ProgrammableMarketing/83sciences.ai (opens in a new tab)
PDF

Views

  • Plan
  • Roadmap
  • Research

On this page

  • The brief
  • Start here
  • Idea portfolio
  • Effort vs impact
  • This week · 7
  • This quarter · 11
  • Big bets · 4
  • Wildcards · 2
Generated Nov 10, 2026
  1. Programmable
  2. Playbooks
  3. 83 Sciences
83 Sciences logo

growth playbook · Nov 10, 2026

83sciences.ai (opens in a new tab)

83 Sciences growth playbook

Diagnosis

83 Sciences is a three-person Y Combinator Summer 2026 company with a strong scientific record (six JACS papers from the team), a working product (the Dalton platform and an iOS companion app) and a clear pitch: about 90% of experiments are never published, and that data can drive discovery. The public website tells a different story.

It presents a materials developer with four focus areas, never mentions the lab brain, does not link the platform page and offers only email addresses as a way in. The three public descriptions (homepage, YC profile, careers listing) also give different statistics and vertical lists.

This playbook first makes the site say what the company sells and gives labs a way to join: a partner path, a proper request-access form, aligned messaging, the app under the company name and the publications as proof. It then builds the partner engine (advisor and tech-transfer channels, a one-notebook pilot, onboarding, a weekly PI digest, alumni referrals, a discovery-contract offer). The big bets are a public dark-data benchmark, the RL-environment line, a cross-lab failed-experiment network and instrument integrations.

Growth Score

6/10

Biggest Opportunity

Turn the first cohort of university lab partners into a repeatable, visible partner program, so every lab that hears the "90% never published" pitch can see what it gets and apply in one step.

Biggest Risk

Well-funded R&D data platforms (Citrine, Uncountable, Albert Invent, Benchling) already sell to the same industrial teams, and a three-person company spread across four focus areas and two customer types may struggle to show depth in any one of them.

01

Start here

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

  1. 1

    Put the lab brain and a "become a lab partner" path on the homepage; today the site hides the product it sells

    Write the three-line homepage section and the /partners criteria from the founders' current partner pitch.

    High impact First result in 2 weeks
  2. 2

    Rewrite the title and meta description around the lab brain, and add a share image

    Write the new homepage title and description.

    Medium impact First result in 2 weeks
  3. 3

    Replace the Google Form waitlist on the platform page with a request-access form that qualifies labs

    Write the seven questions and the scoring rule.

    Medium impact First result in 2 weeks

02

Your idea portfolio

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

  • 9 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 · 1Big bets · 9Fill-ins · 8Reconsider · 6
Effort →

Showing 24 of 24 ideas

This week 7

Quick wins with a clear first action.

  • 01

    Everywhere else, 83 Sciences describes itself as a lab-data company. The YC profile says "AI-native materials discovery powered by unpublished experimental data". The open job describes a queryable "lab brain" and a first cohort of university lab partners. The iOS app is "Voice notes for your lab". The homepage says none of this. It reads "We are creating the materials of the future" and lists four materials focus areas. It never mentions Dalton or the lab brain, never links the /platform page and offers no way in except two email addresses. A PI who hears the pitch and visits the site finds a materials company with no partner offer. Anyone who does reach the platform finds an access-code gate and a Google Form.

    High impactLow effort First result in 2 weeks CEO and Co Founder

    First action

    Write the three-line homepage section and the /partners criteria from the founders' current partner pitch.

    How to measure it

    Qualified lab-partner applications per month

    Now
    –
    Target
    10 a month
    By
    90 days

    Steps

    1. 1

      Write the offer

      Who qualifies, what the lab gives, what it gets, the timeline, from the current partner pitch.

      CEO and Co-Founder 1 day

    2. 2

      Add the homepage section

      Three lines, a Dalton screenshot and one "Become a lab partner" button under the focus areas.

      CTO and Co-Founder 1 day

    3. 3

      Fix the navigation

      Add Platform and Partners to the header; link /platform from the homepage.

      CTO and Co-Founder 1 hour

    4. 4

      Build /partners

      Criteria, benefits, data ownership summary, timeline, FAQ and a short application form.

      COO and Co-Founder 3 days

    5. 5

      Split the audiences

      A separate "Discovery projects" link for industrial R&D teams.

