed theme was pricing without a sales call ("let me sign up and give it a spin"); commenters also asked about free use for open-source maintainers, accuracy checks and keeping private discussions out of public docs, and one argued AI IDEs could add the feature easily.Source
  • 08

    Promptless's own ROI guide says the docs product has "no built-in ROI dashboard or metrics export": customers piece together doc updates shipped, share of doc PRs and review throughput from three places, and estimate time saved themselves. Yet the October 2026 changelog shipped an Impact page for Agent Instructions that shows time and tokens recovered per fix. The best docs numbers Promptless has (Vellum's 54% of PRs, Mautic's 98 to 27 days) were calculated by hand.

    High impactMedium effort First result in 1 quarter Co founder & CTO

    First action

    Calculate code-to-docs time for three current customers by hand, as Mautic did.

    How to measure it

    Docs renewals with an Impact summary sent

    Now
    –
    Target
    90% of renewals
    By
    2 quarters

    Steps

    1. 1

      Pick metrics

      Merged suggestions, share of doc PRs, code-to-docs days, hours saved.

      Co-founder & CTO 2 days

    2. 2

      Compute baselines

      Pre-Promptless code-to-docs time from git history.

      Founding Engineer 1 week

    3. 3

      Build the page

      Reuse the Agent Instructions Impact layout.

      Founding Engineer 2 weeks

    4. 4

      Add the summary

      Monthly email the champion can forward.

      Founding Engineer 1 week

    5. 5

      Review with customers

      Walk through the page at check-ins.

      Forward-deployed technical writer ongoing

    6. 6

      Use in renewals

      Send the summary 30 days before each renewal.

      Co-founder & CEO ongoing

    First result
    1 quarter
    Cost
    About a month of engineering
    Owner
    Co founder & CTO

    Smallest test

    Send three customers a hand-made impact email this month.

    Why it works

    Build a docs Impact page from data Promptless already holds: suggestions merged, share of the repo's doc PRs, median days from code merge to docs merge (before and after Promptless), triggers caught, and estimated hours saved with an editable per-update estimate. Add a one-click monthly summary the champion can forward to their manager before renewal.

    Champions renew what they can prove to their boss, and the same numbers become case studies without extra interviews.

    Risks and how to handle them

    • Hours saved looks inflated

      Let customers set the per-update estimate and show ranges, as the ROI guide advises.

    Evidence

    • Your websiteReporting and ROI doc: Promptless for Docs has no built-in ROI dashboard or metrics export; customers assemble doc updates shipped, share of doc PRs, trigger coverage and review throughput from the suggestions list, triggers page and git history, and time saved needs their own estimate. By contrast the October 2026 changelog added an Impact page for Agent Instructions showing agent time and tokens recovered per session 30 days after each merged fix.Source
    • Your websiteCustomer story: at Vellum, 54% of docs PRs in the last month came from Promptless suggestions, 60% of the product and engineering team uses Promptless to draft doc updates each month, after about six months of use; Vellum plans to have Promptless generate personalised product update summaries for key clients. It is the only customer story on the site.Source
    • PressMautic's own write-up (September 3, 2026): 129 Promptless docs PRs merged in about four months, against 392 human PRs over four-plus years; in a sample, Promptless updates took 27 days from code merge to docs merge versus 98 days by hand (about 3.5 times faster); the docs, stuck at version 5.2 while the app was at 7.0, began catching up. Human review stays: "a fast first draft, with a person making the final call."Source
  • 09

    Guided pilots open with Promptless running on the last 30 days of PRs, so the first workshop reviews real suggestions; the pilot clock starts only at the first suggestion. Self-serve Startup sign-ups get a flat 14-day trial and have to wait for new PRs to see anything. A small team may open only a few doc-relevant PRs in two weeks, below the "under 20 a month" line the pilot guide calls too early.

    High impactMedium effort First result in 1 month Founding Engineer

    First action

    Measure how many days self-serve trials take to see their first suggestion today.

    How to measure it

    Self-serve trial conversion

    Now
    –
    Target
    +30%
    By
    2 quarters

    Steps

    1. 1

      Measure first value

      Days from sign-up to first suggestion for recent trials.

      Founding Engineer 1 day

    2. 2

      Reuse the pilot backfill

      Run the historical-events job automatically on connect.

      Founding Engineer 2 weeks

    3. 3

      Rank results

      Show the five strongest suggestions first.

      Forward-deployed AI engineer 1 week

    4. 4

      Move the clock

      Trial starts at the first suggestion.

      Founding Engineer 1 day

    5. 5

      Morning summary

      Email what Promptless found overnight.

      Forward-deployed technical writer 2 days

    6. 6

      Compare cohorts

      Conversion before and after.

      Co-founder & CEO monthly

    First result
    1 month
    Cost
    Engineering time and model compute per trial
    Owner
    Founding Engineer

    Smallest test

    Run the backfill by hand for the next ten self-serve sign-ups.

    Why it works

    As soon as a self-serve user connects GitHub and their docs repo, replay the last 30 days of merged PRs and show the suggestions Promptless would have made, ranked by confidence, with citations. Start the 14-day clock at the first suggestion, as in pilots, and email a summary the next morning.

    Time to first value decides trial conversion, and backfilled suggestions on the team's own code are the most convincing demo possible, with no new PRs needed.

    Risks and how to handle them

    • Compute cost on poor-fit sign-ups

      Cap the backfill at 30 PRs per trial.

    Evidence

    • Your websiteRun a pilot: 14 days for Growth and Enterprise, with the clock starting at the first suggestion; a 90-minute onboarding workshop after Promptless runs on historical events such as the last 30 days of pull requests; weekly 15-30 minute check-ins. Benchmarks: 20-40 suggestions a month for one writer, 30-100 for teams of 2-5 writers, under 20 a month suggests the team may be too early; quality target at least 8/10.Source
    • Your websitePricing page: docs Startup "$500/mo" for up to 200 pages with Slack and GitHub, unlimited contributors and a "Sign up" button; Growth priced by page bundle from $500-$1,000/mo (200-500 pages) to $2,000-$4,000/mo (2,000-5,000 pages), adding Linear, Jira, Confluence and Notion, translations and Promptless Capture screenshots, behind "Book demo"; Enterprise custom. All three Agent Instructions tiers (under 50, 50-200 and over 200 instructions) say "Contact us". "All plans include a 14-day free trial." The page has no FAQ and does not mention Starport; it links an OSS program for non-commercial projects.Source
  • 10

    llms.txt opens with "Set up Promptless from a coding agent" and says the MCP page covers agent sign-up and claiming, but the MCP docs require "an organization whose Promptless setup is finished" and an active membership. A developer who asks their coding agent to "set up Promptless for our docs" hits a wall, even though the server already offers 11 tools including submitting documentation tasks.

