ProgrammableMarketing/braintrust.dev (opens in a new tab)
PDF

Views

  • Plan
  • Roadmap
  • Research

On this page

  • The brief
  • Start here
  • Idea portfolio
  • Effort vs impact
  • This week · 6
  • This quarter · 9
  • Big bets · 3
  • Wildcards · 3
  • How we can help
Generated Oct 23, 2026
  1. Programmable
  2. Playbooks
  3. Braintrust
Braintrust logo

growth playbook · Oct 23, 2026

braintrust.dev (opens in a new tab)

Braintrust growth playbook

Diagnosis

Braintrust has the developer reach most rivals want (about a million weekly npm downloads, a large research and encyclopedia library) and an $80 million Series B to scale go-to-market and physical presence.

Its hiring shows the plan: a first Head of Field Marketing, AEs and SDRs across five US cities and London, a founding data engineer for a PLG-plus-sales funnel, and engineers for SDKs and billing.

The gap is between that reach and the sales map: events are online-only at US hours, sign-ups carry no city or role, and self-serve usage is not yet scored for sales. This playbook starts with a city-aware events hub, assessment routing, EMEA sessions and the startup program's expiry date, then builds usage scoring, industry pages, partner evenings and a yearly State of AI evals report.

Growth Score

8/10

Biggest Opportunity

Connect the huge developer audience to the new field and sales teams: city-tagged events, usage-scored accounts and research-led outreach.

Biggest Risk

LangSmith, Arize, Galileo, Langfuse and Datadog all now claim agent observability and evals, so messaging converges; Braintrust has to win on depth with engineers and proof in each region.

01

Start here

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

  1. 1

    Turn the online workshop page into an events hub for the cities sales is staffing

    Add city and role fields to the workshop notify form today.

    High impact First result in 1 month
  2. 2

    Route eval maturity assessment results to the right next step

    Write the three bands and the opening line for each.

    High impact First result in 1 month
  3. 3

    Give EMEA prospects workshops in their own hours, including one in French

    Schedule one 10:00 London-time workshop next month.

    Medium impact First result in 1 month

02

Your idea portfolio

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

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

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

Effort vs impactEach number is an idea. Select one to open it.
Impact →
Quick wins · 2Big bets · 8Fill-ins · 6Reconsider · 5
Effort →

Showing 21 of 21 ideas

This week 6

Quick wins with a clear first action.

  • 01

    Braintrust is hiring a Head of Field Marketing to decide "where we show up" across cities, and AEs and SDRs in San Francisco, New York, Chicago, Austin, Denver and London. But the only events surface on braintrust.dev is the workshops page, labelled "Online events": every session runs at 10:00 AM PT (6 PM in London, 1 AM in Singapore), the sign-up asks only for an email, and /events returns 404. The site cannot tell the new hire which engineers live in which city, or invite them to anything in person.

    High impactMedium effort First result in 1 month Head of Field Marketing

    First action

    Add city and role fields to the workshop notify form today.

    How to measure it

    Pipeline sourced or influenced by events

    Now
    –
    Target
    20 opportunities
    By
    1 quarter

    Steps

    1. 1

      Add the fields

      City, role and company on the notify form.

      Web team 1 day

    2. 2

      Sync to the CRM

      Sign-ups land on the right account and owner.

      Founding Data Engineer 3 days

    3. 3

      Build /events

      Online and in-person sessions, filterable by city.

      Web team 1 week

    4. 4

      Pick two cities

      Where sign-ups and open pipeline overlap.

      Head of Field Marketing 2 days

    5. 5

      Run the first room

      A hands-on eval workshop for 30 engineers.

      Head of Field Marketing 3 weeks

    6. 6

      Add an EMEA session

      Repeat Eval foundations at a London-friendly time.

      Developer Advocate 2 weeks

    7. 7

      Follow up with Sales

      Each attendee gets a note from their AE within two days.

      Sales Development Representative ongoing

    First result
    1 month
    Cost
    Venue and catering for two small workshops
    Owner
    Head of Field Marketing

    Smallest test

    Add the city field and count sign-ups by city for two weeks.

    Why it works

    Publish /events with three filters (city, format, audience) and list the online workshops beside the first in-person sessions. Change the "Get notified" form to ask for city, role (engineer or PM) and company, and push answers to the CRM so each AE sees the local engineers who want to come. Run the first in-person "Eval foundations" in the two regions with the most sign-ups, and repeat one online workshop at an EMEA-friendly hour.

    The new team starts with a real list of people asking for events in each territory, instead of guessing, and every guest is already tied to an account and an AE. The Series B was raised partly to expand physical presence; this is the front door for it.

    Ready to use

    ChecklistEvents hub launch list

    [ ] /events page listing online and in-person sessions
    [ ] Notify form: email, city, role, company
    [ ] CRM sync to account and owner
    [ ] City filter and calendar links
    [ ] One EMEA-time workshop
    [ ] First in-person workshop in the top city
    [ ] AE follow-up within 48 hours
    [ ] Pipeline report by event

    Risks and how to handle them

    • Too few sign-ups per city at first

      Promote the city filter in workshop emails and SDK release notes.

    • In-person events drain the small team

      Start with two cities and one format.

    Evidence

    • HiringHead of Field Marketing (San Francisco, posted October 1, 2026): "Events are about people"; the role builds "the field marketing strategy. Decide where we show up, who we bring together and what we do for them", across "audiences, cities and formats", "from a small technical workshop to a party with hundreds of guests"; "Build with Sales. Know their accounts, territories and deals" and "Own the plan for invitations and follow-up"; "Bring partners into the program"; and "Track the accounts and relationships we're building, the pipeline we're creating and the deals we're helping move". Wanted: someone who has "built field marketing programs at a devtool or infrastructure company".Source
    • Your websiteWorkshops page, labelled "Online events" in the menu (checked October 22, 2026): nine sessions from June to November 2026, every one at 10:00 AM PT except one at 7:30 AM PT; one upcoming ("Eval foundations: Answering questions about your agent", November 4); "Eval foundations" has run on August 13, September 9 and October 7. The sign-up box ("Get notified when new events, live workshops, and on-demand sessions are added") asks only for an email address. There is no in-person events page: /events returns 404.Source
    • HiringSales roles open now: Enterprise Account Executive (Central Region, Denver, posted October 16, 2026; "Our primary users are software engineers"), Account Executive, Digital Natives (West Region, San Francisco, September 1), plus Strategic and Enterprise AEs listed for San Francisco, New York City, Chicago, Austin and London (EMEA), and Solutions Engineers for the West and East regions.Source
    • News coverageBraintrust raised an $80 million Series B in February 2026 led by Iconiq, with Andreessen Horowitz, Greylock and others, at an $800 million valuation; the money is to scale engineering and go-to-market teams and expand its physical presence.Source
    • HiringSales Development Representatives: "our first London-based Sales Development Representative" (posted October 8, 2026), with "French language proficiency" required, and an SDR in San Francisco (October 6). Both "Research, identify, and reach out to prospects within targeted accounts through multi-channel outreach (calls, email, LinkedIn)" against "monthly meeting targets", and find "key decision makers and champions within new large enterprise and strategic accounts".Source
  • 02

    The eval maturity assessment asks teams which practices they have, from "Vibe checks" to "Drift alerts", and ends with a generic "Talk to our team". The SDR team is measured on meetings; a team that just said it has no CI gates has told you exactly what to talk about.

    High impactLow effort First result in 1 month Sales Development Representat…

    First action

    Write the three bands and the opening line for each.

    How to measure it

    Meetings from assessments

    Now
    –
    Target
    +40%
    By
    60 days

    Steps

    1. 1

      Define bands

      Map answers to early, building and advanced.

      Sales Development Representative 1 day

    2. 2

      Store answers

      Send each answer to the CRM.

      Founding Data Engineer 3 days

    3. 3

      Write sequences

      One per band, quoting the team's own gaps.

      Sales Development Representative 2 days

    4. 4

      Route advanced teams

      To the regional AE and Solutions Engineer.

      Sales Development Representative 1 day

    5. 5

      Link the workshops

      Early teams get the next live session.

      Developer Advocate 1 day

    6. 6

      Review results

      Meeting rate by band every two weeks.

      Sales Development Representative biweekly

    First result
    1 month
    Cost
    Staff time only
    Owner
    Sales Development Representative

    Smallest test

    Hand-write follow-ups to last week's assessments.