      CEO and Co-Founder 1 day

    6. 6

      Route applications

      Form answers go to a shared inbox with a two-day reply target.

      COO and Co-Founder 2 hours

    7. 7

      Use the link

      Put /partners in every outreach email, talk slide and the YC profile.

      CEO and Co-Founder ongoing

    8. 8

      Review monthly

      Applications, qualification rate and partner starts.

      COO and Co-Founder monthly

    First result
    2 weeks
    Cost
    Founder time only
    Owner
    CEO and Co Founder

    Smallest test

    Add the homepage section and a mailto "Become a lab partner" button today, and count clicks for two weeks.

    Why it works

    Add a second homepage section under the focus areas: "Your lab's unpublished data, working for you". Give it three lines (capture by voice and photo at the bench, one queryable record of every experiment including failed runs, new materials and papers from your own history), a screenshot of Dalton and one button, "Become a lab partner". Add "Platform" to the navigation. Build /partners with who qualifies (materials and chemistry labs with years of notebooks and instrument files), what the lab gives (data access, an hour a week), what it gets (a structured lab brain, the Companion app, co-authored papers and commercialization upside), how data is owned and protected, and the timeline (the company says under six weeks from a partner's data to a first new material). End with a short application form. Industrial R&D teams get a separate "Discovery projects" link.

    The company's growth right now depends on landing lab partners, and the site is the first place a PI, department head or tech-transfer officer checks after hearing about it. A named offer with criteria and a form turns curiosity into applications. It also lets the founders send one link instead of writing the pitch again in every email.

    Ready to use

    CopyHomepage partner section

    Your lab's unpublished data, working for you
    
    Most experiments never reach a paper. We help your lab keep all of them.
    - Capture by voice and photo at the bench with the Companion app
    - One queryable record of every experiment, including failed runs
    - New materials, papers and commercialization from your own history
    
    [Become a lab partner]   [Industry discovery projects]

    Risks and how to handle them

    • More applications than the team can onboard

      State the cohort size and a start date on the page, and keep a ranked waitlist.

    • Losing the materials-company story

      Keep the focus areas first; the partner section explains where the discoveries come from.

    Evidence

    • Your websiteHomepage (checked November 10, 2026): headline "We are creating the materials of the future"; four focus areas: magnetic materials, critical mineral extraction materials, semiconductor packaging materials, and "Materials science RL environments" "With real experimental data"; navigation is Focus, Team & Open Roles and Publications; the only contact options are two email links; "Backed by Y Combinator"; no link to the /platform page, no customer or partner logos and no request-access button.Source
    • Your websitePlatform page: "Ask your lab brain anything through Dalton AI" and "Your historical corpus becomes one structured and queryable database"; one link to platform.83sciences.ai; no request-access form; not linked from the homepage.Source
    • HiringFounding AI Engineer (San Francisco, New York or remote US; $120K-$250K and 0.50%-5.00% equity; the only open role on the Y Combinator jobs page): "90% of experiments never make it to publication"; the company structures that data into a queryable "lab brain" and is working with its first cohort of university lab partners; the role builds data capture, structured records and querying, vision for handwritten notebooks and drawn structures, speech-to-text at the bench, lab-history reasoning agents and models that suggest process conditions, working directly with research partners.Source
    • Y Combinator profile"AI-native materials discovery powered by unpublished experimental data"; Summer 2026 batch, team of 3, San Francisco; target users are industrial R&D teams and academic labs in batteries, semiconductors, membranes, catalysts and pharma; example applications include PFAS replacements and a silicon anode binder compatible with existing battery manufacturing; says it has already discovered a novel material and drafted a manuscript; the CEO was previously an MIT fellow focused on commercializing IP.Source
    • Careers listingThird-party careers listing: "More than 85%" of experimental data is discarded, worth over $100B in R&D value a year; under six weeks from a partner's data to the first new material discovered; tools capture voice notes and instrument output and keep a record of every experiment, including failed runs; discovery contracts with industry, publications and commercialization with academia; priority verticals: critical minerals and metals, energy storage, catalysis and chemicals, life sciences solid forms, semiconductors.Source
  • 02

    Search results and link previews show "83 Sciences | Novel Materials" and a description listing magnetic materials, critical mineral extraction, semiconductor packaging and "reinforcement learning environments". Nothing in it mentions unpublished lab data, the platform or lab partners. The homepage has no social share image, so links posted after the YC launch, in Slack or on LinkedIn appear without a picture.