    Medium impactMedium effort First result in 1 quarter Co founder & CTO

    First action

    Fix the llms.txt wording today so it matches what the MCP page supports.

    How to measure it

    MCP-created trial organisations

    Now
    –
    Target
    20 a month
    By
    2 quarters

    Steps

    1. 1

      Align the wording

      Make llms.txt and the MCP page say the same thing.

      Forward-deployed technical writer 1 hour

    2. 2

      Design the claim flow

      Agent starts, human confirms by email or OAuth.

      Co-founder & CTO 3 days

    3. 3

      Build the tool

      A sign-up tool that returns a claim link.

      Founding Engineer 2 weeks

    4. 4

      Guide the setup

      Next tools for repo connection and first task.

      Founding Engineer 1 week

    5. 5

      Rate-limit

      Prevent abuse of trial creation.

      Co-founder & CTO 2 days

    6. 6

      Track

      Agent-created trials and their conversion.

      Co-founder & CEO monthly

    First result
    1 quarter
    Cost
    About three engineer-weeks
    Owner
    Co founder & CTO

    Smallest test

    Let agents request a trial by email address and handle the rest by hand.

    Why it works

    Add an unauthenticated sign-up tool to the MCP server that creates a trial organisation, returns a claim link for a human to confirm in the browser, and then walks the agent through connecting the docs repo. Document it on the MCP page and in llms.txt so the promise matches the product.

    Developers increasingly try tools by asking their agent to install them. Being the doc agent that can be set up in one conversation is a distribution channel competitors have not claimed.

    Risks and how to handle them

    • Spam sign-ups

      Require human confirmation before any repository access.

    Evidence

    • Your websiteMCP setup: the Promptless MCP server offers 11 tools (for example submit_documentation_task and request_changes) and the September 2026 changelog added following tasks and editing docs from the editor; prerequisites are "an organization whose Promptless setup is finished" and an active membership, so there is no way to sign up from a coding agent, although llms.txt says the page covers agent sign-up and claiming.Source
  • 11

    The Agent Score leaderboard has only 23 sites, mostly ones that came to be scanned. The tool runs in about a minute, every report is public and the history is kept. Scanning the docs of the 500 best-known developer products yourself turns a quiet tool into a dataset nobody else has.

    High impactMedium effort First result in 1 quarter Co founder & CEO

    First action

    Build the list of 500 docs sites by category.

    How to measure it

    Conversations from the index

    Now
    –
    Target
    40 per quarter
    By
    2 quarters

    Steps

    1. 1

      Build the list

      500 public developer docs sites by category.

      Co-founder & CEO 2 days

    2. 2

      Scan

      Batch the scans within rate limits.

      Founding Engineer 1 week

    3. 3

      Analyse

      Most common failures and category averages.

      Forward-deployed AI engineer 1 week

    4. 4

      Publish

      Index page with HTML and markdown versions.

      Forward-deployed technical writer 1 week

    5. 5

      Tell each team

      Personal email with their report and top fix.

      Co-founder & CEO 2 weeks

    6. 6

      Rescan quarterly

      Publish the biggest movers.

      Founding Engineer quarterly

    First result
    1 quarter
    Cost
    Compute and about three weeks of team time
    Owner
    Co founder & CEO

    Smallest test

    Scan 50 API docs sites and publish one post.

    Why it works

    Scan 500 public developer docs sites, publish an index by category (APIs, infrastructure, AI, fintech) with the most common failures and fixes, and email each team its report. Rescan every quarter and publish the movers. Each team that improves becomes a warm conversation about content accuracy, which Agent Score can't measure and Promptless can.

    A ranked index earns press, social shares and backlinks, and every site in it receives a personal reason to talk to Promptless.

    Risks and how to handle them

    • Low scorers feel shamed

      Lead with category averages and fixes, and let teams rescan before publication.

    Builds on

    Evidence

    • Your websiteAgent Score: a free tool, "no signup", that scores from 0 to 100 how well AI agents can read a docs site, using 28 checks in 7 categories from the open-source afdocs project; every report is public, with fixes, score history and an SVG badge. The leaderboard lists 23 scored sites. Promptless's own docs score 100 (A+, scan of October 19, 2026) but fail the markdown and HTML parity check on 2 of 50 pages. Agent Score "checks infrastructure, not content quality", and the page points to a content strategy audit with a free 14-day trial.Source
    • Your websiteBlog index (llms.txt): 39 technical posts (for example documentation debt, developer documentation ROI, agent context files, help center vs docs-as-code, support agent deflection and the DoGBench benchmark for doc agents), 15 product updates, one newsletter (September 2026) and one customer story; no posts carry dates.Source
  • 12

    Promptless built an "AI Tells Bingo" field game and leaderboard for LavaCon, and runs Write the Docs writing days. Technical writers are a small, tight community that meets at a handful of conferences, and a game with a public leaderboard gives them a reason to visit promptless.ai during the event.

    Medium impactMedium effort First result in 1 quarter Forward deployed technical wr…

    First action

    Count LavaCon players and how many left an email.

    How to measure it

    Event-sourced trials

    Now
    –
    Target
    10 per event
    By
    2 quarters

    Steps

    1. 1

      Review LavaCon

      Players, emails, follow-ups and trials.

      Co-founder & CEO 1 day

    2. 2

      Make it reusable

      Board and leaderboard per event slug.

      Founding Engineer 1 week

    3. 3

      Pick events

      Next three writing conferences.

      Forward-deployed technical writer 1 day

    4. 4

      Offer organisers

      Let them host the game for attendees.

      Co-founder & CEO 2 weeks

    5. 5

      Follow up

      Agent Score report and trial offer to players.

      Forward-deployed technical writer 1 week per event

    6. 6

      Measure

      Trials and deals per event.

      Co-founder & CEO per event

    First result
    1 quarter
    Cost
    Travel, prizes and about a week of engineering
    Owner
    Forward deployed technical writer

    Smallest test

    Offer the game to one Write the Docs meetup.