    Why it works

    Score each completed assessment into three bands. Early teams get the next Eval foundations workshop and an Academy path; teams with traces but no evals get a 20-minute "first eval on your data" call; advanced teams go straight to an AE and Solutions Engineer. SDR emails open with the two gaps the team picked.

    Specific follow-up turns a self-assessment into a meeting, and the bands stop SDRs spending time on teams that only need a tutorial.

    Risks and how to handle them

    • Small teams flood the queue

      Only route teams with a work email and three or more engineers.

    Evidence

    • Your websiteEval maturity assessment: visitors "Mark the quality practices your team already has in motion" from "No evals setup", "Manual QA only" and "Vibe checks" up to "Automatic CI gates", "Drift alerts" and "Online scores", then "Set confidence"; the page ends with "Talk to our team". The financial services workshop (August 26) offered "a reference architecture for financial services teams running Braintrust".Source
    • HiringSales Development Representatives: "our first London-based Sales Development Representative" (posted October 8, 2026), with "French language proficiency" required, and an SDR in San Francisco (October 6). Both "Research, identify, and reach out to prospects within targeted accounts through multi-channel outreach (calls, email, LinkedIn)" against "monthly meeting targets", and find "key decision makers and champions within new large enterprise and strategic accounts".Source
    • Your websiteWorkshops page, labelled "Online events" in the menu (checked October 22, 2026): nine sessions from June to November 2026, every one at 10:00 AM PT except one at 7:30 AM PT; one upcoming ("Eval foundations: Answering questions about your agent", November 4); "Eval foundations" has run on August 13, September 9 and October 7. The sign-up box ("Get notified when new events, live workshops, and on-demand sessions are added") asks only for an email address. There is no in-person events page: /events returns 404.Source
  • 03

    Braintrust's first London SDR must speak French, and EMEA AEs and London support engineers are being hired. Yet every live workshop is scheduled for US mornings, so a Paris or London engineer joins at 6 PM or watches a recording.

    Medium impactLow effort First result in 1 month Sales Development Representat…

    First action

    Schedule one 10:00 London-time workshop next month.

    How to measure it

    EMEA meetings from workshops

    Now
    –
    Target
    15 a quarter
    By
    1 quarter

    Steps

    1. 1

      Pick the slot

      Monthly, London morning.

      Developer Advocate 1 day

    2. 2

      Find a French speaker

      Customer or team member.

      Sales Development Representative (London) 2 weeks

    3. 3

      Promote

      SDR outreach, meetups and LinkedIn.

      Sales Development Representative (London) 2 weeks

    4. 4

      Run it

      Same agenda as the US session.

      Developer Advocate monthly

    5. 5

      Follow up

      SDR call within two days.

      Sales Development Representative (London) monthly

    6. 6

      Measure

      Attendees, meetings and pipeline by session.

      Founding Data Engineer monthly

    First result
    1 month
    Cost
    Staff time only
    Owner
    Sales Development Representative (London)

    Smallest test

    Run one London-time session and compare attendance.

    Why it works

    Run Eval foundations monthly at 10:00 London time, and once a quarter in French with a French-speaking customer or engineer. Promote it through the London SDR's outreach, French AI meetups and LinkedIn, and track meetings booked from attendees.

    A live session in the prospect's hours and language gives the new SDR a reason to reach out that is not a cold pitch.

    Risks and how to handle them

    • Low turnout at first

      Co-host with a local meetup.

    Evidence

    • HiringSales Development Representatives: "our first London-based Sales Development Representative" (posted October 8, 2026), with "French language proficiency" required, and an SDR in San Francisco (October 6). Both "Research, identify, and reach out to prospects within targeted accounts through multi-channel outreach (calls, email, LinkedIn)" against "monthly meeting targets", and find "key decision makers and champions within new large enterprise and strategic accounts".Source
    • Your websiteWorkshops page, labelled "Online events" in the menu (checked October 22, 2026): nine sessions from June to November 2026, every one at 10:00 AM PT except one at 7:30 AM PT; one upcoming ("Eval foundations: Answering questions about your agent", November 4); "Eval foundations" has run on August 13, September 9 and October 7. The sign-up box ("Get notified when new events, live workshops, and on-demand sessions are added") asks only for an email address. There is no in-person events page: /events returns 404.Source
    • HiringSales roles open now: Enterprise Account Executive (Central Region, Denver, posted October 16, 2026; "Our primary users are software engineers"), Account Executive, Digital Natives (West Region, San Francisco, September 1), plus Strategic and Enterprise AEs listed for San Francisco, New York City, Chicago, Austin and London (EMEA), and Solutions Engineers for the West and East regions.Source
    • HiringSoftware Engineer, SDK (remote, posted October 8, 2026): builds "the open-source libraries that customers use to observe, understand, and optimize how AI operates", "embedded in thousands of applications"; owns "design, implementation, release, documentation, and community support"; integrates with "OpenAI, Anthropic, and Gemini"; and should "Be a great community ambassador". Support engineers are also being hired in London and Singapore.Source
  • 04

    Braintrust's sitemap lists 480 URLs but leaves out live pages linked from every footer: /startups, the three /product pages, /resources/for-pms and /newsletter. These are some of the highest-intent pages on the site.

    Medium impactLow effort First result in 2 weeks Web team

    First action

    Add the six URLs to the sitemap.

    How to measure it

    Missing pages indexed

    Now
    –
    Target
    6 of 6
    By
    30 days

    Steps

    1. 1

      List gaps

      Compare footer links with the sitemap.

      Web team 1 hour

    2. 2

      Fix the generator

      Include product and resource routes.

      Web team 2 hours

    3. 3

      Check tags

      Canonical and description on each page.

      Web team 1 hour

    4. 4

      Resubmit

      Search Console.

      Web team 15 minutes

    5. 5

      Add a test

      Fail the build when a linked page is missing.

      Web team 2 hours

    6. 6

      Review

      Indexing and search clicks after two weeks.

      Web team 2 weeks

    First result
    2 weeks
    Cost
    $0
    Owner
    Web team

    Smallest test

    Add /startups today and check indexing next week.

    Why it works

    Add the missing pages to the sitemap generator, check that each has a canonical tag and description, resubmit in Search Console and check indexing a week later. Add a test so new top-level pages are not left out again.

    Search engines and AI assistants find and refresh listed pages faster, and the startup program is a page founders search for.

    Risks and how to handle them

    • None significant

      Watch for duplicate URLs with and without trailing slashes.

    Evidence

    • Your websiteSitemap (480 URLs, checked October 22, 2026): 174 articles (including comparison pages such as braintrust-vs-datadog-llm-observability and arize-phoenix-vs-braintrust), 99 encyclopedia entries, 97 blog posts, 29 eval research studies, 15 Academy lessons, 15 customer stories and 9 workshops. Live pages linked from the homepage are missing from it: /startups, /product/observe, /product/evaluate, /product/discover, /resources/for-pms and /newsletter. There are no industry pages.Source
    • Your websiteStartups page: any new customer "with at least $100k in funding is eligible for six months of Pro plan access, while those funded by our partner VC firms at Series A or earlier are eligible for twelve months"; partners include Accel, Founders Fund, Khosla Ventures, Kleiner Perkins and Y Combinator.Source
  • 05

    Startups get six or twelve months of Pro free, and an AE is being hired for digital natives. The month a free period ends is when a fast-growing AI startup either upgrades, shrinks to Starter or leaves, and it is predictable to the day.

    High impactLow effort First result in 1 month Account Executive

    First action

    List startup accounts whose free period ends in the next 90 days.

    How to measure it

    Startup conversion to paid

    Now
    –
    Target
    50%
    By
    2 quarters

    Steps

    1. 1

      List accounts

      End dates and current usage.

      Founding Data Engineer 2 days

    2. 2

      Build the summary

      Usage and value in one email.

      Software Engineer, Billing 1 week

    3. 3

      Assign AEs

      Digital natives AE owns the list.

      Account Executive, Digital Natives 1 day

    4. 4

      Send at day -60

      Summary plus review offer.

      Account Executive, Digital Natives ongoing

    5. 5

      Flag large users

      Above-Pro usage to Enterprise.

      Account Executive, Digital Natives ongoing

    6. 6

      Track

      Conversion by cohort.

