ProgrammableMarketing/valency.io (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 · 10
  • Big bets · 3
  • Wildcards · 2
  • How we can help
Generated Oct 23, 2026
  1. Programmable
  2. Playbooks
  3. Valency
Valency logo

growth playbook · Oct 23, 2026

valency.io (opens in a new tab)

Valency growth playbook

Diagnosis

Valency launched Bond for all users and opened Hub in early October 2026, and its hiring says the next job is community: a first Community Manager to bring researchers to Hub, a GTM Engineer for demos, landing pages and POCs with national labs, and a Technical PM balancing enterprise and product-led growth.

The gaps are where researchers decide whether to trust and use it: Hub papers are not set up for Google Scholar, the corpus is described as 500 million, 44.5 million and "tens of millions", the events page still shows summer events as upcoming, and only researchers have their own page.

This playbook fixes those first, then builds peer stories, journal clubs, campus ambassadors, a quickstart repo, self-serve Teams and a POC kit, with bigger bets on Hub as a home for work held back by arXiv's cap, a quarterly data report and campus licenses.

Growth Score

7/10

Biggest Opportunity

Give researchers a reason to publish on Hub (Scholar indexing and author invitations) and a peer-led path into Bond, run by the new community team.

Biggest Risk

Researchers distrust anything that looks like hype, and well-funded tools (Elicit, Consensus) and open platforms compete for the same attention.

01

Start here

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

  1. 1

    Make every Hub paper findable in Google Scholar, so researchers have a reason to publish there

    Add the citation tags to the work page template and test five pages.

    High impact First result in 1 month
  2. 2

    Use one corpus number everywhere: 500 million, 44.5 million or "tens of millions"

    Write a one-line definition of the corpus count.

    High impact First result in 2 weeks
  3. 3

    Fix the events page, which still lists July and August events as upcoming

    Move the two old events to "Past events" today.

    Medium impact First result in 2 weeks

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 · 4Big bets · 9Fill-ins · 6Reconsider · 2
Effort →

Showing 21 of 21 ideas

This week 6

Quick wins with a clear first action.

  • 01

    Valency is hiring its first Community Manager to "recruit researchers to publish" on Hub. But Hub work pages carry none of the citation meta tags (citation_title, citation_author, citation_pdf_url) that Google Scholar reads, and the Hub sitemap lists no individual works. For a researcher, a paper that does not show up in Scholar barely counts, and that is the first thing the Community Manager will be asked.

    High impactLow effort First result in 1 month Technical Product Manager

    First action

    Add the citation tags to the work page template and test five pages.

    How to measure it

    Hub-native works found in Scholar

    Now
    –
    Target
    90%
    By
    90 days

    Steps

    1. 1

      Add tags

      citation_title, citation_author, citation_publication_date, citation_pdf_url.

      Software Engineer (Full-stack) 2 days

    2. 2

      Add structured data

      ScholarlyArticle with authors and versions.

      Software Engineer (Full-stack) 1 day

    3. 3

      Sitemap the works

      Monthly sitemap files of every work.

      Software Engineer (Full-stack) 1 day

    4. 4

      Handle duplicates

      Link to the arXiv or DOI version where one exists.

      Technical Product Manager 2 days

    5. 5

      Submit to Scholar

      Inclusion request and test searches.

      Technical Product Manager 1 week

    6. 6

      Tell researchers

      "Your Hub paper is in Scholar" in the publish flow and outreach.

      Community Manager ongoing

    First result
    1 month
    Cost
    Engineering time
    Owner
    Technical Product Manager

    Smallest test

    Tag ten Hub-native works and search for them in Scholar after three weeks.

    Why it works

    Add the standard Highwire citation tags and ScholarlyArticle structured data to every work and version page, link the PDF, list each work in a sitemap (split by month as it grows), and point works that also live on arXiv or a journal to that version so Scholar groups them. Then submit Hub to Scholar's inclusion process and check that works appear.

    Researchers publish where their work gets found and cited. Scholar indexing turns Hub from an extra place to post into part of a researcher's record, and the Community Manager gets an honest answer to "why publish here?".

    Ready to use

    ChecklistScholar-ready work page

    [ ] citation_title
    [ ] citation_author (one per author)
    [ ] citation_publication_date
    [ ] citation_pdf_url to a public PDF
    [ ] citation_doi or a link to the arXiv version when there is one
    [ ] ScholarlyArticle structured data
    [ ] Work listed in a sitemap
    [ ] Abstract visible in the HTML

    Risks and how to handle them

    • Scholar takes weeks to crawl

      Start now and check sample works each week.

    • Duplicates of arXiv papers

      Mark the original version so Scholar clusters them.

    Evidence

    • Valency HubHub work pages (checked October 22, 2026, two samples) are server-rendered with title, authors, abstract and share-card images, but carry no citation_title, citation_author or citation_pdf_url meta tags and no ScholarlyArticle structured data; one sample showed "1 view". The Hub sitemap lists 13 URLs (home, works index, submit, collections and docs) and no individual works.Source
    • HiringCommunity Manager (Berkeley, posted October 9, 2026), Valency's first: "Researchers can smell a sales pitch from a mile away. They trust tools their peers use, papers their peers share"; "Valency Hub just opened as a public alpha"; duties include "Recruit researchers to publish, discover, and discuss on Valency Hub", run Bluesky, X, LinkedIn, YouTube and Discord, host "office hours, AMAs with researchers, demos, and journal-club style discussions", in-person meetups "across Bay Area campuses", and build "Clear attribution from community activity to sign-ups and usage". The post says MCP Bond connects AI systems "to tens of millions of papers and preprints".Source
    • Valency HubHub homepage: "Publish your AI-assisted work here. It's versioned like code. Checked by machines before you publish. Reviewed and endorsed by verified humans, in the open"; featured collections include papers by or about the 2026 Nobel laureates (30 works) and recent AI-assisted math and computer science results; sign-up with Google, Microsoft or GitHub.Source
    • Your websiteHub page on valency.io: "Think GitHub, for science: publish in the open, follow and fork any paper"; "In some fields, up to 90% of new papers show signs of" AI assistance; free for individual researchers, "Private workspaces and deployments for teams and institutions".Source
  • 02

    The homepage and Bond page say "500+ million papers and preprints", the October 1 launch post says "over 44.5 million papers", and the Community Manager post says "tens of millions". Valency's own job post says researchers "can smell a sales pitch"; three different numbers are what they notice first.

    High impactLow effort First result in 2 weeks Head of Science

    First action

    Write a one-line definition of the corpus count.

    How to measure it

    Pages aligned

    Now
    –
    Target
    All
    By
    2 weeks

    Steps

    1. 1

      Define

      What counts as a paper.

      Head of Science 1 day

    2. 2

      Count

      Current totals by type.

      Software Engineer (Full-stack) 1 day

    3. 3

      Publish

      Number, definition and date on Bond.

      Community Manager 1 day

    4. 4

      Update

      Homepage, blog, job posts.

      Community Manager 1 day

    5. 5

      Automate

      Live counter from the index.

      Software Engineer (Full-stack) 1 week

    6. 6

      Review

      Monthly.

      Head of Science monthly

    First result
    2 weeks
    Cost
    $0
    Owner
    Head of Science

    Smallest test

    Fix the homepage and Bond page today.

    Why it works

    Define what is counted (unique works, full text, metadata records, preprints) and publish one figure with a definition and a date, ideally a live counter on the Bond page. Update the homepage, blog, pricing and job posts to match.

    A precise, sourced number earns trust with skeptical scientists and gives journalists a figure they can quote.

    Risks and how to handle them

    • The defined number is smaller

      Show both: works with full text and records indexed.