    Why it works

    Package the bingo game as a reusable kit: an event-specific board, a leaderboard per conference, a prize, and a follow-up email offering each player an Agent Score report and a trial. Run it at the next Write the Docs and technical-communication conferences, and invite organisers to host it themselves.

    Writers are the daily users and champions of Promptless; meeting them where they gather, with something playful, builds the word of mouth that the testimonials show.

    Risks and how to handle them

    • Players are not buyers

      Ask for company and docs tool on the leaderboard sign-up.

    Evidence

    • Your websiteSitemap (234 URLs, checked November 9, 2026): 58 blog posts, of which one is a customer story (Vellum); 25 monthly changelog entries from October 2024 to October 2026; 135 docs pages; one use-case page (/use-cases/technical-writer-power-tool/); free tools Agent Score and a Broken Link Report; event pages /lavacon/, /bingo/ and /bingo-leaderboard/; no comparison or alternatives pages.Source
    • Your websiteHomepage: headline "Continuously improve your AI workforce and docs" with two products, Promptless for Docs ("Write the docs. Skip the busywork.") and Promptless for Agent Instructions; 14 customer logos including Megaport, Mezmo, Vitess, Helm, Aptible, Runpod, Flatfile, Latitude, Mautic and Vellum; testimonials from a Senior Technical Writer at Runpod ("My team literally calls me a 10x tech writer"), the CTO of Latitude.sh ("basically pays for itself right away") and the Head of Developer Docs at Prove ("a solo tech writer's godsend"); buttons that open AI assistants to ask about Promptless; "Every plan includes a 14-day free trial"; Agent Instructions claims typical 30-day results of token spend down 18%, first-attempt completion up 15%, wall-clock time down 32% and human interruptions down 42%, with "white-glove onboarding with a forward-deployed AI engineer and technical writer".Source
  • 13

    Promptless's strongest testimonials come from individual writers ("a solo tech writer's godsend", "My team literally calls me a 10x tech writer"). Technical writers often work as the only writer on a team and move between developer-tool companies. Each move takes a convinced user to a new docs team that has never heard of Promptless.

    Medium impactLow effort First result in 1 quarter Co founder & CEO

    First action

    List every champion and their current company.

    How to measure it

    Champion-move pilots

    Now
    –
    Target
    4 per quarter
    By
    2 quarters

    Steps

    1. 1

      List champions

      Main user at every customer and OSS project.

      Co-founder & CEO 1 day

    2. 2

      Ask permission

      Opt-in to contact after a job move.

      Forward-deployed technical writer ongoing

    3. 3

      Watch for moves

      Monthly check of public profiles and bounced emails.

      Co-founder & CEO monthly

    4. 4

      Offer the pilot

      30 days with set-up done for them.

      Forward-deployed technical writer per move

    5. 5

      Thank them

      A gift or donation when the pilot converts.

      Co-founder & CEO per deal

    6. 6

      Track

      Pilots and revenue from moves.

      Co-founder & CEO quarterly

    First result
    1 quarter
    Cost
    Staff time and small thank-you gifts
    Owner
    Co founder & CEO

    Smallest test

    Email the champions you already know have moved.

    Why it works

    Keep a list of every active champion. When one changes company, send a congratulations note with a free 30-day pilot for the new team, set up by the forward-deployed writer, and a small thank-you when it converts. Ask champions at onboarding whether they are happy to be contacted after a move.

    A champion arriving at a new job is looking for quick wins and already knows the product works, which makes this the cheapest warm lead a docs tool can get.

    Risks and how to handle them

    • Feels intrusive

      Contact only champions who opted in.

    Evidence

    • Your websiteHomepage: headline "Continuously improve your AI workforce and docs" with two products, Promptless for Docs ("Write the docs. Skip the busywork.") and Promptless for Agent Instructions; 14 customer logos including Megaport, Mezmo, Vitess, Helm, Aptible, Runpod, Flatfile, Latitude, Mautic and Vellum; testimonials from a Senior Technical Writer at Runpod ("My team literally calls me a 10x tech writer"), the CTO of Latitude.sh ("basically pays for itself right away") and the Head of Developer Docs at Prove ("a solo tech writer's godsend"); buttons that open AI assistants to ask about Promptless; "Every plan includes a 14-day free trial"; Agent Instructions claims typical 30-day results of token spend down 18%, first-attempt completion up 15%, wall-clock time down 32% and human interruptions down 42%, with "white-glove onboarding with a forward-deployed AI engineer and technical writer".Source
    • Your websiteBlog index (llms.txt): 39 technical posts (for example documentation debt, developer documentation ROI, agent context files, help center vs docs-as-code, support agent deflection and the DoGBench benchmark for doc agents), 15 product updates, one newsletter (September 2026) and one customer story; no posts carry dates.Source
  • 14

    The pilot guide sets clear health lines: 20-40 suggestions a month for one writer, 30-100 for teams of 2-5, quality of at least 8/10, and "under 20 a month suggests the team may be too early". Those lines are used to judge pilots but not, as far as the docs show, to watch paying accounts after the pilot ends.

    Medium impactLow effort First result in 1 month Forward deployed technical wr…

    First action

    Run the benchmark against every paying account this week.

    How to measure it

    Healthy accounts

    Now
    –
    Target
    85%
    By
    2 quarters

    Steps

    1. 1

      Set bands

      Per account by number of writers.

      Forward-deployed technical writer 1 day

    2. 2

      Compute weekly

      Suggestions, acceptance and closed-without-merge rates.

      Founding Engineer 3 days

    3. 3

      Alert

      Slack alert after two weeks below band.

      Founding Engineer 1 day

    4. 4

      Tune up

      Check triggers, filters and sources with the customer.

      Forward-deployed technical writer per alert

    5. 5

      Share

      Show the band on the customer's Impact page.

      Founding Engineer 1 week

    6. 6

      Review

      Churn against alerts each quarter.

      Co-founder & CEO quarterly

    First result
    1 month
    Cost
    A few engineer-days
    Owner
    Forward deployed technical writer

    Smallest test

    Check this month's numbers by hand for every account.

    Why it works

    Compute suggestions per month and acceptance rate for every account weekly. When an account drops below its band for two weeks, the forward-deployed writer checks triggers, noise filters and connected sources, and books a 15-minute tune-up. Share the same numbers with the customer on the Impact page.

    Usage drops weeks before cancellation, and a small team can only save accounts it sees early.