      Founding Data Engineer monthly

    First result
    1 month
    Cost
    Staff time only
    Owner
    Account Executive, Digital Natives

    Smallest test

    Call the five largest startups ending this quarter.

    Why it works

    Sixty days before each free period ends, send the account a usage summary (traces, scores, experiments, regressions caught) and an AE offer: a review of their workflow and a plan sized to their usage. Flag accounts whose usage is above Pro limits for Enterprise.

    The free months create habits; a timed, personal conversion step keeps them as paying customers instead of letting a billing email decide.

    Risks and how to handle them

    • Startups feel pressured

      Lead with their own usage and a downgrade option.

    Evidence

    • Your websiteStartups page: any new customer "with at least $100k in funding is eligible for six months of Pro plan access, while those funded by our partner VC firms at Series A or earlier are eligible for twelve months"; partners include Accel, Founders Fund, Khosla Ventures, Kleiner Perkins and Y Combinator.Source
    • Your websitePricing: Starter $0 a month ($10 credits, 1 GB processed data then $4/GB, 10k scores then $2.50 per 1,000, 14-day retention); Pro $249 a month ($100 credits, 5 GB then $3/GB, 50k scores then $1.50 per 1,000, 30-day retention); Enterprise custom. Unlimited users on every plan. A usage calculator estimates traces and quality checks from tokens per month. "6-12 months free for qualifying startups".Source
    • HiringSales roles open now: Enterprise Account Executive (Central Region, Denver, posted October 16, 2026; "Our primary users are software engineers"), Account Executive, Digital Natives (West Region, San Francisco, September 1), plus Strategic and Enterprise AEs listed for San Francisco, New York City, Chicago, Austin and London (EMEA), and Solutions Engineers for the West and East regions.Source
    • HiringSoftware Engineer, Billing (posted October 13, 2026): builds the systems "that make Braintrust's pricing, usage, and payments accurate, scalable, and transparent", billing "processed data, evaluation scores, model usage, and data retention"; helps "launch pricing models that let customers adopt the platform with confidence as their AI workloads grow" and builds tools "helping customers understand their account state".Source
  • 06

    The pricing page has a usage calculator, but the estimate disappears when the visitor leaves. Usage-based prices are usually approved by someone else, and the billing engineer is hired to make pricing "transparent".

    Medium impactLow effort First result in 1 month Software Engineer

    First action

    Add a shareable URL that stores the calculator inputs.

    How to measure it

    Pricing page to Pro or contact

    Now
    –
    Target
    +15%
    By
    60 days

    Steps

    1. 1

      Encode inputs

      Calculator state in the URL.

      Web team 2 days

    2. 2

      Add email option

      Send the estimate and breakdown.

      Web team 2 days

    3. 3

      Recommend a plan

      Starter, Pro or Enterprise from the numbers.

      Software Engineer, Billing 2 days

    4. 4

      Sync to CRM

      Estimate on the lead record.

      Founding Data Engineer 2 days

    5. 5

      Test

      Against the current page.

      Web team 3 weeks

    6. 6

      Brief AEs

      Open calls with the estimate.

      Account Executive, Digital Natives 1 day

    First result
    1 month
    Cost
    Engineering time
    Owner
    Software Engineer, Billing

    Smallest test

    Add the shareable link only.

    Why it works

    Add "Email me this estimate" and a shareable link that keeps the inputs, the recommended plan and a breakdown of data, scores and retention. Pass the estimate to the CRM so the AE can see it before the call.

    A shareable number moves the buying conversation to the team, and the AE starts from the prospect's own volume.

    Risks and how to handle them

    • Estimates are wrong for unusual workloads

      Show the assumptions under each number.

    Evidence

    • Your websitePricing: Starter $0 a month ($10 credits, 1 GB processed data then $4/GB, 10k scores then $2.50 per 1,000, 14-day retention); Pro $249 a month ($100 credits, 5 GB then $3/GB, 50k scores then $1.50 per 1,000, 30-day retention); Enterprise custom. Unlimited users on every plan. A usage calculator estimates traces and quality checks from tokens per month. "6-12 months free for qualifying startups".Source
    • HiringSoftware Engineer, Billing (posted October 13, 2026): builds the systems "that make Braintrust's pricing, usage, and payments accurate, scalable, and transparent", billing "processed data, evaluation scores, model usage, and data retention"; helps "launch pricing models that let customers adopt the platform with confidence as their AI workloads grow" and builds tools "helping customers understand their account state".Source
    • HiringFounding Data Engineer (Marketing department): "Braintrust is growing quickly across enterprise and self-serve customers"; the role defines "the business entities, metrics, and semantics that Product, Finance, and GTM all build on", owns ingestion "from product events, CRM, and other business systems", and works "with Sales, RevOps, and Marketing on the pipeline and customer data". Bonus: "usage-based pricing, consumption models, or PLG-plus-sales-assisted funnels". Stack includes Snowflake and dbt.Source

This quarter 9

The core plays for the next three months.

  • 07

    Braintrust runs a workshop roughly every two weeks, but its YouTube channel has 1,380 subscribers and 109 videos, small next to nearly a million weekly npm downloads. The recordings are the best explainers the company has.

    Medium impactMedium effort First result in 1 quarter Developer Advocate

    First action

    Publish the last Eval foundations recording with chapters.

    How to measure it

    YouTube subscribers

    Now
    –
    Target
    5,000
    By
    2 quarters

    Steps

    1. 1

      Set a template

      Title, chapters, summary, links.

      Developer Advocate 1 week

    2. 2

      Back-fill

      Past nine workshops.

      Developer Advocate 1 month

    3. 3

      Cut clips

      Three per session.

      Developer Advocate biweekly

    4. 4

      Add end cards

      Next workshop and events hub.

      Developer Advocate 1 week

    5. 5

      Share

      Discord, X, LinkedIn.

      Developer Advocate biweekly

    6. 6

      Measure

      Views to sign-ups.

      Founding Data Engineer monthly

    First result
    1 quarter
    Cost
    Editing time or a freelance editor
    Owner
    Developer Advocate

    Smallest test

    Post one workshop with chapters and three clips.

    Why it works

    Upload each workshop with chapters, a written summary and links to the matching docs and Academy lessons; cut three short clips per session for YouTube, X and LinkedIn; add an end card for the next live workshop and the events hub.

    Video meets engineers where they search for "how to eval an agent", and every clip promotes the next session.

    Risks and how to handle them

    • Clips without context

      Link each to the full session.

    Evidence

    • Developer reachThe braintrust npm package has about 976,000 weekly downloads (27 packages, about 2.3 million a week in total, including @braintrust/otel at about 126,000); the Python package was last released on October 21, 2026 (version 0.48.0); GitHub: autoevals has 1,064 stars. By comparison the YouTube channel has 1,380 subscribers, 109 videos and about 118,000 total views, and the community lives on Discord.Source
    • Your websiteWorkshops page, labelled "Online events" in the menu (checked October 22, 2026): nine sessions from June to November 2026, every one at 10:00 AM PT except one at 7:30 AM PT; one upcoming ("Eval foundations: Answering questions about your agent", November 4); "Eval foundations" has run on August 13, September 9 and October 7. The sign-up box ("Get notified when new events, live workshops, and on-demand sessions are added") asks only for an email address. There is no in-person events page: /events returns 404.Source
    • Your websiteSitemap (480 URLs, checked October 22, 2026): 174 articles (including comparison pages such as braintrust-vs-datadog-llm-observability and arize-phoenix-vs-braintrust), 99 encyclopedia entries, 97 blog posts, 29 eval research studies, 15 Academy lessons, 15 customer stories and 9 workshops. Live pages linked from the homepage are missing from it: /startups, /product/observe, /product/evaluate, /product/discover, /resources/for-pms and /newsletter. There are no industry pages.Source
  • 08

    The braintrust npm package is downloaded about 976,000 times a week and the Python SDK ships often, yet the Discord and YouTube audiences are small. Most SDK users never see a workshop invite. The SDK engineer is hired to be a "community ambassador".

    High impactMedium effort First result in 1 quarter Software Engineer

    First action

    Add a workshop link to the next release notes.

    How to measure it

    Sign-ups from SDK and docs

    Now
    –
    Target
    1,000 a quarter
    By
    1 quarter

    Steps

    1. 1

      Write the line

      One sentence and one link.