    Evidence

    • Your websiteHomepage: "The Home for People-Centered, AI-Accelerated Research"; MCP Bond connects "500+ million papers and preprints" to any LLM, "Free for academics and non-profits"; "Roughly 35 million researchers worldwide"; "One connector, 38 tools"; a quote from a senior scientist at Lawrence Berkeley National Laboratory; "Built with a leading national laboratory".Source
    • Valency blogBlog posts: "Valency Bond Is Live for All Users" (October 1, 2026: "instantly access over 44.5 million papers"); "Valency Hub: A new home for research" (October 6); a Hub collection of 722 AI-assisted math results (October 6); "arXiv is tapping the brakes" on "A cap of two submissions a month" (October 1); advisory board posts on the creator of IPython and Jupyter and on arXiv's scientific director (September 15).Source
    • HiringCommunity Manager (Berkeley, posted October 9, 2026), Valency's first: "Researchers can smell a sales pitch from a mile away. They trust tools their peers use, papers their peers share"; "Valency Hub just opened as a public alpha"; duties include "Recruit researchers to publish, discover, and discuss on Valency Hub", run Bluesky, X, LinkedIn, YouTube and Discord, host "office hours, AMAs with researchers, demos, and journal-club style discussions", in-person meetups "across Bay Area campuses", and build "Clear attribution from community activity to sign-ups and usage". The post says MCP Bond connects AI systems "to tens of millions of papers and preprints".Source
  • 03

    events.valency.io shows "Let's Talk AI & Science" on July 30 and "AI in Science Drinks" in London on August 24 under "Upcoming events", each labelled "in 3 weeks" or "in 4 weeks". The new Community Manager is hired to run a "regular cadence of virtual events"; the page they will link to says nothing has happened since August.

    Medium impactLow effort First result in 2 weeks Community Manager

    First action

    Move the two old events to "Past events" today.

    How to measure it

    Monthly event sign-ups

    Now
    –
    Target
    100
    By
    90 days

    Steps

    1. 1

      Clean up

      Past events moved.

      Community Manager 1 hour

    2. 2

      Fix labels

      Relative dates at load time.

      Software Engineer (Full-stack) 2 hours

    3. 3

      Add dates

      Next two sessions.

      Community Manager 1 day

    4. 4

      Add recordings

      Where available.

      Community Manager 1 week

    5. 5

      Link

      Discord, newsletter, site footer.

      Community Manager 1 day

    6. 6

      Track

      Sign-ups per event.

      Community Manager monthly

    First result
    2 weeks
    Cost
    $0
    Owner
    Community Manager

    Smallest test

    Clean the page today.

    Why it works

    Move past events to a "Past events" list with photos or recordings, add the next two dates (office hours and a journal club), and make sure the relative labels are computed when the page loads. Link the page from Discord and the newsletter.

    A live calendar is the simplest proof of an active community.

    Risks and how to handle them

    • Few events at first

      Monthly office hours keep the page current.

    Evidence

    • Valency eventsevents.valency.io, fetched twice on October 22, 2026: under "Upcoming events below" it lists "Let's Talk AI & Science" (Thu, Jul 30, "in 3 weeks") and "AI in Science Drinks · London" (Mon, Aug 24, "in 4 weeks").Source
    • HiringCommunity Manager (Berkeley, posted October 9, 2026), Valency's first: "Researchers can smell a sales pitch from a mile away. They trust tools their peers use, papers their peers share"; "Valency Hub just opened as a public alpha"; duties include "Recruit researchers to publish, discover, and discuss on Valency Hub", run Bluesky, X, LinkedIn, YouTube and Discord, host "office hours, AMAs with researchers, demos, and journal-club style discussions", in-person meetups "across Bay Area campuses", and build "Clear attribution from community activity to sign-ups and usage". The post says MCP Bond connects AI systems "to tens of millions of papers and preprints".Source
  • 04

    Free Bond gives 100 calls a day unverified and 1,000 a week after verifying a work email. Verified users are the ones Valency can reach, and their email domain shows which universities, labs and companies are trying Bond.

    High impactLow effort First result in 1 month GTM Engineer

    First action

    Measure today's share of free users who verify.

    How to measure it

    Verification rate

    Now
    –
    Target
    50%
    By
    60 days

    Steps

    1. 1

      Measure

      Current rate.

      GTM Engineer 1 day

    2. 2

      Add the nudge

      Calls left and verify link in responses.

      Software Engineer (Full-stack) 3 days

    3. 3

      Email

      One message after first use.

      GTM Engineer 1 day

    4. 4

      Domain report

      Weekly by institution.

      GTM Engineer 2 days

    5. 5

      Flag clusters

      Three or more users at one place.

      GTM Engineer weekly

    6. 6

      Hand to sales

      Institutional outreach.

      Co-founder and CEO weekly

    First result
    1 month
    Cost
    Engineering time
    Owner
    GTM Engineer

    Smallest test

    Add the nudge for one week and compare.

    Why it works

    Show the remaining calls and the verify step inside the first responses, send one email after the first session, and report verified users by institution domain each week to the founders.

    Verification lifts usage, gives the team a contactable user and turns anonymous trials into an account list for institutional sales.

    Risks and how to handle them

    • Nudges annoy users

      Show only when the limit is near.

    Evidence

    • Your websiteBond pricing: Individual is free, "100 calls a day while unverified" and "1,000 calls a week when you verify a work email", "No card needed"; Teams is "$499 /month" billed annually ($599 month-to-month) for "10,000 calls a month" and unlimited seats; Enterprise is custom, "Unlimited calls", with private cloud, on-prem or air-gapped options. Bond names three audiences: researchers, "institutions and R&D teams", and "AI tool builders".Source
    • HiringGTM Engineer (Berkeley, posted July 27, 2026): buyers are "researchers, research computing teams, and technical evaluators at national labs and enterprise R&D orgs who will not take our word for it. They want to see it work, poke at the API"; duties include "sample MCP servers", "Build and develop landing pages and measure success", "Support sales and solutions engineering on technical evaluations, POCs, and RFPs with national labs", GitHub and Discord presence, and "Prototype and test PLG mechanics — self-serve onboarding flows, in-product activation nudges".Source
    • HiringTechnical Product Manager (posted July 31, 2026): Valency "sits between three audiences — skeptical academia, results-driven enterprise, and AI-forward amplifiers"; the role balances "enterprise/national-lab requirements against open-platform, product-led growth models", owns "quality-control and trust safeguards for open-access products", and values "two-sided platform dynamics — marketplace growth, user-generated content quality control". A full-stack engineer was also posted on September 18.Source
  • 05

    Bond names three audiences, and the GTM Engineer is asked to "build and develop landing pages and measure success". Only researchers have their own page, which works well: four real jobs shown as transcripts. National labs, enterprise R&D and tool builders, the paying audiences, have none.

    High impactMedium effort First result in 1 month GTM Engineer

    First action

    Draft the institutions page from the last three POC conversations.

    How to measure it

    Institutional demo requests

    Now
    –
    Target
    10 a month
    By
    90 days

    Steps

    1. 1

      Gather questions

      From POCs and RFPs.

      GTM Engineer 1 week

    2. 2

      Write transcripts

      Real runs for each audience.

      GTM Engineer 1 week

    3. 3

      Build pages

      Same layout as researchers.

      Software Engineer (Full-stack) 1 week

    4. 4

      Add proof

      National lab quote and security facts.

      Head of Science 2 days

    5. 5

      Track

      UTM and form source.

      GTM Engineer 1 day

    6. 6

      Iterate

      Monthly from questions asked.