    Medium impactLow effort First result in 2 weeks CEO and Co Founder

    First action

    Write the new homepage title and description.

    How to measure it

    Search click-through on brand queries

    Now
    –
    Target
    +20%
    By
    60 days

    Steps

    1. 1

      Rewrite the homepage

      Title and description around lab data and discovery.

      CEO and Co-Founder 1 hour

    2. 2

      Make a share image

      1200x630 with logo and one line.

      COO and Co-Founder 2 hours

    3. 3

      Add the tags

      Open Graph and Twitter card tags in the Next.js layout.

      CTO and Co-Founder 1 hour

    4. 4

      Cover the other pages

      Titles and descriptions for /platform, /team and /publications.

      CEO and Co-Founder 1 hour

    5. 5

      Test previews

      Paste links into Slack, LinkedIn and iMessage.

      COO and Co-Founder 30 minutes

    6. 6

      Verify in Search Console

      Submit the homepage and check the snippet.

      CTO and Co-Founder 1 week

    First result
    2 weeks
    Cost
    $0
    Owner
    CEO and Co Founder

    Smallest test

    Ship the share image today and paste the link into one channel.

    Why it works

    Change the title to something like "83 Sciences | Materials discovery from your lab's unpublished data" and rewrite the description around the lab brain, partner labs and the discoveries. Add a 1200x630 share image with the rhombus logo and one line. Give /platform, /team and /publications their own titles and descriptions.

    For a new domain with few links, the snippet and the preview card are most of what people see before they click. A preview that states the offer gets more clicks from shared links.

    Risks and how to handle them

    • Dropping the materials keywords

      Keep one materials phrase in the description.

    Evidence

    • Your websiteHomepage metadata: title "83 Sciences | Novel Materials"; meta description "83 Sciences develops magnetic materials, critical mineral extraction materials, semiconductor packaging materials, and materials science reinforcement learning environments"; no social share image; no blog, pricing page or social profile links; domain registered June 24, 2026.Source
    • Y Combinator profile"AI-native materials discovery powered by unpublished experimental data"; Summer 2026 batch, team of 3, San Francisco; target users are industrial R&D teams and academic labs in batteries, semiconductors, membranes, catalysts and pharma; example applications include PFAS replacements and a silicon anode binder compatible with existing battery manufacturing; says it has already discovered a novel material and drafted a manuscript; the CEO was previously an MIT fellow focused on commercializing IP.Source
  • 03

    A researcher who finds platform.83sciences.ai sees "Enter your early access code". Without a code, the only option is a waitlist on a Google Form, which takes them to another site, and the page has no link back to the main website. A waitlist that collects a name and email tells the founders nothing about which labs to call first.

    Medium impactLow effort First result in 2 weeks COO and Co Founder

    First action

    Write the seven questions and the scoring rule.

    How to measure it

    Request-access submissions that become calls

    Now
    –
    Target
    40%
    By
    60 days

    Steps

    1. 1

      Write questions

      Seven fields, all optional except email and role.

      COO and Co-Founder 2 hours

    2. 2

      Build the form

      Native form on the platform page and /partners.

      CTO and Co-Founder 1 day

    3. 3

      Move the waitlist

      Import existing Google Form entries and retire the form.

      COO and Co-Founder 2 hours

    4. 4

      Score and route

      Strong fits alert the founders.

      COO and Co-Founder half a day

    5. 5

      Confirmation email

      What happens next and when.

      CEO and Co-Founder 1 hour

    6. 6

      Link home

      Logo and link back to 83sciences.ai.

      CTO and Co-Founder 30 minutes

    7. 7

      Review weekly

      Submissions, calls and partner starts.

      COO and Co-Founder weekly

    First result
    2 weeks
    Cost
    Engineering time only
    Owner
    COO and Co Founder

    Smallest test

    Add role and institution to the current waitlist this week.