      Software Engineer, SDK 1 day

    2. 2

      Release notes

      Every SDK release.

      Software Engineer, SDK ongoing

    3. 3

      CLI help

      Link in help output.

      Software Engineer, SDK 1 week

    4. 4

      Docs banner

      Next live session.

      Web team 2 days

    5. 5

      Office hours

      Monthly on Discord.

      Developer Support Engineer monthly

    6. 6

      Measure

      Tagged links.

      Founding Data Engineer monthly

    First result
    1 quarter
    Cost
    $0
    Owner
    Software Engineer, SDK

    Smallest test

    Tag one release note link and count clicks.

    Why it works

    Add a short "learn" line to SDK release notes and the CLI's help output pointing to the next workshop and the Academy; add a docs banner for the events hub; invite engineers who open GitHub issues to the Discord office hours. Keep it opt-in and free of tracking.

    The people already installing the SDK are the warmest audience Braintrust has, and the channel costs nothing.

    Risks and how to handle them

    • Developers dislike noise

      One line, no prompts, no telemetry.

    Evidence

    • Developer reachThe braintrust npm package has about 976,000 weekly downloads (27 packages, about 2.3 million a week in total, including @braintrust/otel at about 126,000); the Python package was last released on October 21, 2026 (version 0.48.0); GitHub: autoevals has 1,064 stars. By comparison the YouTube channel has 1,380 subscribers, 109 videos and about 118,000 total views, and the community lives on Discord.Source
    • HiringSoftware Engineer, SDK (remote, posted October 8, 2026): builds "the open-source libraries that customers use to observe, understand, and optimize how AI operates", "embedded in thousands of applications"; owns "design, implementation, release, documentation, and community support"; integrates with "OpenAI, Anthropic, and Gemini"; and should "Be a great community ambassador". Support engineers are also being hired in London and Singapore.Source
    • Your websiteWorkshops page, labelled "Online events" in the menu (checked October 22, 2026): nine sessions from June to November 2026, every one at 10:00 AM PT except one at 7:30 AM PT; one upcoming ("Eval foundations: Answering questions about your agent", November 4); "Eval foundations" has run on August 13, September 9 and October 7. The sign-up box ("Get notified when new events, live workshops, and on-demand sessions are added") asks only for an email address. There is no in-person events page: /events returns 404.Source
  • 09

    The founding data engineer is asked to define the metrics GTM builds on for a "PLG-plus-sales-assisted" funnel. SDRs need to know which free or Pro workspaces are growing into enterprise accounts.

    High impactHigh effort First result in 1 quarter Founding Data Engineer

    First action

    Pick five usage signals that preceded past Enterprise deals.

    How to measure it

    Pipeline from product-qualified accounts

    Now
    –
    Target
    30% of new pipeline
    By
    2 quarters

    Steps

    1. 1

      Study past deals

      Usage before each Enterprise win.

      Founding Data Engineer 2 weeks

    2. 2

      Model usage

      Account-level tables in dbt.

      Founding Data Engineer 1 month

    3. 3

      Join firmographics

      From the CRM.

      Founding Data Engineer 2 weeks

    4. 4

      Score weekly

      With a readable reason.

      Founding Data Engineer 1 week

    5. 5

      Push to queues

      SDR and AE views.

      Founding Data Engineer 1 week

    6. 6

      Review

      Conversion by score band.

      Founding Data Engineer monthly

    First result
    1 quarter
    Cost
    Data engineering time
    Owner
    Founding Data Engineer

    Smallest test

    Hand-score the top 50 workspaces and give them to SDRs.

    Why it works

    Model account-level usage (traces, scores, users, projects, retention overages) in Snowflake, join it to company size from the CRM, and push a weekly score and the reason ("10 new users this month, hitting the 5 GB limit") to the SDR and AE queues.

    Usage-led signals convert far better than cold lists, and they give the SDR a specific reason to call.

    Risks and how to handle them

    • Score is a black box

      Always show the top two reasons.

    Evidence

    • HiringFounding Data Engineer (Marketing department): "Braintrust is growing quickly across enterprise and self-serve customers"; the role defines "the business entities, metrics, and semantics that Product, Finance, and GTM all build on", owns ingestion "from product events, CRM, and other business systems", and works "with Sales, RevOps, and Marketing on the pipeline and customer data". Bonus: "usage-based pricing, consumption models, or PLG-plus-sales-assisted funnels". Stack includes Snowflake and dbt.Source
    • HiringSales Development Representatives: "our first London-based Sales Development Representative" (posted October 8, 2026), with "French language proficiency" required, and an SDR in San Francisco (October 6). Both "Research, identify, and reach out to prospects within targeted accounts through multi-channel outreach (calls, email, LinkedIn)" against "monthly meeting targets", and find "key decision makers and champions within new large enterprise and strategic accounts".Source
    • Your websitePricing: Starter $0 a month ($10 credits, 1 GB processed data then $4/GB, 10k scores then $2.50 per 1,000, 14-day retention); Pro $249 a month ($100 credits, 5 GB then $3/GB, 50k scores then $1.50 per 1,000, 30-day retention); Enterprise custom. Unlimited users on every plan. A usage calculator estimates traces and quality checks from tokens per month. "6-12 months free for qualifying startups".Source
  • 10

    The field marketing role is asked to "bring partners into the program". Braintrust integrates with OpenAI, Anthropic and Gemini and many frameworks, and engineers come to events where they can meet those teams.

    High impactMedium effort First result in 1 quarter Head of Field Marketing

    First action

    Shortlist three partners and two cities.

    How to measure it

    Target-account attendees

    Now
    –
    Target
    40% of guests
    By
    2 quarters

    Steps

    1. 1

      Shortlist partners

      Model providers and frameworks.

      Head of Field Marketing 2 weeks

    2. 2

      Agree the format

      Talk, demo, panel.

      Head of Field Marketing 2 weeks

    3. 3

      Build guest lists

      From both sales teams.

      Enterprise Account Executive (Central Region) 2 weeks

    4. 4

      Run the evening

      One city first.

      Head of Field Marketing 1 month

    5. 5

      Joint follow-up

      Within two days.

      Sales Development Representative 1 week

    6. 6

      Measure

      Accounts and pipeline.

      Founding Data Engineer monthly

    First result
    1 quarter
    Cost
    Shared venue costs
    Owner
    Head of Field Marketing

    Smallest test

    Run one co-hosted evening in San Francisco.

    Why it works

    Plan one evening per quarter per city with a partner: a short talk from the partner, a live eval or tracing demo from Braintrust, and customers on a panel. Split the guest list between both sales teams and follow up together.

    Partners double the draw and the invitation list at half the cost, and joint events show Braintrust as the neutral layer across models.

    Risks and how to handle them

    • Partner sells over us

      Agree topics and follow-up rules in advance.

    Evidence

    • HiringHead of Field Marketing (San Francisco, posted October 1, 2026): "Events are about people"; the role builds "the field marketing strategy. Decide where we show up, who we bring together and what we do for them", across "audiences, cities and formats", "from a small technical workshop to a party with hundreds of guests"; "Build with Sales. Know their accounts, territories and deals" and "Own the plan for invitations and follow-up"; "Bring partners into the program"; and "Track the accounts and relationships we're building, the pipeline we're creating and the deals we're helping move". Wanted: someone who has "built field marketing programs at a devtool or infrastructure company".Source
    • HiringSoftware Engineer, SDK (remote, posted October 8, 2026): builds "the open-source libraries that customers use to observe, understand, and optimize how AI operates", "embedded in thousands of applications"; owns "design, implementation, release, documentation, and community support"; integrates with "OpenAI, Anthropic, and Gemini"; and should "Be a great community ambassador". Support engineers are also being hired in London and Singapore.Source
    • News coverageBraintrust raised an $80 million Series B in February 2026 led by Iconiq, with Andreessen Horowitz, Greylock and others, at an $800 million valuation; the money is to scale engineering and go-to-market teams and expand its physical presence.Source
  • 11

    The sitemap has no industry pages, yet Braintrust ran a workshop for financial services teams with a reference architecture, and lists HIPAA, SOC 2 and hybrid deployment. Enterprise AEs in the Central and East regions meet banks, insurers and health companies.