      GTM Engineer monthly

    First result
    1 month
    Cost
    Staff time only
    Owner
    GTM Engineer

    Smallest test

    Publish the tool-builder page first.

    Why it works

    Build two pages in the same style: "for institutions and R&D teams" (private corpus, deployment options, security, a POC in two weeks) and "for AI tool builders" (MCP endpoint, token efficiency, sample code, Teams pricing). Each shows real transcripts and a clear next step.

    Paying buyers need to see their own workflow; the researcher page proves the format works.

    Risks and how to handle them

    • Security claims need care

      State only what is in place today.

    Evidence

    • Your websitevalency.io sitemap: 13 URLs; the only audience page is /bond/for/researchers/, which shows four real jobs ("stay current", "build reading lists", "verify claims", "map people") as transcripts; there are no pages for institutions, research computing teams or AI tool builders.Source
    • Your websiteBond pricing: Individual is free, "100 calls a day while unverified" and "1,000 calls a week when you verify a work email", "No card needed"; Teams is "$499 /month" billed annually ($599 month-to-month) for "10,000 calls a month" and unlimited seats; Enterprise is custom, "Unlimited calls", with private cloud, on-prem or air-gapped options. Bond names three audiences: researchers, "institutions and R&D teams", and "AI tool builders".Source
    • HiringGTM Engineer (Berkeley, posted July 27, 2026): buyers are "researchers, research computing teams, and technical evaluators at national labs and enterprise R&D orgs who will not take our word for it. They want to see it work, poke at the API"; duties include "sample MCP servers", "Build and develop landing pages and measure success", "Support sales and solutions engineering on technical evaluations, POCs, and RFPs with national labs", GitHub and Discord presence, and "Prototype and test PLG mechanics — self-serve onboarding flows, in-product activation nudges".Source
    • HiringTechnical Product Manager (posted July 31, 2026): Valency "sits between three audiences — skeptical academia, results-driven enterprise, and AI-forward amplifiers"; the role balances "enterprise/national-lab requirements against open-platform, product-led growth models", owns "quality-control and trust safeguards for open-access products", and values "two-sided platform dynamics — marketplace growth, user-generated content quality control". A full-stack engineer was also posted on September 18.Source
  • 06

    Hub already curates timely collections (the 2026 Nobel laureates' work, recent AI-assisted math results) and adds papers within hours. The Community Manager is asked to turn papers into posts "researchers want to share". Bluesky is where many scientists now discuss papers.

    Medium impactLow effort First result in 1 month Community Manager

    First action

    Write the first thread from the Nobel collection.

    How to measure it

    Weekly social visits

    Now
    –
    Target
    1,000
    By
    90 days

    Steps

    1. 1

      Pick works

      Three a week.

      Community Manager weekly

    2. 2

      Write

      Accurate, no hype.

      Community Manager weekly

    3. 3

      Check with authors

      Tag and invite them.

      Community Manager weekly

    4. 4

      Adapt

      LinkedIn and X versions.

      Community Manager weekly

    5. 5

      Track

      UTM per post.

      Community Manager weekly

    6. 6

      Review

      Which topics drive sign-ups.

      Community Manager monthly

    First result
    1 month
    Cost
    $0
    Owner
    Community Manager

    Smallest test

    Post one thread this week.

    Why it works

    Each week post one thread: three new works with a plain-language line each, one researcher quote and a link to the collection. Repeat the best on LinkedIn and X in each platform's style, and track sign-ups from each post.

    Useful curation earns follows; a weekly rhythm builds the habit of checking Hub.

    Risks and how to handle them

    • Misstating a result

      Quote abstracts and link sources.

    Evidence

    • Valency HubHub homepage: "Publish your AI-assisted work here. It's versioned like code. Checked by machines before you publish. Reviewed and endorsed by verified humans, in the open"; featured collections include papers by or about the 2026 Nobel laureates (30 works) and recent AI-assisted math and computer science results; sign-up with Google, Microsoft or GitHub.Source
    • HiringCommunity Manager (Berkeley, posted October 9, 2026), Valency's first: "Researchers can smell a sales pitch from a mile away. They trust tools their peers use, papers their peers share"; "Valency Hub just opened as a public alpha"; duties include "Recruit researchers to publish, discover, and discuss on Valency Hub", run Bluesky, X, LinkedIn, YouTube and Discord, host "office hours, AMAs with researchers, demos, and journal-club style discussions", in-person meetups "across Bay Area campuses", and build "Clear attribution from community activity to sign-ups and usage". The post says MCP Bond connects AI systems "to tens of millions of papers and preprints".Source
    • Social channelsYouTube @valencyio: 23 subscribers, 38 videos, 1,287 total views. Valency is also on Bluesky, X, LinkedIn and Discord, and runs a newsletter on Buttondown.Source

This quarter 10

The core plays for the next three months.

  • 07

    The homepage quotes a Lawrence Berkeley National Laboratory scientist who went from days of searching to minutes. Researchers "trust tools their peers use"; peer stories are the strongest marketing Valency can make.

    Medium impactLow effort First result in 1 quarter Community Manager

    First action

    Ask three active users for a 30-minute interview.

    How to measure it

    Spotlight sign-ups

    Now
    –
    Target
    50 a month
    By
    2 quarters

    Steps

    1. 1

      Find users

      Active Bond users.

      Community Manager 1 week

    2. 2

      Interview

      30 minutes each.

      Community Manager 2 weeks

    3. 3

      Write

      Transcript and outcome.

      Community Manager 1 week

    4. 4

      Film

      Two-minute video.

      Community Manager 1 week

    5. 5

      Publish

      Blog, YouTube, social.

      Community Manager biweekly

    6. 6

      Measure

      Sign-ups per story.

      Community Manager monthly

    First result
    1 quarter
    Cost
    Staff time only
    Owner
    Community Manager

    Smallest test

    Publish the first spotlight.

    Why it works

    Publish one spotlight every two weeks: the researcher, the question, the transcript and what changed. Cut a two-minute video for YouTube and a short clip for Bluesky and LinkedIn.

    Specific peer stories persuade scientists more than features, and they feed every channel.

    Risks and how to handle them

    • Researchers are busy

      Offer written interviews too.

    Evidence

    • Your websiteHomepage: "The Home for People-Centered, AI-Accelerated Research"; MCP Bond connects "500+ million papers and preprints" to any LLM, "Free for academics and non-profits"; "Roughly 35 million researchers worldwide"; "One connector, 38 tools"; a quote from a senior scientist at Lawrence Berkeley National Laboratory; "Built with a leading national laboratory".Source
    • HiringCommunity Manager (Berkeley, posted October 9, 2026), Valency's first: "Researchers can smell a sales pitch from a mile away. They trust tools their peers use, papers their peers share"; "Valency Hub just opened as a public alpha"; duties include "Recruit researchers to publish, discover, and discuss on Valency Hub", run Bluesky, X, LinkedIn, YouTube and Discord, host "office hours, AMAs with researchers, demos, and journal-club style discussions", in-person meetups "across Bay Area campuses", and build "Clear attribution from community activity to sign-ups and usage". The post says MCP Bond connects AI systems "to tens of millions of papers and preprints".Source
    • Social channelsYouTube @valencyio: 23 subscribers, 38 videos, 1,287 total views. Valency is also on Bluesky, X, LinkedIn and Discord, and runs a newsletter on Buttondown.Source
  • 08

    The Community Manager will host "journal-club style discussions", and Hub collections already group timely work. A recurring session gives researchers a reason to come back to Hub.

    Medium impactLow effort First result in 1 quarter Community Manager

    First action

    Book one author for the first session.

