From 15 conversations to 2 institutional pilots: run scientist relations as a product-led design-partner engine, not sponsorship

The question asked: draft a working plan for Keenable's founding GTM / Scientists Relations motion. Built from three sets: a proposed four-phase 12-week roadmap (4 phases), Keenable's live GTM / Scientists Relations job page (13 charter items), and 8 current external program examples from Consensus, Elicit, Google DeepMind and OpenAI. Targets are proposed, not historical; budget figures are planning assumptions.

The 12-week roadmap proposed targets

primary KPI target (count) funnel narrowing from broad conversations to focused pilots · hover a phase for objective and deliverables

What the visual cannot say alone

Keenable's own role page assigns five core responsibilities — channel voice, technical content, community, events, and end-to-end partnerships — and names a Stanford founders hackathon as a sample problem. app.keenable.ai

Institutional distribution scales: Consensus reports over 5 million users from over 10,000 universities via sitewide licenses and university-email signup, and works directly with 85+ universities. consensus.app · consensus.app

Value-exchange programs work when the participant side is concrete: DeepMind pairs funding and mentorship across seven UK university fellowship programs; OpenAI grants free frontier-model access to roughly 10,000 researchers (third-party claim). deepmind.google · davidborish.com

Verifiable evidence is the positioning that lands with scientists: OpenAI's deep research launch leads with reproducible scores (26.6% on Humanity's Last Exam, SOTA on GAIA), not slogans. openai.com

Charter from Keenable's live job page directly stated by Keenable

Mission

Make the world's knowledge accessible to agents.

Responsibility

Build and manage the developer community, including relationships with SMB and startup customers.
Organize events where strong engineers encounter Keenable, such as meetups, workshops, dinners, hackathons, and conferences.
Own Keenable's voice on X and other channels, including posts, launch threads, and benchmarks.
Own partnerships end-to-end from finding the partner to shipped integration and co-published launch post.
Write technical content drawn from what the team is building, such as blog posts, teardowns, and eval results.

Requirement

A feel for the developer zeitgeist on X.
Exceptional writing with a proven ability to create content that engineers actually shared.
Prefer flat titles and large scope.
Strong engineering foundation with experience shipping software and holding credible technical conversations about retrieval, agents, or production trade-offs.
Want the mission to make the world's knowledge accessible to agents.

Sample problem

Organize a founders hackathon at Stanford that puts Keenable's web search in the hands of builders and turns their projects into an authentic product story.
Turn an internal eval result into a benchmark post that lands on X and Hacker News without reading as marketing.

Caveat: the title says “GTM / Scientists Relations,” but the page body is a founding DevRel charter aimed mainly at agent builders and engineers. Everything below adapts that charter to scientists — recommendations, not claims Keenable already runs them.

The strategic plan around the roadmap recommendations & planning assumptions

Positioning & ICP

“Web access for research agents: live, citable, reproducible, and cheap enough to query freely.” Not another end-user literature-review app — scientists are the wedge, reference community, and source of hard evals for the infrastructure/API.
Initial ICP: computational, AI-native scientists and research engineers who code, use agents or notebooks, need current non-paper web sources, and can install an API/MCP tool without procurement.
Priority verticals: AI/ML, bioinformatics/biotech intelligence, climate/earth observation, open-source scientific tooling. Pick the first by measured activation and retention after interviews.

Value exchange & loops

Participants get free/expanded usage, engineering access, office hours, co-designed integrations, optional visibility. Keenable gets feedback, permitted eval sets, case studies, integrations, referrals. No pay-for-praise; publication and data rights defined up front.
Scientist loop: working workflow → reproducible eval or recipe → technical launch → peer adoption → more product signal.
Partner loop: shipped integration → co-launch → installs/API calls → case study → institutional pilot.

Content & channels

Pillars: reproducible retrieval benchmarks, build logs and teardowns, workflow recipes, scientist spotlights, honest failure analyses. Disclose methods, datasets, costs, latency, limits — evidence before slogans.
X and Hacker News for technical reach; GitHub, docs and integration registries for durable discovery and activation; targeted lab outreach and warm intros for design partners; workshops and office hours for activation; discipline conferences only after a workflow proves pull. No generic community server before repeated use.

Cadence, measurement & ownership

Weekly: product/relations review, office hours, one technical artifact. Biweekly cohort touchpoint; monthly benchmark and changelog; quarterly flagship report or hackathon.
Definitions: qualified = credible fit + named workflow + next meeting; activated = completed a real workflow with Keenable; retained = usage in 3 of 4 weeks; partner = integration shipped and co-marketed; pilot = named owner, success criteria, duration, expansion decision date.
Track API/MCP activation, time-to-first-success, WAU, cohort retention, queries per retained team, referred and content-assisted activations, partner-sourced usage, pilot conversion, product insights shipped.
Scientists Relations owns cohort, content, community, events, partner process, feedback synthesis; Engineering owns product quality, integrations, technical proof; founders/revenue own strategic lab and institutional deals; legal/security supports research and data terms. One shared weekly scorecard.