    High impactMedium effort First result in 1 quarter Enterprise Account Executive

    First action

    Turn the financial services workshop into a page.

    How to measure it

    Industry-page demo requests

    Now
    –
    Target
    20 a quarter
    By
    2 quarters

    Steps

    1. 1

      Draft financial services

      From the workshop.

      Solutions Engineer 2 weeks

    2. 2

      Draft healthcare

      HIPAA and PHI focus.

      Solutions Engineer 3 weeks

    3. 3

      Security review

      Claims match the trust center.

      Security team 1 week

    4. 4

      Publish

      With the architecture download.

      Web team 1 week

    5. 5

      Use in outreach

      Regional AE sequences.

      Enterprise Account Executive (Central Region) ongoing

    6. 6

      Measure

      Demos by page.

      Founding Data Engineer monthly

    First result
    1 quarter
    Cost
    Staff time only
    Owner
    Enterprise Account Executive (Central Region)

    Smallest test

    Publish the financial services page first.

    Why it works

    Publish two industry pages: financial services and healthcare. Each covers the risks (PII, policy, audit), how evals and traces handle them, the hybrid deployment option, the workshop recording and a reference architecture download. Use them in AE outreach.

    Regulated buyers search by industry and need to see their own risks named before they take a meeting.

    Risks and how to handle them

    • Overstated compliance claims

      Link every claim to the trust center.

    Evidence

    • Your websiteSitemap (480 URLs, checked October 22, 2026): 174 articles (including comparison pages such as braintrust-vs-datadog-llm-observability and arize-phoenix-vs-braintrust), 99 encyclopedia entries, 97 blog posts, 29 eval research studies, 15 Academy lessons, 15 customer stories and 9 workshops. Live pages linked from the homepage are missing from it: /startups, /product/observe, /product/evaluate, /product/discover, /resources/for-pms and /newsletter. There are no industry pages.Source
    • Your websiteEval maturity assessment: visitors "Mark the quality practices your team already has in motion" from "No evals setup", "Manual QA only" and "Vibe checks" up to "Automatic CI gates", "Drift alerts" and "Online scores", then "Set confidence"; the page ends with "Talk to our team". The financial services workshop (August 26) offered "a reference architecture for financial services teams running Braintrust".Source
    • Your websiteHomepage: "Agents fail differently than normal software. You need active observability to monitor and fix them." Observability, Evals and Discovery; Loop agent, Patterns, trace-to-dataset, an MCP server and SDKs "for Python, TypeScript, Go, Ruby, C#, and more"; SOC 2 Type II, HIPAA, GDPR and hybrid deployment; Notion ("evaluates AI at scale across 70 engineers") and Graphite stories; a link to "Take the eval maturity assessment".Source
    • HiringSales roles open now: Enterprise Account Executive (Central Region, Denver, posted October 16, 2026; "Our primary users are software engineers"), Account Executive, Digital Natives (West Region, San Francisco, September 1), plus Strategic and Enterprise AEs listed for San Francisco, New York City, Chicago, Austin and London (EMEA), and Solutions Engineers for the West and East regions.Source
  • 12

    Braintrust has published 29 eval research studies (model comparisons, voice, search relevance, shopping assistants, cost efficiency). SDRs must hold "technical conversations" and reach decision makers; a study on the prospect's own use case is a stronger opener than a product pitch.

    Medium impactLow effort First result in 1 month Sales Development Representat…

    First action

    Tag the 29 studies by use case.

    How to measure it

    SDR reply rate

    Now
    –
    Target
    +30%
    By
    1 quarter

    Steps

    1. 1

      Tag studies

      Use case and industry.

      Developer Advocate 2 days

    2. 2

      Write summaries

      One paragraph each.

      Developer Advocate 1 week

    3. 3

      Map to accounts

      Use case per target account.

      Sales Development Representative 1 week

    4. 4

      Build sequences

      Study, summary, offer.

      Sales Development Representative 2 days

    5. 5

      Offer walk-throughs

      With the author.

      Developer Advocate ongoing

    6. 6

      Measure

      Reply and meeting rates.

      Sales Development Representative monthly

    First result
    1 month
    Cost
    Staff time only
    Owner
    Sales Development Representative

    Smallest test

    Send ten research-led emails and compare replies.

    Why it works

    Tag each study by use case and industry, map it to target account types, and write a one-paragraph summary SDRs can send ("we tested X; here is what changed the result"). Offer a 20-minute walk-through with the author.

    Research-led outreach reads as useful to engineers and gets replies from people who ignore sales emails.

    Risks and how to handle them

    • Studies age fast

      Use the newest two per use case.

    Evidence

    • Your websiteSitemap (480 URLs, checked October 22, 2026): 174 articles (including comparison pages such as braintrust-vs-datadog-llm-observability and arize-phoenix-vs-braintrust), 99 encyclopedia entries, 97 blog posts, 29 eval research studies, 15 Academy lessons, 15 customer stories and 9 workshops. Live pages linked from the homepage are missing from it: /startups, /product/observe, /product/evaluate, /product/discover, /resources/for-pms and /newsletter. There are no industry pages.Source
    • HiringSales Development Representatives: "our first London-based Sales Development Representative" (posted October 8, 2026), with "French language proficiency" required, and an SDR in San Francisco (October 6). Both "Research, identify, and reach out to prospects within targeted accounts through multi-channel outreach (calls, email, LinkedIn)" against "monthly meeting targets", and find "key decision makers and champions within new large enterprise and strategic accounts".Source
    • HiringSales roles open now: Enterprise Account Executive (Central Region, Denver, posted October 16, 2026; "Our primary users are software engineers"), Account Executive, Digital Natives (West Region, San Francisco, September 1), plus Strategic and Enterprise AEs listed for San Francisco, New York City, Chicago, Austin and London (EMEA), and Solutions Engineers for the West and East regions.Source
  • 13

    Braintrust is hiring support engineers in San Francisco, London and Singapore. New Pro teams that do not run their first eval on real data early tend to stay on tracing only.

    Medium impactMedium effort First result in 1 quarter Developer Support Engineer

    First action

    Define "first eval on real data" as a tracked event.

    How to measure it

    First eval within 14 days

    Now
    –
    Target
    70%
    By
    2 quarters

    Steps

    1. 1

      Define the milestone

      Event in product data.

      Founding Data Engineer 1 week

    2. 2

      Write the curriculum

      Four steps, one hour.

      Developer Support Engineer 2 weeks

    3. 3

      Schedule cohorts

      SF, London, Singapore.

      Developer Support Engineer 1 week

    4. 4

      Invite

      Every new Pro workspace.

      Developer Support Engineer weekly

    5. 5

      Nudge

      Teams without an eval at day 7.

      Developer Support Engineer weekly

    6. 6

      Measure

      Retention by cohort.

      Founding Data Engineer monthly

    First result
    1 quarter
    Cost
    Support time
    Owner
    Developer Support Engineer

    Smallest test

    Run one cohort for this week's new Pro teams.

    Why it works

    Invite every new Pro workspace to a weekly onboarding cohort in its time zone: connect traces, turn a failure into a dataset, write a scorer, run an experiment. Track which teams reach their first eval within 14 days.

    A guided first eval builds the habit that keeps teams paying, and three time zones cover every customer.

    Risks and how to handle them

    • Low attendance

      Offer a recorded version and office hours.

    Evidence

    • HiringSoftware Engineer, SDK (remote, posted October 8, 2026): builds "the open-source libraries that customers use to observe, understand, and optimize how AI operates", "embedded in thousands of applications"; owns "design, implementation, release, documentation, and community support"; integrates with "OpenAI, Anthropic, and Gemini"; and should "Be a great community ambassador". Support engineers are also being hired in London and Singapore.Source
    • Your websitePricing: Starter $0 a month ($10 credits, 1 GB processed data then $4/GB, 10k scores then $2.50 per 1,000, 14-day retention); Pro $249 a month ($100 credits, 5 GB then $3/GB, 50k scores then $1.50 per 1,000, 30-day retention); Enterprise custom. Unlimited users on every plan. A usage calculator estimates traces and quality checks from tokens per month. "6-12 months free for qualifying startups".Source
    • Your websiteHomepage: "Agents fail differently than normal software. You need active observability to monitor and fix them." Observability, Evals and Discovery; Loop agent, Patterns, trace-to-dataset, an MCP server and SDKs "for Python, TypeScript, Go, Ruby, C#, and more"; SOC 2 Type II, HIPAA, GDPR and hybrid deployment; Notion ("evaluates AI at scale across 70 engineers") and Graphite stories; a link to "Take the eval maturity assessment".Source
  • 14

    Real Chrome visitors to braintrust.dev see the main content in 3.1s at the 75th percentile, above Google's 2.5s "good" limit, while responsiveness and layout are good. For a company selling speed of investigation, the site should feel instant.