    How to measure it

    Repeat attendance

    Now
    –
    Target
    40%
    By
    2 quarters

    Steps

    1. 1

      Pick collections

      Monthly.

      Community Manager monthly

    2. 2

      Invite authors

      One per session.

      Community Manager monthly

    3. 3

      Host

      One hour on video.

      Community Manager monthly

    4. 4

      Continue on Hub

      Conversation tab.

      Community Manager monthly

    5. 5

      Publish

      Recording and summary.

      Community Manager monthly

    6. 6

      Measure

      Attendance and Hub activity.

      Community Manager monthly

    First result
    1 quarter
    Cost
    $0
    Owner
    Community Manager

    Smallest test

    Host one session.

    Why it works

    Pick one collection each month, invite an author or reviewer, discuss on video, and continue the thread in the Hub conversation tab and Discord. Publish the recording and a summary.

    Regular events build habit and fill Hub's open-review threads with real discussion.

    Risks and how to handle them

    • Low turnout

      Partner with a lab group as co-host.

    Evidence

    • HiringCommunity Manager (Berkeley, posted October 9, 2026), Valency's first: "Researchers can smell a sales pitch from a mile away. They trust tools their peers use, papers their peers share"; "Valency Hub just opened as a public alpha"; duties include "Recruit researchers to publish, discover, and discuss on Valency Hub", run Bluesky, X, LinkedIn, YouTube and Discord, host "office hours, AMAs with researchers, demos, and journal-club style discussions", in-person meetups "across Bay Area campuses", and build "Clear attribution from community activity to sign-ups and usage". The post says MCP Bond connects AI systems "to tens of millions of papers and preprints".Source
    • Valency HubHub homepage: "Publish your AI-assisted work here. It's versioned like code. Checked by machines before you publish. Reviewed and endorsed by verified humans, in the open"; featured collections include papers by or about the 2026 Nobel laureates (30 works) and recent AI-assisted math and computer science results; sign-up with Google, Microsoft or GitHub.Source
    • Valency eventsevents.valency.io, fetched twice on October 22, 2026: under "Upcoming events below" it lists "Let's Talk AI & Science" (Thu, Jul 30, "in 3 weeks") and "AI in Science Drinks · London" (Mon, Aug 24, "in 4 weeks").Source
  • 09

    Valency is in Berkeley and wants in-person meetups "across Bay Area campuses". Graduate students and postdocs are heavy literature users and spread tools inside their labs.

    High impactMedium effort First result in 2 quarters Community Manager

    First action

    Recruit two ambassadors in two departments.

    How to measure it

    Ambassador sign-ups

    Now
    –
    Target
    300
    By
    2 quarters

    Steps

    1. 1

      Define the role

      Workshops and feedback.

      Community Manager 1 week

    2. 2

      Recruit

      Eight students.

      Community Manager 1 month

    3. 3

      Train

      Workshop kit.

      GTM Engineer 2 weeks

    4. 4

      Run workshops

      One per lab group.

      Community Manager monthly

    5. 5

      Track

      Codes per ambassador.

      GTM Engineer monthly

    6. 6

      Expand

      Next campuses.

      Community Manager 1 quarter

    First result
    2 quarters
    Cost
    Stipends
    Owner
    Community Manager

    Smallest test

    Run two lab workshops this month.

    Why it works

    Recruit eight student ambassadors across departments, give them verified Bond access, lab workshop materials and a small stipend, and track sign-ups by ambassador code. Expand to other campuses where usage clusters.

    Students carry tools into labs and later into their careers, and peer recommendations beat ads in academia.

    Risks and how to handle them

    • Uneven effort

      Pay per workshop held.

    Evidence

    • HiringCommunity Manager (Berkeley, posted October 9, 2026), Valency's first: "Researchers can smell a sales pitch from a mile away. They trust tools their peers use, papers their peers share"; "Valency Hub just opened as a public alpha"; duties include "Recruit researchers to publish, discover, and discuss on Valency Hub", run Bluesky, X, LinkedIn, YouTube and Discord, host "office hours, AMAs with researchers, demos, and journal-club style discussions", in-person meetups "across Bay Area campuses", and build "Clear attribution from community activity to sign-ups and usage". The post says MCP Bond connects AI systems "to tens of millions of papers and preprints".Source
    • Your websiteBond pricing: Individual is free, "100 calls a day while unverified" and "1,000 calls a week when you verify a work email", "No card needed"; Teams is "$499 /month" billed annually ($599 month-to-month) for "10,000 calls a month" and unlimited seats; Enterprise is custom, "Unlimited calls", with private cloud, on-prem or air-gapped options. Bond names three audiences: researchers, "institutions and R&D teams", and "AI tool builders".Source
    • Your websitePress kit and team: the CEO is a UC Berkeley astronomy professor whose previous AI company was acquired by GE; the Head of Science has founded national open-science and research-software initiatives; media contact [email protected].Source
  • 10

    The GTM Engineer is hired to build "sample MCP servers" and documentation "a skeptical staff scientist would actually use". Technical evaluators want to "poke at the API" before they trust it.

    High impactMedium effort First result in 1 quarter GTM Engineer

    First action

    Publish the first recipe with a notebook.

    How to measure it

    Repo-driven activations

    Now
    –
    Target
    100 a month
    By
    2 quarters

    Steps

    1. 1

      Pick recipes

      Four jobs plus private corpus.

      GTM Engineer 1 week

    2. 2

      Write code

      Notebooks and server.

      GTM Engineer 1 month

    3. 3

      Docs

      Setup per client.

      GTM Engineer 2 weeks

    4. 4

      Publish

      GitHub and Bond page.

      GTM Engineer 1 day

    5. 5

      Promote

      Discord and forums.

      Community Manager ongoing

    6. 6

      Measure

      Stars and sign-ups.

      GTM Engineer monthly

    First result
    1 quarter
    Cost
    Staff time only
    Owner
    GTM Engineer

    Smallest test

    Publish one recipe this month.

    Why it works

    Publish a GitHub repo with recipes for the four researcher jobs, notebooks, a sample server that combines Bond with a private folder of PDFs, and copy-paste setup for common MCP clients. Link it from Bond and track stars and sign-ups.

    Developers and research computing teams evaluate by running code; a good repo shortens time to first value and spreads by itself.

    Risks and how to handle them

    • Code goes stale

      Test the recipes in CI.

    Evidence

    • HiringGTM Engineer (Berkeley, posted July 27, 2026): buyers are "researchers, research computing teams, and technical evaluators at national labs and enterprise R&D orgs who will not take our word for it. They want to see it work, poke at the API"; duties include "sample MCP servers", "Build and develop landing pages and measure success", "Support sales and solutions engineering on technical evaluations, POCs, and RFPs with national labs", GitHub and Discord presence, and "Prototype and test PLG mechanics — self-serve onboarding flows, in-product activation nudges".Source
    • Your websitevalency.io sitemap: 13 URLs; the only audience page is /bond/for/researchers/, which shows four real jobs ("stay current", "build reading lists", "verify claims", "map people") as transcripts; there are no pages for institutions, research computing teams or AI tool builders.Source
    • Your websiteHomepage: "The Home for People-Centered, AI-Accelerated Research"; MCP Bond connects "500+ million papers and preprints" to any LLM, "Free for academics and non-profits"; "Roughly 35 million researchers worldwide"; "One connector, 38 tools"; a quote from a senior scientist at Lawrence Berkeley National Laboratory; "Built with a leading national laboratory".Source
  • 11

    Teams costs $499 a month for 10,000 calls and unlimited seats, but the path is "Talk to us". The Technical PM is asked to balance enterprise sales with product-led growth.