    Risks and how to handle them

    • Low volume is normal for quiet release months

      Compare against the account's own release cadence.

    Builds on

    Evidence

    • Your websiteRun a pilot: 14 days for Growth and Enterprise, with the clock starting at the first suggestion; a 90-minute onboarding workshop after Promptless runs on historical events such as the last 30 days of pull requests; weekly 15-30 minute check-ins. Benchmarks: 20-40 suggestions a month for one writer, 30-100 for teams of 2-5 writers, under 20 a month suggests the team may be too early; quality target at least 8/10.Source
    • Your websiteReporting and ROI doc: Promptless for Docs has no built-in ROI dashboard or metrics export; customers assemble doc updates shipped, share of doc PRs, trigger coverage and review throughput from the suggestions list, triggers page and git history, and time saved needs their own estimate. By contrast the October 2026 changelog added an Impact page for Agent Instructions showing agent time and tokens recovered per session 30 days after each merged fix.Source
  • 15

    The Agent Instructions launch post says customers adapted the docs product to update their agent instructions, which led to the new product. The docs customers already trust Promptless with their repositories and Slack, and their engineering teams run coding agents. Yet Agent Instructions is sold only by demo, with every tier on "Contact us".

    High impactMedium effort First result in 1 quarter Co founder & CEO

    First action

    List docs customers whose engineers use coding agents.

    How to measure it

    Cross-sell rate

    Now
    –
    Target
    20% of docs customers
    By
    2 quarters

    Steps

    1. 1

      Shortlist

      Docs customers with active coding-agent use.

      Co-founder & CEO 2 days

    2. 2

      Get the intro

      Docs champion introduces the engineering lead.

      Forward-deployed technical writer 2 weeks

    3. 3

      Run the diagnostic

      One team, local, on their infrastructure.

      Forward-deployed AI engineer 1 week per account

    4. 4

      Present

      Top five wasted patterns with evidence and fixes.

      Forward-deployed AI engineer 1 hour

    5. 5

      Propose

      Startup tier for that team with a 30-day review.

      Co-founder & CEO 1 day

    6. 6

      Measure

      Cross-sell rate and revenue per customer.

      Co-founder & CEO monthly

    First result
    1 quarter
    Cost
    Forward-deployed engineering time
    Owner
    Co founder & CEO

    Smallest test

    Run the diagnostic for one docs customer this month.

    Why it works

    Offer each docs customer the local instruction-debt diagnostic for free: run it on one team's agent sessions, present the top five wasted patterns with evidence, and propose a paid Startup tier for that team. Use the docs champion to introduce the platform or engineering lead.

    Selling a second product to a happy customer is far cheaper than finding a new one, and the diagnostic shows value before any price is discussed.

    Risks and how to handle them

    • Distracts from docs delivery

      Limit to five accounts per month.

    Evidence

    • Your websiteLaunch post for Agent Instructions: Promptless began with docs, customers adapted it to update agent instructions, and traces showed missing, stale, contradictory or unclear instructions; the product has an Instruction Hub, distribution, a Trace Analyzer and a Governance Platform proposing pull requests; the CTA is to email the founders or request a local diagnostic of "instruction debt".Source
    • Your websitePricing page: docs Startup "$500/mo" for up to 200 pages with Slack and GitHub, unlimited contributors and a "Sign up" button; Growth priced by page bundle from $500-$1,000/mo (200-500 pages) to $2,000-$4,000/mo (2,000-5,000 pages), adding Linear, Jira, Confluence and Notion, translations and Promptless Capture screenshots, behind "Book demo"; Enterprise custom. All three Agent Instructions tiers (under 50, 50-200 and over 200 instructions) say "Contact us". "All plans include a 14-day free trial." The page has no FAQ and does not mention Starport; it links an OSS program for non-commercial projects.Source
  • 16

    Promptless ships every month (25 monthly changelog entries since October 2024) and writes a lot (39 technical posts), but has published only one newsletter post (September 2026), no blog post carries a date, and the homepage has no newsletter sign-up. Readers who come for a post on documentation debt or agent context files have no way to hear from Promptless again.

    Medium impactLow effort First result in 1 month Forward deployed technical wr…

    First action

    Add a sign-up form to the five most-read blog posts.

    How to measure it

    Subscribers

    Now
    –
    Target
    1,000
    By
    2 quarters

    Steps

    1. 1

      Pick a tool

      Email tool connected to the analytics.

      Co-founder & CEO 1 day

    2. 2

      Add forms

      Posts, changelog and Agent Score reports.

      Founding Engineer 1 day

    3. 3

      Add dates

      Publish and updated dates on every post.

      Founding Engineer half a day

    4. 4

      Write the template

      Changelog highlights, one post, one number.

      Forward-deployed technical writer 1 day

    5. 5

      Send monthly

      Alongside the changelog entry.

      Forward-deployed technical writer monthly

    6. 6

      Measure

      Subscribers, opens and trials from email.

      Co-founder & CEO monthly

    First result
    1 month
    Cost
    Email tool subscription
    Owner
    Forward deployed technical writer

    Smallest test

    Send the October changelog to existing trial users.

    Why it works

    Add an email sign-up to every blog post, the changelog and Agent Score reports. Each month, send one email: the changelog highlights, one technical post and one customer number. Add publish dates to posts so readers and AI assistants can tell what is current.

    Technical writers evaluate tools slowly and talk to each other; a monthly email keeps Promptless in mind until the moment a docs backlog becomes urgent.

    Risks and how to handle them

    • Low open rates

      Keep it short and practical, written by a writer for writers.

    Evidence

    • Your websiteBlog index (llms.txt): 39 technical posts (for example documentation debt, developer documentation ROI, agent context files, help center vs docs-as-code, support agent deflection and the DoGBench benchmark for doc agents), 15 product updates, one newsletter (September 2026) and one customer story; no posts carry dates.Source
    • Your websiteSite metadata: page title "Continuously improve your AI workforce and docs | Promptless Documentation"; no newsletter signup, LinkedIn or X links detected on the homepage; GitHub organisation with 73 public repositories (dogbench has 10 stars, promptless-cli is public); analytics are Google Analytics and PostHog.Source
    • Your websiteSitemap (234 URLs, checked November 9, 2026): 58 blog posts, of which one is a customer story (Vellum); 25 monthly changelog entries from October 2024 to October 2026; 135 docs pages; one use-case page (/use-cases/technical-writer-power-tool/); free tools Agent Score and a Broken Link Report; event pages /lavacon/, /bingo/ and /bingo-leaderboard/; no comparison or alternatives pages.Source
  • 18

    The homepage now sells two products to two buyers, technical writers (Docs) and platform or AI engineering leads (Agent Instructions), through tabs on one page. The page title ends "| Promptless Documentation", which makes the company look like a docs site rather than a product. There is only one use-case page, for technical writers.