    Why it works

    Put a short native form on the platform page and on /partners: name, role (PI, postdoc, student, industry), institution or company, field, rough number of years of notebooks, main instruments and how data is stored today. Score each answer, reply to strong fits within two days with a call link, and send everyone else a confirmation that explains next steps. Add a link back to the main site.

    Keeping visitors on the company's own domain looks more trustworthy to institutional buyers. A few qualifying questions let a three-person team spend calls on labs with the most data.

    Risks and how to handle them

    • Longer forms reduce sign-ups

      Keep extra questions optional.

    Builds on

    Evidence

    • Your websitePlatform entry page: "Dalton Electronic Lab Notebook and Data Structuring Platform"; visitors are asked to "Enter your early access code" or sign in; people without a code are pointed to a waitlist on a Google Form; no link back to the main site.Source
    • Your websitePlatform page: "Ask your lab brain anything through Dalton AI" and "Your historical corpus becomes one structured and queryable database"; one link to platform.83sciences.ai; no request-access form; not linked from the homepage.Source
  • 04

    The iOS Companion app ("Voice notes for your lab") records experiment sessions of up to an hour, collects photos and audio, and requires lab membership. Its App Store seller is shown as an individual developer, not 83 Sciences, Inc. The website never mentions the app. University IT and industrial security teams reviewing a tool that records audio in their labs will look for a company behind it.

    Medium impactLow effort First result in 1 month CTO and Co Founder

    First action

    Start Apple's organization enrollment (it needs a D-U-N-S number for the company).

    How to measure it

    Share of partner-lab members using the app weekly

    Now
    –
    Target
    50%
    By
    90 days

    Steps

    1. 1

      Get a D-U-N-S number

      For 83 Sciences, Inc., if it does not have one.

      COO and Co-Founder 1-2 weeks

    2. 2

      Enroll the organization

      Apple Developer Program as a company.

      COO and Co-Founder 1 week

    3. 3

      Transfer the app

      Use App Store Connect's app transfer to the new account.

      CTO and Co-Founder 1 day

    4. 4

      Update the listing

      Support URL, privacy URL and screenshots.

      CTO and Co-Founder half a day

    5. 5

      Add to the site

      App section with badge on /platform and /partners.

      CTO and Co-Founder 2 hours

    6. 6

      Ask for ratings

      In-app prompt after the tenth saved note.

      CTO and Co-Founder 1 day

    First result
    1 month
    Cost
    $99 a year Apple fee plus staff time
    Owner
    CTO and Co Founder

    Smallest test

    Add the App Store badge to /platform today.

    Why it works

    Move the app to an Apple Developer organization account in the company's legal name, so the listing shows 83 Sciences, Inc. as seller. Add an app section to /platform and /partners with screenshots and an App Store badge. Point the listing's support and privacy links at company pages.

    The seller name is one of the first trust checks for an institutional app that handles lab audio and photos. Linking the app from the site also shows visitors that the product exists and works today.

    Risks and how to handle them

    • Transfer disrupts current users

      App transfers keep users and ratings; do it between releases.

    Evidence

    • App Store83 Sciences Companion (iOS, "Voice notes for your lab"): records experiment sessions of up to one hour plus photos and typed notes, with transcripts and suggested titles and notebook destinations; requires an existing account and lab membership; no ratings yet; updated November 4, 2026; the seller is shown as an individual developer rather than 83 Sciences, Inc.; its website link is 83sciences.ai/platform.Source
    • Your websiteHomepage (checked November 10, 2026): headline "We are creating the materials of the future"; four focus areas: magnetic materials, critical mineral extraction materials, semiconductor packaging materials, and "Materials science RL environments" "With real experimental data"; navigation is Focus, Team & Open Roles and Publications; the only contact options are two email links; "Backed by Y Combinator"; no link to the /platform page, no customer or partner logos and no request-access button.Source
  • 05

    The company's three public descriptions disagree. The job post says "90% of experiments never make it to publication", while the careers listing says "More than 85%" of data is discarded. The homepage lists magnetic, critical-mineral, semiconductor-packaging and RL-environment work. The YC profile lists batteries, semiconductors, membranes, catalysts and pharma. The careers listing lists critical minerals and metals, energy storage, catalysis and chemicals, life sciences solid forms and semiconductors. Scientists notice when numbers change between pages.