    High impactMedium effort First result in 1 quarter GTM Engineer

    First action

    Count how many verified users share an email domain.

    How to measure it

    Teams accounts

    Now
    –
    Target
    30
    By
    2 quarters

    Steps

    1. 1

      Define trial

      14 days, card up front.

      Technical Product Manager 1 week

    2. 2

      Build checkout

      Card and invites.

      Software Engineer (Full-stack) 1 month

    3. 3

      Usage view

      Calls per seat.

      Software Engineer (Full-stack) 2 weeks

    4. 4

      Email flow

      Trial days 1, 7, 12.

      GTM Engineer 1 week

    5. 5

      Sales trigger

      Threshold alert.

      GTM Engineer 1 week

    6. 6

      Review

      Conversion monthly.

      Technical Product Manager monthly

    First result
    1 quarter
    Cost
    Engineering time
    Owner
    GTM Engineer

    Smallest test

    Offer manual trials to five teams.

    Why it works

    Let groups start a 14-day Teams trial with a card, invite colleagues, and see usage. Offer a call only when a team passes a usage threshold. Price labs and startups the same way.

    Self-serve lets small teams buy fast, and usage data tells sales which ones to call.

    Risks and how to handle them

    • Trial abuse

      Card required and call caps during trial.

    Evidence

    • Your websiteBond pricing: Individual is free, "100 calls a day while unverified" and "1,000 calls a week when you verify a work email", "No card needed"; Teams is "$499 /month" billed annually ($599 month-to-month) for "10,000 calls a month" and unlimited seats; Enterprise is custom, "Unlimited calls", with private cloud, on-prem or air-gapped options. Bond names three audiences: researchers, "institutions and R&D teams", and "AI tool builders".Source
    • HiringTechnical Product Manager (posted July 31, 2026): Valency "sits between three audiences — skeptical academia, results-driven enterprise, and AI-forward amplifiers"; the role balances "enterprise/national-lab requirements against open-platform, product-led growth models", owns "quality-control and trust safeguards for open-access products", and values "two-sided platform dynamics — marketplace growth, user-generated content quality control". A full-stack engineer was also posted on September 18.Source
    • HiringGTM Engineer (Berkeley, posted July 27, 2026): buyers are "researchers, research computing teams, and technical evaluators at national labs and enterprise R&D orgs who will not take our word for it. They want to see it work, poke at the API"; duties include "sample MCP servers", "Build and develop landing pages and measure success", "Support sales and solutions engineering on technical evaluations, POCs, and RFPs with national labs", GitHub and Discord presence, and "Prototype and test PLG mechanics — self-serve onboarding flows, in-product activation nudges".Source
  • 12

    The GTM Engineer supports "technical evaluations, POCs, and RFPs with national labs". Each evaluation is likely built from scratch, which slows sales.

    High impactMedium effort First result in 2 quarters GTM Engineer

    First action

    Write the success criteria from the last evaluation.

    How to measure it

    POC conversion

    Now
    –
    Target
    50%
    By
    3 quarters

    Steps

    1. 1

      Write criteria

      Standard measures.

      GTM Engineer 1 week

    2. 2

      Test set

      Template for questions.

      Head of Science 1 week

    3. 3

      Security summary

      Current controls only.

      Technical Product Manager 2 weeks

    4. 4

      Report template

      Results and next step.

      GTM Engineer 1 week

    5. 5

      Run

      Next three POCs.

      GTM Engineer 1 quarter

    6. 6

      Publish

      Anonymized results.

      Community Manager 1 quarter

    First result
    2 quarters
    Cost
    Staff time only
    Owner
    GTM Engineer

    Smallest test

    Use the kit for the next POC.

    Why it works

    Package a standard POC: success criteria, a test set of the lab's own questions, a deployment note, a security summary, and a results report template. Run every evaluation from the kit and publish anonymized results.

    Repeatable POCs shorten the sales cycle and make wins comparable.

    Risks and how to handle them

    • Labs need custom work

      Keep a fixed core and a small custom part.

    Evidence

    • HiringGTM Engineer (Berkeley, posted July 27, 2026): buyers are "researchers, research computing teams, and technical evaluators at national labs and enterprise R&D orgs who will not take our word for it. They want to see it work, poke at the API"; duties include "sample MCP servers", "Build and develop landing pages and measure success", "Support sales and solutions engineering on technical evaluations, POCs, and RFPs with national labs", GitHub and Discord presence, and "Prototype and test PLG mechanics — self-serve onboarding flows, in-product activation nudges".Source
    • Your websiteBond pricing: Individual is free, "100 calls a day while unverified" and "1,000 calls a week when you verify a work email", "No card needed"; Teams is "$499 /month" billed annually ($599 month-to-month) for "10,000 calls a month" and unlimited seats; Enterprise is custom, "Unlimited calls", with private cloud, on-prem or air-gapped options. Bond names three audiences: researchers, "institutions and R&D teams", and "AI tool builders".Source
    • Your websiteHomepage: "The Home for People-Centered, AI-Accelerated Research"; MCP Bond connects "500+ million papers and preprints" to any LLM, "Free for academics and non-profits"; "Roughly 35 million researchers worldwide"; "One connector, 38 tools"; a quote from a senior scientist at Lawrence Berkeley National Laboratory; "Built with a leading national laboratory".Source
  • 13

    Hub's promise is open review "by verified humans", and the TPM role mentions two-sided platform dynamics. Hub knows each work's references; the authors of those references are the best possible reviewers, and an invitation about their own field is hard to ignore.

    High impactMedium effort First result in 2 quarters Technical Product Manager

    First action

    Send invites by hand for ten new works.

    How to measure it

    Invite-to-review rate

    Now
    –
    Target
    10%
    By
    2 quarters

    Steps

    1. 1

      Find authors

      From references.

      Software Engineer (Full-stack) 2 weeks

    2. 2

      Write the invite

      Short and specific.

      Community Manager 1 day

    3. 3

      Set limits

      One per person a month.

      Technical Product Manager 1 day

    4. 4

      Manual pilot

      Ten works.

      Community Manager 2 weeks

    5. 5

      Automate

      With opt-out.

      Software Engineer (Full-stack) 1 month

    6. 6

      Measure

      Reviews and sign-ups.

      Technical Product Manager monthly

    First result
    2 quarters
    Cost
    Engineering time
    Owner
    Technical Product Manager

    Smallest test

    Invite authors for ten works by hand.

    Why it works

    When a work is published, email the corresponding authors of a few key references: "A new paper builds on yours; read, comment or endorse." Limit to one invite a month per person, and show endorsements on the work.

    It brings expert reviewers and new users in one step, and each review makes Hub more useful.

    Risks and how to handle them

    • Feels like spam

      Strict limits and an easy opt-out.