    Medium impactMedium effort First result in 1 month Co founder & CEO

    First action

    Change the homepage title tag.

    How to measure it

    Landing-page conversion

    Now
    –
    Target
    +20%
    By
    1 quarter

    Steps

    1. 1

      Fix the title

      Brand and both products in the title tag.

      Founding Engineer 1 hour

    2. 2

      Write two pages

      One per product, with its own proof.

      Co-founder & CEO 1 week

    3. 3

      Build

      Pages with markdown versions and llms.txt entries.

      Founding Engineer 1 week

    4. 4

      Route the homepage

      Two clear paths above the fold.

      Founding Engineer 2 days

    5. 5

      Add use cases

      Docs managers, developer relations, support leaders.

      Forward-deployed technical writer 2 weeks

    6. 6

      Measure

      Conversion by page and product.

      Co-founder & CEO monthly

    First result
    1 month
    Cost
    Staff time only
    Owner
    Co founder & CEO

    Smallest test

    Change the title tag and watch search click-through for a month.

    Why it works

    Create /docs-agent and /agent-instructions landing pages, each with its own headline, proof, pricing and demo route, and make the homepage a short router between them. Change the title to "Promptless: AI agents that keep your docs and agent instructions current". Add use-case pages for docs managers, developer relations and support leaders.

    Each buyer sees their own problem in the first line and their own proof, and search results show the company and product rather than a docs site.

    Risks and how to handle them

    • Splitting dilutes the homepage's search ranking

      Keep the homepage as the hub and link both pages prominently.

    Evidence

    • Your websiteSite metadata: page title "Continuously improve your AI workforce and docs | Promptless Documentation"; no newsletter signup, LinkedIn or X links detected on the homepage; GitHub organisation with 73 public repositories (dogbench has 10 stars, promptless-cli is public); analytics are Google Analytics and PostHog.Source
    • Your websiteHomepage: headline "Continuously improve your AI workforce and docs" with two products, Promptless for Docs ("Write the docs. Skip the busywork.") and Promptless for Agent Instructions; 14 customer logos including Megaport, Mezmo, Vitess, Helm, Aptible, Runpod, Flatfile, Latitude, Mautic and Vellum; testimonials from a Senior Technical Writer at Runpod ("My team literally calls me a 10x tech writer"), the CTO of Latitude.sh ("basically pays for itself right away") and the Head of Developer Docs at Prove ("a solo tech writer's godsend"); buttons that open AI assistants to ask about Promptless; "Every plan includes a 14-day free trial"; Agent Instructions claims typical 30-day results of token spend down 18%, first-attempt completion up 15%, wall-clock time down 32% and human interruptions down 42%, with "white-glove onboarding with a forward-deployed AI engineer and technical writer".Source
    • Your websiteSitemap (234 URLs, checked November 9, 2026): 58 blog posts, of which one is a customer story (Vellum); 25 monthly changelog entries from October 2024 to October 2026; 135 docs pages; one use-case page (/use-cases/technical-writer-power-tool/); free tools Agent Score and a Broken Link Report; event pages /lavacon/, /bingo/ and /bingo-leaderboard/; no comparison or alternatives pages.Source
  • 19

    The open-source program is free for non-commercial projects, but it is applied for by email with no listed criteria, and the proof sits in raw GitHub PR searches linked from llms.txt (Vitess, Helm, Runpod, Mautic, Bazel). Launch HN commenters asked for exactly this kind of free use for maintainers. Open-source docs are public, so every merged Promptless PR is a demo anyone can inspect.

    Medium impactMedium effort First result in 1 quarter Co founder & CEO

    First action

    Write the eligibility criteria in five bullet points.

    How to measure it

    OSS projects with merged Promptless PRs

    Now
    –
    Target
    25
    By
    2 quarters

    Steps

    1. 1

      Set criteria

      Non-commercial, active, public docs repo.

      Co-founder & CEO 1 day

    2. 2

      Build the page

      Criteria, form and showcase.

      Founding Engineer 1 week

    3. 3

      Fill the showcase

      Merged PR counts and examples per project.

      Forward-deployed technical writer 3 days

    4. 4

      Pitch foundations

      Offer to member projects through their programmes.

      Co-founder & CEO 1 month

    5. 5

      Onboard projects

      GitHub App and style setup.

      Forward-deployed technical writer ongoing

    6. 6

      Track

      Projects, PRs merged and commercial leads from them.

      Co-founder & CEO monthly

    First result
    1 quarter
    Cost
    Compute for free projects and staff time
    Owner
    Co founder & CEO

    Smallest test

    Publish the showcase with the five projects already listed.

    Why it works

    Publish /open-source with clear criteria, a short application form, and a live showcase: projects, merged PR counts and three example PRs each. Pitch foundation programmes (CNCF and the Linux Foundation are already named) to offer it to member projects, and ask each project for a line in its contributing guide.

    Maintainers and contributors are often also engineers at commercial companies; seeing Promptless PRs in a project they know is the most credible introduction possible.

    Risks and how to handle them

    • Free projects absorb support time

      Use a community channel and a self-serve setup guide.

    Builds on

    Evidence

    • Your websiteOpen-source program: free for eligible non-commercial projects (CNCF and Linux Foundation given as examples), applied for by email, with no listed criteria; llms.txt links public lists of Promptless pull requests on Vitess, Helm, Runpod, Mautic and Bazel docs repositories.Source
    • PressLaunch HN (February 18, 2025, 107 points, 61 comments): the most pressed theme was pricing without a sales call ("let me sign up and give it a spin"); commenters also asked about free use for open-source maintainers, accuracy checks and keeping private discussions out of public docs, and one argued AI IDEs could add the feature easily.Source

Big bets 3

Larger plays that can change the trajectory.

  • 20

    Starport is already a full docs site (llms.txt, a read-only MCP server, agent-readiness scoring, broken-link checks, Doc Detective tests, hosting) with migration guides, and the blog covers "Help Center vs Docs-as-Code: When to Switch". But Starport has no price, no landing page, and its overview page is unreadable to AI agents. As docs platforms add their own update agents, owning the site as well as the agent protects Promptless from being a feature of someone else's platform.