    Medium impactLow effort First result in 1 week CEO and Co Founder

    First action

    Find the original source for the unpublished-data figure and decide on the wording.

    How to measure it

    Public profiles using the agreed statistic and verticals

    Now
    –
    Target
    100%
    By
    30 days

    Steps

    1. 1

      Source the number

      Find a citable study for the figure, or say "most".

      CTO and Co-Founder half a day

    2. 2

      Agree the verticals

      Three to five, in priority order.

      CEO and Co-Founder 1 hour

    3. 3

      Write a message sheet

      One line, the statistic, the verticals, two proof points.

      CEO and Co-Founder 2 hours

    4. 4

      Update the site

      Homepage focus areas and metadata.

      CTO and Co-Founder 1 hour

    5. 5

      Update profiles

      YC profile, job posts, careers listings and decks.

      COO and Co-Founder half a day

    6. 6

      Review quarterly

      As focus changes.

      CEO and Co-Founder quarterly

    First result
    1 week
    Cost
    $0
    Owner
    CEO and Co Founder

    Smallest test

    Make the job post and the careers listing use the same number today.

    Why it works

    Pick one statistic with a citable source and use it word for word on the homepage, YC profile, job posts and decks. Agree one list of three to five verticals and use it everywhere, in the same order. Keep a one-page message sheet so new hires and partners repeat it.

    A research audience checks claims. One sourced number and one focus list read as rigor, and repetition makes the pitch memorable.

    Risks and how to handle them

    • Narrowing scares off other fields

      Add "and other materials and chemistry labs" after the list.

    Evidence

    • HiringFounding AI Engineer (San Francisco, New York or remote US; $120K-$250K and 0.50%-5.00% equity; the only open role on the Y Combinator jobs page): "90% of experiments never make it to publication"; the company structures that data into a queryable "lab brain" and is working with its first cohort of university lab partners; the role builds data capture, structured records and querying, vision for handwritten notebooks and drawn structures, speech-to-text at the bench, lab-history reasoning agents and models that suggest process conditions, working directly with research partners.Source
    • Careers listingThird-party careers listing: "More than 85%" of experimental data is discarded, worth over $100B in R&D value a year; under six weeks from a partner's data to the first new material discovered; tools capture voice notes and instrument output and keep a record of every experiment, including failed runs; discovery contracts with industry, publications and commercialization with academia; priority verticals: critical minerals and metals, energy storage, catalysis and chemicals, life sciences solid forms, semiconductors.Source
    • Y Combinator profile"AI-native materials discovery powered by unpublished experimental data"; Summer 2026 batch, team of 3, San Francisco; target users are industrial R&D teams and academic labs in batteries, semiconductors, membranes, catalysts and pharma; example applications include PFAS replacements and a silicon anode binder compatible with existing battery manufacturing; says it has already discovered a novel material and drafted a manuscript; the CEO was previously an MIT fellow focused on commercializing IP.Source
    • Your websiteHomepage (checked November 10, 2026): headline "We are creating the materials of the future"; four focus areas: magnetic materials, critical mineral extraction materials, semiconductor packaging materials, and "Materials science RL environments" "With real experimental data"; navigation is Focus, Team & Open Roles and Publications; the only contact options are two email links; "Backed by Y Combinator"; no link to the /platform page, no customer or partner logos and no request-access button.Source
  • 06

    The team's strongest credential is peer-reviewed discovery: six JACS papers, including "MnBi2 Is a Permanent Magnet" and a generative method for solving crystal structures from powder diffraction. The Publications page lists only titles and DOIs, and the homepage does not mention any of them. A department head or industrial R&D lead will not decode the titles.

    Medium impactLow effort First result in 2 weeks CTO and Co Founder

    First action

    Write the two lines for the MnBi2 and powder-diffraction papers.

    How to measure it

    Applications from visitors who viewed Publications

    Now
    –
    Target
    25% of applications
    By
    90 days

    Steps

    1. 1

      Write summaries

      Two lines per paper.

      CTO and Co-Founder half a day

    2. 2

      Pick three

      Strongest for partner relevance.

      CEO and Co-Founder 1 hour

    3. 3

      Add figures

      Where the license allows, or redraw.