    Evidence

    • Valency HubHub homepage: "Publish your AI-assisted work here. It's versioned like code. Checked by machines before you publish. Reviewed and endorsed by verified humans, in the open"; featured collections include papers by or about the 2026 Nobel laureates (30 works) and recent AI-assisted math and computer science results; sign-up with Google, Microsoft or GitHub.Source
    • HiringTechnical Product Manager (posted July 31, 2026): Valency "sits between three audiences — skeptical academia, results-driven enterprise, and AI-forward amplifiers"; the role balances "enterprise/national-lab requirements against open-platform, product-led growth models", owns "quality-control and trust safeguards for open-access products", and values "two-sided platform dynamics — marketplace growth, user-generated content quality control". A full-stack engineer was also posted on September 18.Source
    • Your websiteHub page on valency.io: "Think GitHub, for science: publish in the open, follow and fork any paper"; "In some fields, up to 90% of new papers show signs of" AI assistance; free for individual researchers, "Private workspaces and deployments for teams and institutions".Source
    • HiringCommunity Manager (Berkeley, posted October 9, 2026), Valency's first: "Researchers can smell a sales pitch from a mile away. They trust tools their peers use, papers their peers share"; "Valency Hub just opened as a public alpha"; duties include "Recruit researchers to publish, discover, and discuss on Valency Hub", run Bluesky, X, LinkedIn, YouTube and Discord, host "office hours, AMAs with researchers, demos, and journal-club style discussions", in-person meetups "across Bay Area campuses", and build "Clear attribution from community activity to sign-ups and usage". The post says MCP Bond connects AI systems "to tens of millions of papers and preprints".Source
  • 14

    The YouTube channel has 38 videos but 23 subscribers and 1,287 total views. The Community Manager now owns YouTube along with Bluesky, LinkedIn and X.

    Medium impactLow effort First result in 1 quarter Community Manager

    First action

    Pick the ten best videos by watch time.

    How to measure it

    Monthly views

    Now
    –
    Target
    5,000
    By
    2 quarters

    Steps

    1. 1

      Rank videos

      Watch time.

      Community Manager 1 day

    2. 2

      Cut clips

      Ten one-minute clips.

      Community Manager 2 weeks

    3. 3

      Add transcripts

      And chapters.

      Community Manager 1 week

    4. 4

      Post clips

      Other channels.

      Community Manager weekly

    5. 5

      Start a series

      Paper of the week.

      Community Manager weekly

    6. 6

      Measure

      Views and subscribers.

      Community Manager monthly

    First result
    1 quarter
    Cost
    $0
    Owner
    Community Manager

    Smallest test

    Post three clips this week.

    Why it works

    Re-cut the best ten videos into one-minute clips, add transcripts and chapters, post the clips on the other channels, and start one recurring format (a five-minute "paper of the week").

    The content already exists; short formats and a series give it an audience.

    Risks and how to handle them

    • Low production quality

      Clear audio matters more than visuals.

    Evidence

    • Social channelsYouTube @valencyio: 23 subscribers, 38 videos, 1,287 total views. Valency is also on Bluesky, X, LinkedIn and Discord, and runs a newsletter on Buttondown.Source
    • HiringCommunity Manager (Berkeley, posted October 9, 2026), Valency's first: "Researchers can smell a sales pitch from a mile away. They trust tools their peers use, papers their peers share"; "Valency Hub just opened as a public alpha"; duties include "Recruit researchers to publish, discover, and discuss on Valency Hub", run Bluesky, X, LinkedIn, YouTube and Discord, host "office hours, AMAs with researchers, demos, and journal-club style discussions", in-person meetups "across Bay Area campuses", and build "Clear attribution from community activity to sign-ups and usage". The post says MCP Bond connects AI systems "to tens of millions of papers and preprints".Source
  • 15

    The Community Manager's success measure includes "clear attribution from community activity to sign-ups and usage". Without it, events and posts cannot be compared.

    High impactLow effort First result in 1 month GTM Engineer

    First action

    Add the onboarding question this week.

    How to measure it

    Attributed sign-ups

    Now
    –
    Target
    80%
    By
    2 months

    Steps

    1. 1

      Define sources

      Channels and events.

      GTM Engineer 1 day

    2. 2

      Add question

      At sign-up.

      Software Engineer (Full-stack) 1 day

    3. 3

      UTM templates

      Per channel.

      Community Manager 1 day

    4. 4

      Event codes

      Per event.

      Community Manager 1 day

    5. 5

      Dashboard

      Weekly.

      GTM Engineer 1 week

    6. 6

      Review

      Weekly read-out.

      Community Manager weekly

    First result
    1 month
    Cost
    $0
    Owner
    GTM Engineer

    Smallest test

    Add the question today.

    Why it works

    Add UTM links to every post and event, give each event and ambassador a code, ask "where did you hear about Valency?" at sign-up, and build a simple weekly dashboard.

    Knowing which activities bring active users lets a small team spend time where it works.

    Risks and how to handle them

    • Self-reported answers vary

      Combine with UTM data.

    Evidence

    • HiringCommunity Manager (Berkeley, posted October 9, 2026), Valency's first: "Researchers can smell a sales pitch from a mile away. They trust tools their peers use, papers their peers share"; "Valency Hub just opened as a public alpha"; duties include "Recruit researchers to publish, discover, and discuss on Valency Hub", run Bluesky, X, LinkedIn, YouTube and Discord, host "office hours, AMAs with researchers, demos, and journal-club style discussions", in-person meetups "across Bay Area campuses", and build "Clear attribution from community activity to sign-ups and usage". The post says MCP Bond connects AI systems "to tens of millions of papers and preprints".Source
    • HiringGTM Engineer (Berkeley, posted July 27, 2026): buyers are "researchers, research computing teams, and technical evaluators at national labs and enterprise R&D orgs who will not take our word for it. They want to see it work, poke at the API"; duties include "sample MCP servers", "Build and develop landing pages and measure success", "Support sales and solutions engineering on technical evaluations, POCs, and RFPs with national labs", GitHub and Discord presence, and "Prototype and test PLG mechanics — self-serve onboarding flows, in-product activation nudges".Source
  • 16

    The Community Manager will "represent Valency at scientific conferences and hackathons". Researchers want to see Bond answer their own question before trusting it.

    Medium impactMedium effort First result in 2 quarters Community Manager

    First action

    List the next six relevant meetings and costs.

    How to measure it

    Sign-ups per conference

    Now
    –
    Target
    150
    By
    2 quarters

    Steps

    1. 1

      List meetings

      Six options.

      Community Manager 1 week

    2. 2

      Pick three

      By field and cost.

      Co-founder and CEO 1 week

    3. 3

      Prepare demo

      Bring-your-question flow.

      GTM Engineer 2 weeks

    4. 4

      Run

      Table or workshop.

      Community Manager per event

    5. 5

      Hackathon prize

      Best integration.

      GTM Engineer 1 event

    6. 6

      Follow up

      Within 48 hours.

      Community Manager per event

    First result
    2 quarters
    Cost
    Travel and sponsorship
    Owner
    Community Manager

    Smallest test

    Run one workshop at a local meeting.

    Why it works

    Pick three meetings across Valency's strongest fields (astronomy, computational biology, machine learning), run a "bring your question" live demo table or workshop, and sponsor one hackathon prize for the best Bond integration.

    In-person demos on attendees' own questions convert skeptics, and hackathons create example code.

    Risks and how to handle them

    • High cost per lead

      Prefer workshops over booths.

    Evidence

    • HiringCommunity Manager (Berkeley, posted October 9, 2026), Valency's first: "Researchers can smell a sales pitch from a mile away. They trust tools their peers use, papers their peers share"; "Valency Hub just opened as a public alpha"; duties include "Recruit researchers to publish, discover, and discuss on Valency Hub", run Bluesky, X, LinkedIn, YouTube and Discord, host "office hours, AMAs with researchers, demos, and journal-club style discussions", in-person meetups "across Bay Area campuses", and build "Clear attribution from community activity to sign-ups and usage". The post says MCP Bond connects AI systems "to tens of millions of papers and preprints".Source
    • HiringGTM Engineer (Berkeley, posted July 27, 2026): buyers are "researchers, research computing teams, and technical evaluators at national labs and enterprise R&D orgs who will not take our word for it. They want to see it work, poke at the API"; duties include "sample MCP servers", "Build and develop landing pages and measure success", "Support sales and solutions engineering on technical evaluations, POCs, and RFPs with national labs", GitHub and Discord presence, and "Prototype and test PLG mechanics — self-serve onboarding flows, in-product activation nudges".Source
    • Your websitePress kit and team: the CEO is a UC Berkeley astronomy professor whose previous AI company was acquired by GE; the Head of Science has founded national open-science and research-software initiatives; media contact [email protected].Source

Big bets 3

Larger plays that can change the trajectory.