    High impactHigh effort First result in 1 quarter Co founder & CTO

    First action

    Find two customers or prospects planning a docs migration this year.

    How to measure it

    Paying Starport sites

    Now
    –
    Target
    10
    By
    3 quarters

    Steps

    1. 1

      Find design partners

      Two teams planning a migration.

      Co-founder & CEO 1 month

    2. 2

      Fix the overview

      Publish the Starport overview in HTML and markdown.

      Forward-deployed technical writer 2 days

    3. 3

      Price it

      Fixed migration fee plus hosting and agent.

      Co-founder & CEO 1 week

    4. 4

      Run two migrations

      URLs and redirects preserved, scores before and after.

      Forward-deployed AI engineer 2 months

    5. 5

      Build the landing page

      Offer, results and migration checklist.

      Founding Engineer 2 weeks

    6. 6

      Productise

      Importers for the most common source platforms.

      Co-founder & CTO 1 quarter

    First result
    1 quarter
    Cost
    Two quarters of a part-time engineering focus
    Owner
    Co founder & CTO

    Smallest test

    Migrate one design partner by hand and measure the time it takes.

    Why it works

    Package Starport as "move your docs in 30 days, with an agent keeping them current from day one": a fixed-price migration (URLs and redirects preserved), hosting, and Promptless for Docs included. Target teams on help centers and teams whose docs platform price is rising. Launch with two migrations done for design partners and publish their before-and-after Agent Scores.

    A migration is a moment when teams are willing to change tools, and bundling the site with the agent creates a higher-value, stickier contract.

    Risks and how to handle them

    • Docs platforms become hostile

      Keep publishing to every platform; Starport is an option, not a requirement.

    • Hosting support load

      Use static hosting and a clear support boundary.

    Builds on

    Evidence

    • Your websiteStarport: a docs site built on the open-source Astro Starlight framework, shipping llms.txt files, a read-only MCP server for AI clients, copy-as-Markdown page actions, agent-readiness scoring, broken-link checks, redirects, site search, PDF export, Doc Detective testing and hosting on Vercel. No price is shown, and the docs include migration guides (choose a platform, migrate, preserve URLs and redirects).Source
    • Your websiteAI-reader check (November 9, 2026, each URL fetched at least twice, with and without a trailing slash): /blog, /changelog and the Agent Score tool page /agent-score returned HTTP 404, and /docs/for-docs/starport returned "404 – Page Not Found" with "This page does not exist in its markdown form." The same reader got full content from /pricing, /pricing.md, /agent-score.md, individual blog posts, each monthly changelog and public Agent Score reports. /blog/, /changelog/ and /agent-score/ are in the sitemap, Blog and Changelog are in the site navigation, and llms.txt lists Starport and Agent Score as pages.Source
    • Competitor researchRivals' own pages: Mintlify lists "Autopilot: Always-on agent for self-updating docs" and says that on paid plans "the agent proposes documentation updates through pull requests", with a free Starter tier; GitBook offers stale-content detection that "flags it for review" and gap tracking; Warp's Documentation Agent watches merged PRs and opens doc PRs for Mintlify or GitBook; DeepDocs is a GitHub agent, freemium with paid plans up to $25 a month (AlternativeTo).Source
    • Your websiteBlog index (llms.txt): 39 technical posts (for example documentation debt, developer documentation ROI, agent context files, help center vs docs-as-code, support agent deflection and the DoGBench benchmark for doc agents), 15 product updates, one newsletter (September 2026) and one customer story; no posts carry dates.Source
  • 21

    Promptless published DoGBench ("Can agents meet expert standards for user-facing documentation?") and a post on "The task-completion gap in AI-written docs", and the dogbench repository is its most-starred on GitHub, with only 10 stars. Competitors are all claiming self-updating docs; whoever defines how doc agents are measured shapes how buyers compare them.

    High impactHigh effort First result in 2 quarters Forward deployed AI engineer

    First action

    Write the submission rules and invite three outside agents to run it.

    How to measure it

    External submissions

    Now
    –
    Target
    8 agents
    By
    3 quarters

    Steps

    1. 1

      Set the rules

      Tasks, scoring and submission process.

      Forward-deployed AI engineer 2 weeks

    2. 2

      Recruit graders

      Expert technical writers paid per review.

      Forward-deployed technical writer 1 month

    3. 3

      Run baselines

      General coding agents and Promptless.

      Forward-deployed AI engineer 1 month

    4. 4

      Publish the leaderboard

      With methods and raw results.

      Founding Engineer 2 weeks

    5. 5

      Invite submissions

      Doc-agent vendors and open-source projects.

      Co-founder & CEO ongoing

    6. 6

      Report

      Twice-yearly findings at conferences.

      Co-founder & CEO 6 months

    First result
    2 quarters
    Cost
    Grader fees and compute
    Owner
    Forward deployed AI engineer

    Smallest test

    Publish a first leaderboard with three agents.

    Why it works

    Turn DoGBench into a maintained benchmark: a public leaderboard of doc agents and general coding agents on real documentation tasks, a submission process, expert-reviewed scoring and a twice-yearly report. Invite technical writers to contribute tasks and graders, and present it at writing conferences.

    A neutral, rigorous benchmark earns links, talks and trust from the writers who make buying decisions, and gives Promptless a fair way to show where it leads.

    Risks and how to handle them

    • Seen as biased toward Promptless

      Publish methods, use outside graders and report where Promptless loses.

    Evidence

    • Your websiteBlog index (llms.txt): 39 technical posts (for example documentation debt, developer documentation ROI, agent context files, help center vs docs-as-code, support agent deflection and the DoGBench benchmark for doc agents), 15 product updates, one newsletter (September 2026) and one customer story; no posts carry dates.Source
    • Your websiteSite metadata: page title "Continuously improve your AI workforce and docs | Promptless Documentation"; no newsletter signup, LinkedIn or X links detected on the homepage; GitHub organisation with 73 public repositories (dogbench has 10 stars, promptless-cli is public); analytics are Google Analytics and PostHog.Source
    • Competitor researchRivals' own pages: Mintlify lists "Autopilot: Always-on agent for self-updating docs" and says that on paid plans "the agent proposes documentation updates through pull requests", with a free Starter tier; GitBook offers stale-content detection that "flags it for review" and gap tracking; Warp's Documentation Agent watches merged PRs and opens doc PRs for Mintlify or GitBook; DeepDocs is a GitHub agent, freemium with paid plans up to $25 a month (AlternativeTo).Source
  • 22

    Agent Instructions starts with a demo, then a local diagnostic that finds the highest-impact gaps, and every tier is "Contact us". The homepage already claims typical 30-day results (interruptions down 42%, wall-clock time down 32%), traces never leave the customer's systems, and promptless-cli is already public on GitHub. A free local diagnostic is the Agent Score of this product: it would let any engineer see their own instruction debt without a call.