      CTO and Co-Founder 1 day

    4. 4

      Homepage strip

      "Discoveries by our team" with links.

      CTO and Co-Founder half a day

    5. 5

      Link to /partners

      A call to action under each summary.

      CTO and Co-Founder 1 hour

    6. 6

      Add the next paper

      When the partner manuscript is public.

      CTO and Co-Founder ongoing

    First result
    2 weeks
    Cost
    Founder time only
    Owner
    CTO and Co Founder

    Smallest test

    Add summaries for two papers this week.

    Why it works

    For each paper, add one plain-language sentence on what was found and one on how the same skill helps a partner lab (for example, solving structures from old diffraction files in a lab's archive). Put the three strongest papers on the homepage as a "Discoveries by our team" strip. Add a figure from each paper where licensing allows.

    Published discoveries are proof a three-person startup can actually find materials. Explaining them in plain language makes that proof usable by non-specialist buyers.

    Risks and how to handle them

    • Overclaiming what the papers show

      Have co-authors check each summary.

    Evidence

    • Your websitePublications page: seven papers, six in the Journal of the American Chemical Society (2022-2026), including "MnBi2 Is a Permanent Magnet" (2025) and "Crystal Structure Determination from Powder Diffraction Patterns with Generative Machine Learning" (2024), plus a 2026 ChemRxiv preprint on an agentic workflow for high-throughput synthesis; titles and DOI links only, no summaries.Source
    • Your websiteHomepage (checked November 10, 2026): headline "We are creating the materials of the future"; four focus areas: magnetic materials, critical mineral extraction materials, semiconductor packaging materials, and "Materials science RL environments" "With real experimental data"; navigation is Focus, Team & Open Roles and Publications; the only contact options are two email links; "Backed by Y Combinator"; no link to the /platform page, no customer or partner logos and no request-access button.Source
  • 07

    The site is already fast: PageSpeed scores 97 on mobile and 100 on desktop, and real visitors pass Core Web Vitals. Two small items remain: about 314 KB of image savings (most of the page weight) and touch targets too small or close together on mobile, the page's only accessibility issue.

    Low impactLow effort First result in 1 day CTO and Co Founder

    First action

    Switch the hero logo to next/image.

    How to measure it

    Mobile accessibility score

    Now
    –
    Target
    100
    By
    2 weeks

    Steps

    1. 1

      Convert images

      next/image with AVIF or WebP.

      CTO and Co-Founder 1 hour

    2. 2

      Size them

      Serve at displayed dimensions.

      CTO and Co-Founder 30 minutes

    3. 3

      Enlarge tap targets

      44px minimum with spacing.

      CTO and Co-Founder 30 minutes

    4. 4

      Check on a phone

      Nav, email and partner buttons.

      COO and Co-Founder 15 minutes

    5. 5

      Re-run PageSpeed

      Confirm the issues are gone.

      CTO and Co-Founder 15 minutes

    6. 6

      Repeat on new pages

      /partners and /platform when built.

      CTO and Co-Founder ongoing

    First result
    1 day
    Cost
    $0
    Owner
    CTO and Co Founder

    Smallest test

    Fix the tap targets in one commit today.

    Why it works

    Serve the logo and hero images through next/image as AVIF or WebP at the displayed size, and give the nav and email links at least 44px tap areas with spacing. Re-run PageSpeed after.

    Many visitors open the link from a phone after a talk or an email. Bigger tap targets make the contact and partner links easy to hit, and smaller images keep the site fast on conference Wi-Fi.

    Risks and how to handle them

    • Iridescent logo loses quality

      Compare quality settings side by side.

    Evidence

    • PageSpeedGoogle PageSpeed (November 10, 2026): mobile 97, desktop 100; real visitors pass Core Web Vitals (LCP 0.9 s, INP 49 ms, CLS 0.01 at the 75th percentile); about 314 KB of image savings available; one accessibility issue: "Touch targets do not have sufficient size or spacing."Source

This quarter 11

The core plays for the next three months.

  • 08

    83 Sciences says it has already discovered a novel material and drafted a manuscript, and that it takes under six weeks from a partner's data to a first new material. The site has no case study or partner story, so this claim is invisible to the next lab or R&D team.