  • 17

    Valency's own blog says arXiv now caps authors at two submissions a month and calls it "a sea change". Productive groups, especially those using AI tools, will have more work than slots.

    High impactHigh effort First result in 3 quarters Co founder and CEO

    First action

    Ask ten prolific authors how the cap affects them.

    How to measure it

    Monthly new works

    Now
    –
    Target
    500
    By
    3 quarters

    Steps

    1. 1

      Interview authors

      Ten.

      Community Manager 2 weeks

    2. 2

      Add DOIs

      Via a registration agency.

      Technical Product Manager 1 quarter

    3. 3

      Scholar indexing

      Idea 1.

      Software Engineer (Full-stack) 1 month

    4. 4

      Path to arXiv

      Export when slots free up.

      Software Engineer (Full-stack) 1 month

    5. 5

      Write the post

      With researchers.

      Co-founder and CEO 2 weeks

    6. 6

      Measure

      Submissions per month.

      Technical Product Manager monthly

    First result
    3 quarters
    Cost
    Engineering time and DOI fees
    Owner
    Co founder and CEO

    Smallest test

    Publish a short guide for authors this month.

    Why it works

    Make Hub the documented home for additional and versioned work: clear licensing, DOIs, Scholar indexing, and a one-click path to post on arXiv later. Explain it in a post written with researchers, not against arXiv.

    A real constraint creates demand for a trusted alternative, and Hub is built for versioned, reviewed work.

    Risks and how to handle them

    • Seen as competing with arXiv

      Advisory board input and a cooperative tone.

    Builds on

    Evidence

    • Valency blogBlog posts: "Valency Bond Is Live for All Users" (October 1, 2026: "instantly access over 44.5 million papers"); "Valency Hub: A new home for research" (October 6); a Hub collection of 722 AI-assisted math results (October 6); "arXiv is tapping the brakes" on "A cap of two submissions a month" (October 1); advisory board posts on the creator of IPython and Jupyter and on arXiv's scientific director (September 15).Source
    • Your websiteHub page on valency.io: "Think GitHub, for science: publish in the open, follow and fork any paper"; "In some fields, up to 90% of new papers show signs of" AI assistance; free for individual researchers, "Private workspaces and deployments for teams and institutions".Source
    • Valency HubHub work pages (checked October 22, 2026, two samples) are server-rendered with title, authors, abstract and share-card images, but carry no citation_title, citation_author or citation_pdf_url meta tags and no ScholarlyArticle structured data; one sample showed "1 view". The Hub sitemap lists 13 URLs (home, works index, submit, collections and docs) and no individual works.Source
    • Valency HubHub homepage: "Publish your AI-assisted work here. It's versioned like code. Checked by machines before you publish. Reviewed and endorsed by verified humans, in the open"; featured collections include papers by or about the 2026 Nobel laureates (30 works) and recent AI-assisted math and computer science results; sign-up with Google, Microsoft or GitHub.Source
  • 18

    Valency sees which fields publish AI-assisted work, how it is reviewed and what researchers ask Bond. The Hub page already cites that up to 90% of new papers in some fields show signs of AI help.

    High impactMedium effort First result in 3 quarters Head of Science

    First action

    Draft five charts from current data.

    How to measure it

    Report readers

    Now
    –
    Target
    10,000
    By
    1 year

    Steps

    1. 1

      Pick questions

      Five.

      Head of Science 1 week

    2. 2

      Pull data

      Aggregated.

      Software Engineer (Full-stack) 1 month

    3. 3

      Privacy check

      No user identifiable.

      Technical Product Manager 2 weeks

    4. 4

      Write

      Report and charts.

      Head of Science 1 month

    5. 5

      Launch

      Webinar and press.

      Co-founder and CEO 2 weeks

    6. 6

      Repeat

      Quarterly.

      Head of Science quarterly

    First result
    3 quarters
    Cost
    Analyst time
    Owner
    Head of Science

    Smallest test

    Publish one chart this month.

    Why it works

    Publish a short quarterly report: AI-assisted works by field, review activity, common questions, how authors credit AI. Release it with a webinar and offer the data to journalists.

    Original data makes Valency the reference on a hot topic and brings press, links and AI citations.

    Risks and how to handle them

    • Small early data

      Start with Hub plus public sources.

    Evidence

    • Your websiteHub page on valency.io: "Think GitHub, for science: publish in the open, follow and fork any paper"; "In some fields, up to 90% of new papers show signs of" AI assistance; free for individual researchers, "Private workspaces and deployments for teams and institutions".Source
    • Your websiteResources page: a directory of 31 AI systems for research (co-scientists, AI reviewers, autonomous labs) filterable by field and access, ending "come back and publish it on Valency Hub", plus a guide on "How to credit the AI you used".Source
    • Valency HubHub homepage: "Publish your AI-assisted work here. It's versioned like code. Checked by machines before you publish. Reviewed and endorsed by verified humans, in the open"; featured collections include papers by or about the 2026 Nobel laureates (30 works) and recent AI-assisted math and computer science results; sign-up with Google, Microsoft or GitHub.Source
    • Your websitePress kit and team: the CEO is a UC Berkeley astronomy professor whose previous AI company was acquired by GE; the Head of Science has founded national open-science and research-software initiatives; media contact [email protected].Source
  • 19

    Free access works for individuals, and Teams fits groups. Universities buy research tools through libraries and research computing, and Valency's team has led national open-science initiatives.

    High impactHigh effort First result in 4 quarters Co founder and CEO

    First action

    Find the two campuses with the most verified users.

    How to measure it

    Campus licenses

    Now
    –
    Target
    5
    By
    4 quarters

    Steps

    1. 1

      Find clusters

      Verified domains.

      GTM Engineer 1 week

    2. 2

      Design license

      Price and terms.

      Co-founder and CEO 1 month

    3. 3

      Librarian guide

      Setup and training.

      Community Manager 1 month

    4. 4

      Pilot

      Two campuses.

      Co-founder and CEO 2 quarters

    5. 5

      Report usage

      To the library.

      Software Engineer (Full-stack) 1 month

    6. 6

      Scale

      Next campuses.

      Co-founder and CEO 2 quarters

    First result
    4 quarters
    Cost
    Sales time
    Owner
    Co founder and CEO

    Smallest test

    Meet one library team.

    Why it works

    Design a campus license (single sign-on, usage reporting, training) priced per institution, pilot it with two universities where verified usage clusters, and publish a librarian guide.

    One campus deal covers thousands of users and turns grassroots use into revenue.

    Risks and how to handle them

    • Slow budget cycles

      Start pilots before budget season.

    Evidence

    • Your websiteBond pricing: Individual is free, "100 calls a day while unverified" and "1,000 calls a week when you verify a work email", "No card needed"; Teams is "$499 /month" billed annually ($599 month-to-month) for "10,000 calls a month" and unlimited seats; Enterprise is custom, "Unlimited calls", with private cloud, on-prem or air-gapped options. Bond names three audiences: researchers, "institutions and R&D teams", and "AI tool builders".Source
    • Your websitePress kit and team: the CEO is a UC Berkeley astronomy professor whose previous AI company was acquired by GE; the Head of Science has founded national open-science and research-software initiatives; media contact [email protected].Source
    • HiringTechnical Product Manager (posted July 31, 2026): Valency "sits between three audiences — skeptical academia, results-driven enterprise, and AI-forward amplifiers"; the role balances "enterprise/national-lab requirements against open-platform, product-led growth models", owns "quality-control and trust safeguards for open-access products", and values "two-sided platform dynamics — marketplace growth, user-generated content quality control". A full-stack engineer was also posted on September 18.Source

Wildcards 2

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

  • 20

    Valency already explains "How to credit the AI you used" and lists 31 AI systems for research. Authors need a clear, honest way to disclose AI help.