    High impactHigh effort First result in 1 quarter Co founder & CTO

    First action

    Strip the current diagnostic down to what can run on one laptop.

    How to measure it

    Teams starting a trial from the CLI

    Now
    –
    Target
    30
    By
    3 quarters

    Steps

    1. 1

      Scope the local version

      Which checks run on one developer's sessions.

      Co-founder & CTO 2 weeks

    2. 2

      Build the report

      Top wasted patterns with evidence and time lost.

      Forward-deployed AI engineer 1 month

    3. 3

      Package

      Ship in promptless-cli with clear privacy notes.

      Founding Engineer 2 weeks

    4. 4

      Add the upgrade

      Share to an Instruction Hub, which starts a trial.

      Founding Engineer 2 weeks

    5. 5

      Launch

      Post with real anonymised reports from dogfooding.

      Co-founder & CEO 1 week

    6. 6

      Measure

      Runs, shares and trials.

      Co-founder & CEO monthly

    First result
    1 quarter
    Cost
    About two engineer-months
    Owner
    Co founder & CTO

    Smallest test

    Give the local script to five developers and read their reports with them.

    Why it works

    Release the diagnostic as a free, open-source command that runs locally over a developer's own agent session logs and prints an "instruction debt report": repeated corrections, stale instructions, conflicting skills and estimated time lost. Offer a one-click upgrade to share the report with the team's Instruction Hub, which starts a trial.

    Engineers trust what they can run on their own machine, and a local tool answers the security question before it is asked. Each report shared inside a company is a warm lead for a team-wide deal.

    Risks and how to handle them

    • A free tool replaces the paid diagnostic

      Keep fleet-wide analysis, governance and evals in the paid product.

    Evidence

    • Your websiteLaunch post for Agent Instructions: Promptless began with docs, customers adapted it to update agent instructions, and traces showed missing, stale, contradictory or unclear instructions; the product has an Instruction Hub, distribution, a Trace Analyzer and a Governance Platform proposing pull requests; the CTA is to email the founders or request a local diagnostic of "instruction debt".Source
    • Your websiteHomepage: headline "Continuously improve your AI workforce and docs" with two products, Promptless for Docs ("Write the docs. Skip the busywork.") and Promptless for Agent Instructions; 14 customer logos including Megaport, Mezmo, Vitess, Helm, Aptible, Runpod, Flatfile, Latitude, Mautic and Vellum; testimonials from a Senior Technical Writer at Runpod ("My team literally calls me a 10x tech writer"), the CTO of Latitude.sh ("basically pays for itself right away") and the Head of Developer Docs at Prove ("a solo tech writer's godsend"); buttons that open AI assistants to ask about Promptless; "Every plan includes a 14-day free trial"; Agent Instructions claims typical 30-day results of token spend down 18%, first-attempt completion up 15%, wall-clock time down 32% and human interruptions down 42%, with "white-glove onboarding with a forward-deployed AI engineer and technical writer".Source
    • Your websiteSite metadata: page title "Continuously improve your AI workforce and docs | Promptless Documentation"; no newsletter signup, LinkedIn or X links detected on the homepage; GitHub organisation with 73 public repositories (dogbench has 10 stars, promptless-cli is public); analytics are Google Analytics and PostHog.Source
    • Your websitePricing page: docs Startup "$500/mo" for up to 200 pages with Slack and GitHub, unlimited contributors and a "Sign up" button; Growth priced by page bundle from $500-$1,000/mo (200-500 pages) to $2,000-$4,000/mo (2,000-5,000 pages), adding Linear, Jira, Confluence and Notion, translations and Promptless Capture screenshots, behind "Book demo"; Enterprise custom. All three Agent Instructions tiers (under 50, 50-200 and over 200 instructions) say "Contact us". "All plans include a 14-day free trial." The page has no FAQ and does not mention Starport; it links an OSS program for non-commercial projects.Source

Wildcards 2

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

  • 23

    Promptless already drafts updates from support tickets (Zendesk and Intercom are listed integrations), Vellum used it to turn troubleshooting tips from support channels into doc examples, and the blog has a post on raising support-agent deflection. But every page speaks to writers and engineers; support leaders, who own a budget and a deflection target, are not addressed anywhere.

    Medium impactMedium effort First result in 1 quarter Co founder & CEO

    First action

    Ask two current customers whose support team reads the docs most.

    How to measure it

    Pilots sponsored by support leaders

    Now
    –
    Target
    3
    By
    2 quarters

    Steps

    1. 1

      Find the sponsor

      One customer's support lead willing to pilot.

      Co-founder & CEO 2 weeks

    2. 2

      Set the trigger

      Resolved tickets with no matching doc section.

      Forward-deployed AI engineer 2 weeks

    3. 3

      Agree the measure

      Tickets per topic before and after.

      Forward-deployed technical writer 1 week

    4. 4

      Run the pilot

      60 days with weekly reviews.

      Forward-deployed technical writer 2 months

    5. 5

      Write the page

      Landing page for support leaders with pilot results.

      Co-founder & CEO 1 week

    6. 6

      Decide

      Scale or stop based on the numbers.

      Co-founder & CEO 1 day

    First result
    1 quarter
    Cost
    Forward-deployed time for one pilot
    Owner
    Co founder & CEO

    Smallest test

    Run the trigger on one month of a customer's resolved tickets and count the gaps found.

    Why it works

    Build a support-led offer: when a ticket is resolved with an answer the docs don't contain, Promptless drafts the missing section for review. Create a landing page for support leaders, measure tickets on topics before and after, and run a pilot with one existing customer's support team.

    Support leaders feel the cost of missing docs every day and can attach a number to it, which opens a second buyer and budget inside the same companies.

    Risks and how to handle them

    • Private ticket details reach public docs

      Keep human review and the existing private-data safeguards in the loop.