    Medium impactMedium effort First result in 1 quarter Web team

    First action

    Run PageSpeed on the five most visited pages.

    How to measure it

    LCP at the 75th percentile

    Now
    –
    Target
    Under 2.5s
    By
    2 quarters

    Steps

    1. 1

      Measure pages

      Top five by traffic.

      Web team 1 day

    2. 2

      Find the element

      Largest paint on each.

      Web team 2 days

    3. 3

      Fix

      Preload, compress, defer embeds.

      Web team 2 weeks

    4. 4

      Test

      Lab and field.

      Web team 2 weeks

    5. 5

      Guard

      Performance budget in CI.

      Web team 1 week

    6. 6

      Review

      Chrome UX Report monthly.

      Web team monthly

    First result
    1 quarter
    Cost
    Engineering time
    Owner
    Web team

    Smallest test

    Fix the homepage first and compare field data in a month.

    Why it works

    Find the largest element on the homepage, pricing and customer pages, preload or simplify it, defer the animated benchmark and video embeds, and check the Chrome UX Report monthly.

    Faster pages lift sign-ups and help search rankings, and the fix is concentrated in a few page elements.

    Risks and how to handle them

    • Design changes

      Keep the benchmark, load it later.

    Evidence

    • Google PageSpeedChrome UX Report (real visitors to braintrust.dev): Largest Contentful Paint at the 75th percentile is 3.1s, above Google's 2.5s "good" limit; responsiveness (INP 143 ms) and layout stability (CLS 0.02) are good.Source
    • Your websiteHomepage: "Agents fail differently than normal software. You need active observability to monitor and fix them." Observability, Evals and Discovery; Loop agent, Patterns, trace-to-dataset, an MCP server and SDKs "for Python, TypeScript, Go, Ruby, C#, and more"; SOC 2 Type II, HIPAA, GDPR and hybrid deployment; Notion ("evaluates AI at scale across 70 engineers") and Graphite stories; a link to "Take the eval maturity assessment".Source
  • 15

    Braintrust has 15 customer stories, including Notion evaluating AI across 70 engineers. Engineers trust other engineers, and the field marketing role is asked to invest in "the communities we do business in".

    Medium impactMedium effort First result in 1 quarter Head of Field Marketing

    First action

    Ask three customer engineers to speak.

    How to measure it

    Peer guests per event

    Now
    –
    Target
    10
    By
    2 quarters

    Steps

    1. 1

      Pick speakers

      From customer stories.

      Head of Field Marketing 2 weeks

    2. 2

      Agree topics

      Real eval problems.

      Developer Advocate 2 weeks

    3. 3

      Invite peers

      Two each.

      Head of Field Marketing per event

    4. 4

      Thank speakers

      Gift or donation.

      Head of Field Marketing per event

    5. 5

      Follow up peers

      SDR note.

      Sales Development Representative per event

    6. 6

      Measure

      Peers to meetings.

      Founding Data Engineer quarterly

    First result
    1 quarter
    Cost
    Speaker gifts and travel
    Owner
    Head of Field Marketing

    Smallest test

    Book one customer for the next workshop.

    Why it works

    Invite customer engineers to speak at city events and workshops, give them a thank-you (a donation, travel, merch that "isn't trash"), and ask each to bring two peers from other companies.

    Customer speakers draw their peers, and every invited peer is a referral.

    Risks and how to handle them

    • Customers are busy

      Offer short talks and panels.

    Evidence

    • HiringHead of Field Marketing (San Francisco, posted October 1, 2026): "Events are about people"; the role builds "the field marketing strategy. Decide where we show up, who we bring together and what we do for them", across "audiences, cities and formats", "from a small technical workshop to a party with hundreds of guests"; "Build with Sales. Know their accounts, territories and deals" and "Own the plan for invitations and follow-up"; "Bring partners into the program"; and "Track the accounts and relationships we're building, the pipeline we're creating and the deals we're helping move". Wanted: someone who has "built field marketing programs at a devtool or infrastructure company".Source
    • Your websiteHomepage: "Agents fail differently than normal software. You need active observability to monitor and fix them." Observability, Evals and Discovery; Loop agent, Patterns, trace-to-dataset, an MCP server and SDKs "for Python, TypeScript, Go, Ruby, C#, and more"; SOC 2 Type II, HIPAA, GDPR and hybrid deployment; Notion ("evaluates AI at scale across 70 engineers") and Graphite stories; a link to "Take the eval maturity assessment".Source
    • Your websiteSitemap (480 URLs, checked October 22, 2026): 174 articles (including comparison pages such as braintrust-vs-datadog-llm-observability and arize-phoenix-vs-braintrust), 99 encyclopedia entries, 97 blog posts, 29 eval research studies, 15 Academy lessons, 15 customer stories and 9 workshops. Live pages linked from the homepage are missing from it: /startups, /product/observe, /product/evaluate, /product/discover, /resources/for-pms and /newsletter. There are no industry pages.Source

Big bets 3

Larger plays that can change the trajectory.

  • 16

    Every eval maturity assessment records which practices teams have: vibe checks, datasets, CI gates, drift alerts. Aggregated, that is a picture of how AI teams test their agents that no rival owns.

    High impactMedium effort First result in 2 quarters Founding Data Engineer

    First action

    Count assessment completions to date.

    How to measure it

    Assessments completed

    Now
    –
    Target
    2,000
    By
    3 quarters

    Steps

    1. 1

      Count responses

      Enough for a fair sample.

      Founding Data Engineer 1 week

    2. 2

      Add questions

      Company size and industry.

      Web team 1 week

    3. 3

      Survey events

      Same questions in person.

      Head of Field Marketing 1 quarter

    4. 4

      Analyze

      Anonymized aggregates only.

      Founding Data Engineer 1 month

    5. 5

      Publish

      Report, page and charts.

      Developer Advocate 1 month

    6. 6

      Use in sales

      Benchmark slide.

      Enterprise Account Executive (Central Region) ongoing

    First result
    2 quarters
    Cost
    Design and analysis time
    Owner
    Founding Data Engineer

    Smallest test

    Publish one chart from current answers.

    Why it works

    Publish a yearly report from anonymized assessment answers plus a short survey at events: share of teams with CI gates, online scoring and human review, by company size and industry. Launch it at a city event and give AEs a benchmark slide.

    Original data earns press and links, gets cited by AI assistants, and lets every prospect compare themselves to peers.

    Risks and how to handle them

    • Small sample

      Publish ranges and sample sizes.

    Evidence

    • Your websiteEval maturity assessment: visitors "Mark the quality practices your team already has in motion" from "No evals setup", "Manual QA only" and "Vibe checks" up to "Automatic CI gates", "Drift alerts" and "Online scores", then "Set confidence"; the page ends with "Talk to our team". The financial services workshop (August 26) offered "a reference architecture for financial services teams running Braintrust".Source
    • HiringFounding Data Engineer (Marketing department): "Braintrust is growing quickly across enterprise and self-serve customers"; the role defines "the business entities, metrics, and semantics that Product, Finance, and GTM all build on", owns ingestion "from product events, CRM, and other business systems", and works "with Sales, RevOps, and Marketing on the pipeline and customer data". Bonus: "usage-based pricing, consumption models, or PLG-plus-sales-assisted funnels". Stack includes Snowflake and dbt.Source
    • HiringHead of Field Marketing (San Francisco, posted October 1, 2026): "Events are about people"; the role builds "the field marketing strategy. Decide where we show up, who we bring together and what we do for them", across "audiences, cities and formats", "from a small technical workshop to a party with hundreds of guests"; "Build with Sales. Know their accounts, territories and deals" and "Own the plan for invitations and follow-up"; "Bring partners into the program"; and "Track the accounts and relationships we're building, the pipeline we're creating and the deals we're helping move". Wanted: someone who has "built field marketing programs at a devtool or infrastructure company".Source
    • Competitor researchRivals' own homepages: LangSmith, "Agent & LLM Observability Platform"; Arize, "Agent Observability, Evaluation & Improvement Platform"; Galileo, evaluation and monitoring "at enterprise scale"; Langfuse, "Open Source Agent Evals & Observability"; Datadog, agent and LLM observability inside its monitoring suite.Source
  • 17

    The Head of Field Marketing will own events "to a party with hundreds of guests", and Braintrust has customers like Notion, Stripe and Cloudflare to put on stage. No rival owns a conference about evals.