    Medium impactLow effort First result in 1 quarter Community Manager

    First action

    Design the badge and the citation line.

    How to measure it

    Badged papers

    Now
    –
    Target
    200
    By
    2 quarters

    Steps

    1. 1

      Design

      Badge and text.

      Community Manager 1 week

    2. 2

      Hub record

      Disclosure page per paper.

      Software Engineer (Full-stack) 2 weeks

    3. 3

      Guide

      How to add it.

      Community Manager 1 week

    4. 4

      Seed

      Ten early authors.

      Community Manager 1 month

    5. 5

      Promote

      Directory and social.

      Community Manager ongoing

    6. 6

      Track

      Badge links.

      GTM Engineer monthly

    First result
    1 quarter
    Cost
    Staff time only
    Owner
    Community Manager

    Smallest test

    Ask ten authors to use the badge.

    Why it works

    Offer a small badge and a citation line authors can add to any paper, listing the AI systems used, with a link to a Hub record. Promote it with the resources directory.

    A useful disclosure standard spreads on its own, and every badge links back to Hub.

    Risks and how to handle them

    • Seen as self-promotion

      Keep it neutral and open.

    Evidence

    • Your websiteResources page: a directory of 31 AI systems for research (co-scientists, AI reviewers, autonomous labs) filterable by field and access, ending "come back and publish it on Valency Hub", plus a guide on "How to credit the AI you used".Source
    • Your websiteHub page on valency.io: "Think GitHub, for science: publish in the open, follow and fork any paper"; "In some fields, up to 90% of new papers show signs of" AI assistance; free for individual researchers, "Private workspaces and deployments for teams and institutions".Source
    • Valency HubHub homepage: "Publish your AI-assisted work here. It's versioned like code. Checked by machines before you publish. Reviewed and endorsed by verified humans, in the open"; featured collections include papers by or about the 2026 Nobel laureates (30 works) and recent AI-assisted math and computer science results; sign-up with Google, Microsoft or GitHub.Source
  • 21

    Bond indexes new preprints within hours and the researcher page shows "stay current" as job one. Many researchers would take that as an email without opening an assistant.

    Medium impactMedium effort First result in 1 quarter GTM Engineer

    First action

    Send a manual digest to 20 users.

    How to measure it

    Digest opens

    Now
    –
    Target
    50%
    By
    2 quarters

    Steps

    1. 1

      Pick topics

      At sign-up.

      Software Engineer (Full-stack) 1 week

    2. 2

      Build digest

      Weekly query.

      GTM Engineer 2 weeks

    3. 3

      Send

      Newsletter tool.

      GTM Engineer 1 week

    4. 4

      Link to Hub

      Every item.

      Software Engineer (Full-stack) 1 day

    5. 5

      Measure

      Opens and visits.

      GTM Engineer weekly

    6. 6

      Improve

      Topic tuning.

      GTM Engineer monthly

    First result
    1 quarter
    Cost
    Engineering time
    Owner
    GTM Engineer

    Smallest test

    Send 20 manual digests.

    Why it works

    Let users pick keywords or authors at sign-up and send a weekly digest of new work with one-line summaries and links to Hub. Use the newsletter tool already in place.

    A weekly email keeps Valency in front of researchers who do not use AI assistants daily.

    Risks and how to handle them

    • Noisy results

      Let users rate items.

    Evidence

    • Your websitevalency.io sitemap: 13 URLs; the only audience page is /bond/for/researchers/, which shows four real jobs ("stay current", "build reading lists", "verify claims", "map people") as transcripts; there are no pages for institutions, research computing teams or AI tool builders.Source
    • Social channelsYouTube @valencyio: 23 subscribers, 38 videos, 1,287 total views. Valency is also on Bluesky, X, LinkedIn and Discord, and runs a newsletter on Buttondown.Source
    • Valency blogBlog posts: "Valency Bond Is Live for All Users" (October 1, 2026: "instantly access over 44.5 million papers"); "Valency Hub: A new home for research" (October 6); a Hub collection of 722 AI-assisted math results (October 6); "arXiv is tapping the brakes" on "A cap of two submissions a month" (October 1); advisory board posts on the creator of IPython and Jupyter and on arXiv's scientific director (September 15).Source

03

How Programmable can help

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

ProgrammableDigital

Show it, don't just describe it

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

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

More growth playbooks

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

  • GLG growth strategyThe expert network: expert calls (including AI-moderated calls), surveys, an Expert Content Library, events, advisory and board…21 growth ideas
  • Verisoul growth strategyA fake account and fraud detection platform: device fingerprinting, account linking, bot detection, email and phone intelligence, proxy…23 growth ideas
  • Xero growth strategyCloud accounting for small businesses and their accountants: accounting, payroll, invoicing, payments and bill pay (powered by Melio in…21 growth ideas
  • Astronomer growth strategyThe company behind Astro, a managed orchestration platform for data and AI powered by Apache Airflow, with Astro Private Cloud, Astro…21 growth ideas
  • Palatial growth strategyA platform and services team that turns photos, CAD, scans, text and video into validated, simulation-ready assets and scenes for robot…21 growth ideas
  • Brickanta growth strategyAn agentic AI platform for construction teams that reads project documents from tender to handover: bid/no-bid tender reports, contract…22 growth ideas
  • Socure growth strategyAI-native identity verification and fraud prevention: RiskOS decisioning and orchestration, document and biometric verification, fraud…22 growth ideas
  • Dispatch growth strategyA San Francisco space logistics startup (Y Combinator, Spring 2026) building Spark-1, a refurbishable, autonomous reentry vehicle that…23 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. Valency did not commission, review or endorse it, and Programmable is not affiliated with Valency. 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 Valency runs its business.

Something wrong, or do you represent Valency? 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

    Scholar-ready Hub pages, one corpus number, the events page, email verification, pages for institutions and builders, and the weekly Hub thread; add attribution.

  2. Days 31 to 6011 ideas

    Spotlights, journal clubs, campus ambassadors, the quickstart repo, Teams trials, the POC kit, author invites, YouTube clips and conference demos; the AI badge and digest.

  3. Days 61 to 903 ideas

    Commit to the big bets: Hub for work held back by the arXiv cap, the quarterly report and campus licenses.

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 literature reviews in minutes instead of days. The most cited problem: AI tools that hallucinate citations.

  • Every answer tied to a real, citable source
  • Keeping up with fast-moving fields
  • In their words: “grounded”

Market

Competitive landscape

4 competitors compared: Elicit, Consensus, ResearchHub and arXiv and journals.

  • Gap to exploit: Grounded literature access inside any AI assistant
  • Elicit: Not an MCP-first connector or a publishing venue
  • Consensus: Search app rather than infrastructure

Hiring signals

Valency has moved from building to adoption.

  • Community Manager: Recruit researchers to Hub; events; attribution. Ideas 1, 3, 6, 7, 8, 9, 15.
  • GTM Engineer: Landing pages, sample MCP servers, POCs, PLG. Ideas 4, 5, 10, 12.
  • Technical Product Manager: Enterprise vs product-led; two-sided platform. Ideas 11, 13.

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