    Evidence

    • PressY Combinator profile: Winter 2025 batch, founded 2024, team size 5, San Francisco; integrations listed include Zendesk and Intercom; named customers include Vellum, Clay and an unnamed Fortune 500 company.Source
    • Your websiteCustomer story: at Vellum, 54% of docs PRs in the last month came from Promptless suggestions, 60% of the product and engineering team uses Promptless to draft doc updates each month, after about six months of use; Vellum plans to have Promptless generate personalised product update summaries for key clients. It is the only customer story on the site.Source
    • Your websiteBlog index (llms.txt): 39 technical posts (for example documentation debt, developer documentation ROI, agent context files, help center vs docs-as-code, support agent deflection and the DoGBench benchmark for doc agents), 15 product updates, one newsletter (September 2026) and one customer story; no posts carry dates.Source
  • 24

    Vellum's story ends with its next step: having Promptless generate personalised product update summaries for key clients. Promptless already knows every product change, with citations, and which doc pages changed. Customer success teams at developer-tool companies write these summaries by hand for their biggest accounts.

    Medium impactMedium effort First result in 1 quarter Co founder & CTO

    First action

    Ask Vellum for three example summaries they would want.

    How to measure it

    Add-on adoption

    Now
    –
    Target
    5 customers
    By
    3 quarters

    Steps

    1. 1

      Gather examples

      Three summaries Vellum would send.

      Forward-deployed technical writer 1 week

    2. 2

      Design inputs

      Which account data maps to which features.

      Co-founder & CTO 1 week

    3. 3

      Build the draft

      Monthly per-account summary with citations.

      Founding Engineer 1 month

    4. 4

      Pilot

      Vellum reviews and sends for two months.

      Forward-deployed technical writer 2 months

    5. 5

      Price

      Add-on for Growth and Enterprise.

      Co-founder & CEO 1 week

    6. 6

      Publish the story

      Update the Vellum story with results.

      Forward-deployed technical writer 1 week

    First result
    1 quarter
    Cost
    About one engineer-month
    Owner
    Co founder & CTO

    Smallest test

    Draft one summary by hand for Vellum's largest client.

    Why it works

    Add a feature that, each month, drafts a per-account summary of changes relevant to that account (based on the features or API areas it uses), with links to the updated docs, for the customer success manager to review and send. Pilot it with Vellum, then price it as an add-on to Growth and Enterprise.

    It brings a new user (customer success) into each account, raises contract value and makes Promptless harder to remove, using data it already holds.

    Risks and how to handle them

    • Wrong account data in a client email

      Always a draft for a person to approve.

    Evidence

    • Your websiteCustomer story: at Vellum, 54% of docs PRs in the last month came from Promptless suggestions, 60% of the product and engineering team uses Promptless to draft doc updates each month, after about six months of use; Vellum plans to have Promptless generate personalised product update summaries for key clients. It is the only customer story on the site.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.

ProgrammableInfo

Turn the research into a list you can act on

These ideas start with finding the right companies. Info finds them and the right person at each, researches every one and drafts the outreach, then asks before it sends.

  • Build this list in Info for idea #14 (opens programmable.info in a new tab)

New accounts: 100,000 free credits with code BETA100K

Find the companies (opens programmable.info in a new tab)
ProgrammableTalk

Fill the events in this plan

These ideas are events. Talk gets registrants for them and tells you who's coming and what their company does before the day.

Plan an event (opens programmable.talk 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.

  • MochaCare growth strategyAI tools plus hands-on services that take admin off home care agencies: caregiver hiring and onboarding, one inbox for calls, texts…21 growth ideas
  • ROLLER growth strategyAll-in-one venue management software for attractions: online ticketing, point of sale, self-serve kiosks, memberships, digital waivers…21 growth ideas
  • Overflow growth strategyA giving platform for churches and nonprofits: card, ACH, Apple and Google Pay, stock, crypto, DAF and RoundUps gifts in one dashboard…20 growth ideas
  • Zatanna growth strategyTurns software workflows that only work through a human interface (EHR and dental practice systems, ERPs, carrier, insurance and other…23 growth ideas
  • Blue growth strategyA phone with an AI chief of staff built in: BlueOS on a Google Pixel reads notifications, texts and email, then handles them with an FYI…22 growth ideas
  • Altruistiq growth strategyAn enterprise sustainability platform for food and beverage companies: corporate carbon footprints, product carbon footprints at scale…22 growth ideas
  • Redis growth strategyThe real-time data platform: Redis open source, Redis Cloud and Redis Software for caching, streaming, search and vectors, and Redis Iris…22 growth ideas
  • Serval growth strategyAn AI-native IT and enterprise service management platform: Catalyst builds automations from plain language, a help desk agent resolves…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. Promptless did not commission, review or endorse it, and Programmable is not affiliated with Promptless. 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 Promptless runs its business.

Something wrong, or do you represent Promptless? 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 every page readable by AI agents, then fix the demo page, put the Mautic results on site, price Starport and Agent Instructions, connect Agent Score reports to a trial, launch the badge, and speed up the mobile homepage.

  2. Days 31 to 6014 ideas

    Comparison pages, a docs Impact report, backfilled trials, agent sign-up over MCP, the Agent Score index, events, champion referrals, account health, cross-sell, the newsletter, split product pages, the OSS showcase, support-led docs and client update summaries.

  3. Days 61 to 903 ideas

    Commit to the big bets: Starport as a migration offer, DoGBench as the public standard for doc agents, and a self-serve instruction-debt diagnostic.

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 faster first drafts. The most cited problem: docs lag behind releases.

  • Help for solo writers ("a solo tech writer's godsend")
  • Context lost between engineers and the docs owner
  • In their words: “docs drift”

Market

Competitive landscape

5 competitors compared: Mintlify, GitBook, Warp Documentation Agent, DeepDocs and 1 more.

  • Gap to exploit: Platform-neutral doc agents that publish to any docs tool
  • Mintlify: Update agent tied to its own platform and paid plans
  • GitBook: Flags content for review rather than drafting the fix

Hiring signals

Promptless is a team of five and is hiring one role: a Founding Engineer in San Francisco ($140k-$200k plus 0.40%-2.00% equity) to build the agents, integrations and infrastructure and to talk with customers.

  • Founding Engineer (All Levels): Only open role; engineering builds the growth surface too. Ideas 1, 9, 10.

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