    High impactHigh effort First result in 3 quarters Head of Field Marketing

    First action

    Hold a date and venue for next year.

    How to measure it

    Conference-sourced pipeline

    Now
    –
    Target
    100 opportunities
    By
    3 quarters

    Steps

    1. 1

      Set goals

      Accounts, attendees, pipeline.

      Head of Field Marketing 2 weeks

    2. 2

      Book venue

      San Francisco first.

      Head of Field Marketing 1 month

    3. 3

      Program

      Customers and workshops.

      Developer Advocate 2 months

    4. 4

      Invite

      Target accounts with AEs.

      Enterprise Account Executive (Central Region) 2 months

    5. 5

      Run

      One day.

      Head of Field Marketing 1 day

    6. 6

      Follow up

      Every account within a week.

      Sales Development Representative 1 week

    7. 7

      Repeat

      London.

      Head of Field Marketing 6 months

    First result
    3 quarters
    Cost
    Conference budget
    Owner
    Head of Field Marketing

    Smallest test

    Run a half-day version after the first city events.

    Why it works

    Run a one-day conference: customer talks on real eval failures, hands-on workshops, the State of AI evals report and product launches. Repeat it in London six months later. Price tickets low and give free seats to target accounts.

    A flagship event gathers a year of pipeline in a day, and positions Braintrust as the home of the eval discipline.

    Risks and how to handle them

    • Too big too soon

      Start with 300 seats.

    Evidence

    • HiringHead of Field Marketing (San Francisco, posted October 1, 2026): "Events are about people"; the role builds "the field marketing strategy. Decide where we show up, who we bring together and what we do for them", across "audiences, cities and formats", "from a small technical workshop to a party with hundreds of guests"; "Build with Sales. Know their accounts, territories and deals" and "Own the plan for invitations and follow-up"; "Bring partners into the program"; and "Track the accounts and relationships we're building, the pipeline we're creating and the deals we're helping move". Wanted: someone who has "built field marketing programs at a devtool or infrastructure company".Source
    • Your websiteHomepage: "Agents fail differently than normal software. You need active observability to monitor and fix them." Observability, Evals and Discovery; Loop agent, Patterns, trace-to-dataset, an MCP server and SDKs "for Python, TypeScript, Go, Ruby, C#, and more"; SOC 2 Type II, HIPAA, GDPR and hybrid deployment; Notion ("evaluates AI at scale across 70 engineers") and Graphite stories; a link to "Take the eval maturity assessment".Source
    • News coverageBraintrust raised an $80 million Series B in February 2026 led by Iconiq, with Andreessen Horowitz, Greylock and others, at an $800 million valuation; the money is to scale engineering and go-to-market teams and expand its physical presence.Source
    • HiringSales roles open now: Enterprise Account Executive (Central Region, Denver, posted October 16, 2026; "Our primary users are software engineers"), Account Executive, Digital Natives (West Region, San Francisco, September 1), plus Strategic and Enterprise AEs listed for San Francisco, New York City, Chicago, Austin and London (EMEA), and Solutions Engineers for the West and East regions.Source
  • 18

    The @braintrust/otel package is downloaded about 126,000 times a week, and rivals from Datadog to Langfuse compete on tracing. Many engineering teams already emit OpenTelemetry and will not add a second SDK just to try evals.

    High impactHigh effort First result in 2 quarters Software Engineer

    First action

    List the five most common OTel setups among new workspaces.

    How to measure it

    OTel-first workspaces

    Now
    –
    Target
    25% of new
    By
    3 quarters

    Steps

    1. 1

      Audit setups

      Common frameworks.

      Software Engineer, SDK 2 weeks

    2. 2

      Write the guide

      One hour to first eval.

      Developer Advocate 1 month

    3. 3

      Fix rough edges

      Mapping and attributes.

      Software Engineer, SDK 1 quarter

    4. 4

      Publish

      Page and docs.

      Web team 1 week

    5. 5

      Promote

      Workshops and SDK notes.

      Developer Advocate ongoing

    6. 6

      Measure

      Activation and upgrade.

      Founding Data Engineer monthly

    First result
    2 quarters
    Cost
    Engineering time
    Owner
    Software Engineer, SDK

    Smallest test

    Write the guide for one framework.

    Why it works

    Publish a page and guide: send your existing OpenTelemetry traces to Braintrust and run your first eval in an hour, with examples for the most common frameworks. Track teams that start with OTel and later add the SDK.

    Removing the rewrite lowers the cost of switching or trying Braintrust beside an incumbent.

    Risks and how to handle them

    • Feature gaps vs the SDK

      Show what the SDK adds.

    Evidence

    • Developer reachThe braintrust npm package has about 976,000 weekly downloads (27 packages, about 2.3 million a week in total, including @braintrust/otel at about 126,000); the Python package was last released on October 21, 2026 (version 0.48.0); GitHub: autoevals has 1,064 stars. By comparison the YouTube channel has 1,380 subscribers, 109 videos and about 118,000 total views, and the community lives on Discord.Source
    • HiringSoftware Engineer, SDK (remote, posted October 8, 2026): builds "the open-source libraries that customers use to observe, understand, and optimize how AI operates", "embedded in thousands of applications"; owns "design, implementation, release, documentation, and community support"; integrates with "OpenAI, Anthropic, and Gemini"; and should "Be a great community ambassador". Support engineers are also being hired in London and Singapore.Source
    • Competitor researchRivals' own homepages: LangSmith, "Agent & LLM Observability Platform"; Arize, "Agent Observability, Evaluation & Improvement Platform"; Galileo, evaluation and monitoring "at enterprise scale"; Langfuse, "Open Source Agent Evals & Observability"; Datadog, agent and LLM observability inside its monitoring suite.Source

Wildcards 3

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

  • 19

    The field marketing role asks what "makes an engineer want to spend an evening with us". Engineers love breaking things; Braintrust's product is about finding where agents fail.

    Medium impactLow effort First result in 1 quarter Head of Field Marketing

    First action

    Pick one city and three volunteer agents.

    How to measure it

    Meetings per event

    Now
    –
    Target
    10
    By
    2 quarters

    Steps

    1. 1

      Find agents

      Customers or open-source projects.

      Developer Advocate 2 weeks

    2. 2

      Set rules

      Scope and safety.

      Developer Advocate 1 week

    3. 3

      Book venue

      Small and loud.

      Head of Field Marketing 2 weeks

    4. 4

      Run

      Tracing on screen.

      Developer Advocate 1 evening

    5. 5

      Publish

      Best failures.

      Developer Advocate 1 week

    6. 6

      Follow up

      SDR notes.

      Sales Development Representative 1 week

    First result
    1 quarter
    Cost
    Venue, food and prizes
    Owner
    Head of Field Marketing

    Smallest test

    Run one in San Francisco.

    Why it works

    Host an evening where teams bring an agent and others try to break it, with Braintrust tracing every attempt and turning failures into eval datasets live. Publish the best failures (with permission) as a short write-up.

    It is memorable, hands-on and shows the product working without a pitch.

    Risks and how to handle them

    • Embarrasses a guest

      Only volunteer agents, with consent.

    Evidence

    • HiringHead of Field Marketing (San Francisco, posted October 1, 2026): "Events are about people"; the role builds "the field marketing strategy. Decide where we show up, who we bring together and what we do for them", across "audiences, cities and formats", "from a small technical workshop to a party with hundreds of guests"; "Build with Sales. Know their accounts, territories and deals" and "Own the plan for invitations and follow-up"; "Bring partners into the program"; and "Track the accounts and relationships we're building, the pipeline we're creating and the deals we're helping move". Wanted: someone who has "built field marketing programs at a devtool or infrastructure company".Source
    • Your websiteHomepage: "Agents fail differently than normal software. You need active observability to monitor and fix them." Observability, Evals and Discovery; Loop agent, Patterns, trace-to-dataset, an MCP server and SDKs "for Python, TypeScript, Go, Ruby, C#, and more"; SOC 2 Type II, HIPAA, GDPR and hybrid deployment; Notion ("evaluates AI at scale across 70 engineers") and Graphite stories; a link to "Take the eval maturity assessment".Source
  • 20

    The billing engineer will help customers "understand their account state" as usage-based bills grow. A bill shows cost; it does not show what the team got.