Decision gates

End of week 2: proceed only if ≥10 interviews reveal 3 repeated high-pain workflows.
End of week 4: continue a segment only if ≥50% of onboarded partners activate.
End of week 8: scale only if ≥8 teams retain, or a clear alternative segment outperforms.
End of week 12: choose one primary scientist vertical and one repeatable acquisition loop.

Lean 90-day budget planning assumption

$20k–$50k total, excluding salary and normal product usage; release spend by milestone, never upfront.
$5k–$12k small events and workshops · $5k–$15k participant grants, credits, travel · $3k–$8k benchmark and content production · $2k–$5k tooling, design, miscellaneous.

Do not do

No vanity-follower goal and no broad ambassador program at launch.
No conference sponsorship without an activation plan.
No benchmark that cannot be reproduced.
No mixing scientist relations with enterprise quota ownership.
No collection or publication of research data without explicit consent.

External pattern evidence

Model	Value exchange	Mechanism	Proof	Source
Consensus	an AI search engine for academic research	universities can provide a sitewide license, and users can create an account with their university email address	over 5 million students, researchers and faculty members from over 10,000 different universities	consensus.app
Consensus	an AI-powered search engine for academic research that is a search engine first, not a chatbot	institutional library trials or subscriptions, and user account creation with a university email	working directly with 85+ universities	consensus.app
Elicit	the academic literature specialist	—	5M+ total researchers; institutional adoption at NASA, Stanford, Takeda, Unilever, and B. Braun with 40+ named customers on the customer stories page	theaiagentindex.com
Elicit	A research assistant built specifically for academic literature review	—	Searches across roughly 125 million+ papers	pulserevops.com
Google DeepMind	Financial support, mentorship from senior Google DeepMind researchers, opportunities to attend conferences and events, in exchange for fostering AI talent and research communities	Postdoctoral fellowships, postgraduate scholarships, Masters programs, undergraduate research ready programs, and teacher training programs	So far, seven universities in the UK have launched fellowship programs.	deepmind.google
OpenAI	Researchers at selected universities get free access to frontier models, GPT-5.6 Sol Pro across ChatGPT, ChatGPT Work, and Codex, with expanded deep research, higher usage limits, larger context windows, business-grade privacy protections, and data excluded from model training by default.	Grants, free access program, workspace access, skills, and connectors.	10,000 researchers this summer, with access already live at the Institute for Advanced Study and École normale supérieure.	davidborish.com
OpenAI deep research	A practical research agent and agentic research workflow that can be scoped, supervised, redirected, and audited	—	—	youngju.dev
OpenAI deep research	a new agentic capability that conducts multi-step research on the internet for complex tasks	in ChatGPT, select 'deep research' in the message composer and enter your query	scores a new high at 26.6% accuracy on Humanity's Last Exam, and reaches a new state of the art (SOTA) on GAIA	openai.com

Primary pages: consensus.app, deepmind.google, openai.com. Third-party claims: theaiagentindex.com, pulserevops.com, davidborish.com, youngju.dev.

Roadmap rows

Phase	Weeks	Objective	Deliverables	Primary KPI	Target	Source
1. Instrument & recruit	Weeks 1–2	Establish baselines and recruit a 10–15 person scientist design-partner cohort.	Scientist landing page and onboarding; event taxonomy; CRM tags; 50-account target list; interview script; first benchmark draft.	Qualified scientist conversations	15	app.keenable.ai
2. Prove value	Weeks 3–4	Run concierge pilots on concrete workflows and publish reproducible evidence.	Three workflow recipes; two benchmark/eval posts; weekly office hours; public issue-and-feedback loop.	Activated design partners	10	app.keenable.ai
3. Create loops	Weeks 5–8	Turn successful use into reusable integrations, stories, and peer referrals.	Two shipped integrations; four scientist spotlights or case studies; one workshop; referral program; monthly changelog for research users.	Weekly retained scientist teams	8	consensus.app
4. Scale winners	Weeks 9–12	Focus resources on the highest-retention discipline and convert institutional demand.	Flagship benchmark report; two campus or lab partnerships; one focused hackathon; institutional pilot package; next-quarter segment plan.	Institutional pilots	2	deepmind.google

Method: assembled from three result sets — a 4-row proposed 12-week roadmap (weeks, objective, deliverables, KPI, target), 13 charter items from Keenable's live GTM / Scientists Relations careers page, and 8 rows of external program examples with audience, value exchange, mechanism and proof. Targets (15 → 10 → 8 → 2) are proposed goals, not measured results; budget ranges are planning assumptions. Keenable materials describe web search and content access for agents via MCP, CLI and REST API, targeting AI labs, inference platforms and researchers, with keyless use within limits. Some external figures come from third-party pages and are labeled as such. Audience column of the evidence set folded into prose for space.