    Medium impactMedium effort First result in 1 quarter Software Engineer

    First action

    Draft the receipt for ten customers by hand.

    How to measure it

    Billing disputes

    Now
    –
    Target
    -30%
    By
    2 quarters

    Steps

    1. 1

      Pick metrics

      Value and cost.

      Software Engineer, Billing 1 week

    2. 2

      Build the query

      Per workspace.

      Founding Data Engineer 2 weeks

    3. 3

      Design the email

      One screen.

      Web team 1 week

    4. 4

      Send

      Before each invoice.

      Software Engineer, Billing monthly

    5. 5

      Add a tip

      Cost-saving setting.

      Developer Support Engineer monthly

    6. 6

      Measure

      Disputes and expansion.

      Founding Data Engineer quarterly

    First result
    1 quarter
    Cost
    Engineering time
    Owner
    Software Engineer, Billing

    Smallest test

    Email ten receipts and ask for feedback.

    Why it works

    Send each paying workspace a monthly note before the invoice: traces logged, evals run, regressions caught before release, patterns found, and spend against the plan. Include one tip to reduce cost.

    Customers who see value next to cost argue less about bills and expand more.

    Risks and how to handle them

    • Shows low usage

      Pair it with a help offer.

    Evidence

    • HiringSoftware Engineer, Billing (posted October 13, 2026): builds the systems "that make Braintrust's pricing, usage, and payments accurate, scalable, and transparent", billing "processed data, evaluation scores, model usage, and data retention"; helps "launch pricing models that let customers adopt the platform with confidence as their AI workloads grow" and builds tools "helping customers understand their account state".Source
    • Your websitePricing: Starter $0 a month ($10 credits, 1 GB processed data then $4/GB, 10k scores then $2.50 per 1,000, 14-day retention); Pro $249 a month ($100 credits, 5 GB then $3/GB, 50k scores then $1.50 per 1,000, 30-day retention); Enterprise custom. Unlimited users on every plan. A usage calculator estimates traces and quality checks from tokens per month. "6-12 months free for qualifying startups".Source
    • Your websiteHomepage: "Agents fail differently than normal software. You need active observability to monitor and fix them." Observability, Evals and Discovery; Loop agent, Patterns, trace-to-dataset, an MCP server and SDKs "for Python, TypeScript, Go, Ruby, C#, and more"; SOC 2 Type II, HIPAA, GDPR and hybrid deployment; Notion ("evaluates AI at scale across 70 engineers") and Graphite stories; a link to "Take the eval maturity assessment".Source
  • 21

    Braintrust has 99 encyclopedia entries and 15 Academy lessons. Engineers now ask AI assistants "what is an LLM-as-a-judge" before they search; the answer they get often has no source.

    Medium impactLow effort First result in 1 month Developer Advocate

    First action

    Check what assistants cite for ten eval terms.

    How to measure it

    Top-20 terms citing Braintrust

    Now
    –
    Target
    10
    By
    2 quarters

    Steps

    1. 1

      Baseline

      Ten terms, four assistants.

      Developer Advocate 2 days

    2. 2

      Add definitions

      First sentence of each entry.

      Developer Advocate 2 weeks

    3. 3

      Structured data

      Defined-term markup.

      Web team 1 week

    4. 4

      Link examples

      To docs and Academy.

      Developer Advocate 2 weeks

    5. 5

      Monitor

      Monthly checks.

      Developer Advocate monthly

    6. 6

      Expand

      Add missing terms.

      Developer Advocate quarterly

    First result
    1 month
    Cost
    Writing time
    Owner
    Developer Advocate

    Smallest test

    Rewrite ten entries and recheck in a month.

    Why it works

    Add a one-sentence definition at the top of each entry, structured data, links between related terms and to a working example in Braintrust, and check monthly which sources assistants cite for the top 20 terms.

    Being the cited definition puts Braintrust's name into the first answer an engineer reads.

    Risks and how to handle them

    • Assistants change sources

      Track monthly and adjust.

    Evidence

    • Your websiteSitemap (480 URLs, checked October 22, 2026): 174 articles (including comparison pages such as braintrust-vs-datadog-llm-observability and arize-phoenix-vs-braintrust), 99 encyclopedia entries, 97 blog posts, 29 eval research studies, 15 Academy lessons, 15 customer stories and 9 workshops. Live pages linked from the homepage are missing from it: /startups, /product/observe, /product/evaluate, /product/discover, /resources/for-pms and /newsletter. There are no industry pages.Source
    • Competitor researchRivals' own homepages: LangSmith, "Agent & LLM Observability Platform"; Arize, "Agent Observability, Evaluation & Improvement Platform"; Galileo, evaluation and monitoring "at enterprise scale"; Langfuse, "Open Source Agent Evals & Observability"; Datadog, agent and LLM observability inside its monitoring suite.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.

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)
ProgrammableDigital

Show it, don't just describe it

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

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

More growth playbooks

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

  • CyberCube growth strategyCyber risk analytics for the insurance industry: underwriting (Account Manager), portfolio and catastrophe modelling (Portfolio Manager…22 growth ideas
  • Ashna Rana (a.r.) growth strategyFreelance conversion rate optimization and creative strategy: on-site testing programs, landing pages, UX research and design, data…21 growth ideas
  • AutoSitu growth strategyAn AI system for real estate development that connects site feasibility, design review and QA/QC, code compliance, submittal completeness…24 growth ideas
  • Code and Theory growth strategyA digital-first creative and technology agency (50% creatives, 50% engineers) that designs, builds and markets for large companies: brand…22 growth ideas
  • Ergo growth strategyAI revenue infrastructure: Ergo captures every call, email and Slack message a sales team has with customers and turns it into CRM…21 growth ideas
  • Misprint growth strategyA real-time bid/ask marketplace for trading cards: graded and ungraded Pokémon and One Piece singles, slabs and sealed product, with…23 growth ideas
  • Zafin growth strategyAn AI platform company for regulated institutions: Zafin AIOS orchestrates and governs agentic work; the Zafin Banking Platform runs…20 growth ideas
  • Granted growth strategyAn AI-native patient advocate that reviews medical bills, fights denials, finds financial assistance and negotiates with providers…21 growth ideas

See every playbook and website review

Independent research. Programmable wrote this playbook from public sources, as they stood on October 23, 2026, and cites them throughout. Braintrust did not commission, review or endorse it, and Programmable is not affiliated with Braintrust. 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 Braintrust runs its business.

Something wrong, or do you represent Braintrust? 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

    Events hub with city sign-ups, assessment routing, EMEA and French sessions, sitemap fixes, startup expiry outreach and a shareable pricing estimate.

  2. Days 31 to 6011 ideas

    Workshop video, SDK-channel invites, usage scoring, partner evenings, industry pages, research-led SDR emails, onboarding cohorts, site speed and customer speakers; red-team nights and value receipts.

  3. Days 61 to 903 ideas

    Commit to the big bets: the State of AI evals report, a flagship conference and an OpenTelemetry-first on-ramp.

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 one place for traces, datasets and evals. The most cited problem: agents regress silently.

  • Finding failure patterns they would not have found
  • Manual review does not scale
  • In their words: “evals”

Market

Competitive landscape

5 competitors compared: LangSmith, Arize, Galileo, Langfuse and 1 more.

  • Gap to exploit: In-person eval education for engineers
  • LangSmith: Tied in many minds to LangChain
  • Arize: Broader ML focus

Hiring signals

Braintrust is turning a developer-led product into a field and sales machine.

  • Head of Field Marketing: Cities, formats, partners and event pipeline. Ideas 1, 10, 15, 17, 19.
  • Sales Development Representative (London, French): First London SDR. Idea 3.
  • Sales Development Representative (San Francisco): Meeting targets in enterprise accounts. Ideas 2, 12.

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 first (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)
  • 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