    High impactMedium effort First result in 1 quarter CEO and Co Founder

    First action

    Agree with the partner lab what can be shared and when.

    How to measure it

    Applications after reading the story

    Now
    –
    Target
    15 in the quarter
    By
    1 quarter after publication

    Steps

    1. 1

      Agree terms

      What can be shared, embargo and credit.

      CEO and Co-Founder 1 week

    2. 2

      Interview the PI

      Before-and-after and the surprise.

      CEO and Co-Founder 1 hour

    3. 3

      Write the story

      Data, method, finding, timeline.

      CTO and Co-Founder 3 days

    4. 4

      Check the science

      Co-authors review.

      CTO and Co-Founder 1 week

    5. 5

      Publish

      On /partners after the paper is public.

      CTO and Co-Founder 1 day

    6. 6

      Reuse

      Talk, post and outreach slide.

      COO and Co-Founder 1 week

    First result
    1 quarter
    Cost
    Founder time only
    Owner
    CEO and Co Founder

    Smallest test

    Write an anonymized version with the field and timeline only.

    Why it works

    With the partner lab's agreement, write the story of the first discovery: what data the lab had (years of notebooks, failed runs), how it was captured, what the lab brain surfaced, the material found and the timeline. Publish it on /partners with a quote from the PI, and turn it into a talk, a LinkedIn post and a slide for every outreach.

    A real discovery from a real lab's unused data is the whole pitch in one example. Other PIs trust a peer's account more than a startup's claim.

    Risks and how to handle them

    • Publishing before the paper harms priority

      Wait for the preprint or acceptance.

    Builds on

    Evidence

    • Y Combinator profile"AI-native materials discovery powered by unpublished experimental data"; Summer 2026 batch, team of 3, San Francisco; target users are industrial R&D teams and academic labs in batteries, semiconductors, membranes, catalysts and pharma; example applications include PFAS replacements and a silicon anode binder compatible with existing battery manufacturing; says it has already discovered a novel material and drafted a manuscript; the CEO was previously an MIT fellow focused on commercializing IP.Source
    • Careers listingThird-party careers listing: "More than 85%" of experimental data is discarded, worth over $100B in R&D value a year; under six weeks from a partner's data to the first new material discovered; tools capture voice notes and instrument output and keep a record of every experiment, including failed runs; discovery contracts with industry, publications and commercialization with academia; priority verticals: critical minerals and metals, energy storage, catalysis and chemicals, life sciences solid forms, semiconductors.Source
  • 09

    83 Sciences' advisors include a CNSI director at UC Santa Barbara and an MIT associate professor, and the founders come from Stanford, MIT, Harvard and Columbia labs. Each of these departments has many groups sitting on years of unpublished notebooks. A warm introduction from a respected colleague is the fastest way to a PI's calendar.

    High impactMedium effort First result in 1 month CEO and Co Founder

    First action

    Write the forwardable note.

    How to measure it

    Partner labs started through advisor introductions

    Now
    –
    Target
    5
    By
    1 quarter

    Steps

    1. 1

      Write the note

      Two paragraphs plus /partners.

      CEO and Co-Founder 2 hours

    2. 2

      Brief advisors

      Ideal lab profile and fields.

      CEO and Co-Founder 1 week

    3. 3

      Book talks

      One department talk per advisor.

      COO and Co-Founder 1 month

    4. 4

      Follow up

      Within two days of each introduction.

      CEO and Co-Founder ongoing

    5. 5

      Alumni labs

      Founders' former groups and colleagues.

      CTO and Co-Founder 1 month

    6. 6

      Track

      Introduction to call to partner.

      COO and Co-Founder weekly

    First result
    1 month
    Cost
    Travel for talks
    Owner
    CEO and Co Founder

    Smallest test

    Ask one advisor for one introduction this week.

    Why it works

    Give each advisor a two-paragraph forwardable note and the /partners link, and ask for three introductions to groups with long data histories in the company's priority fields. Offer a department lunch talk on unpublished data. Track each introduction to a partner start.

    PIs ignore cold startup emails but answer colleagues. Department talks reach many labs in one hour.

    Risks and how to handle them

    • Advisor fatigue

      Ask for three, not open-ended help, and report back on each.

    Builds on

    Evidence

    • Your websiteTeam page: three co-founders (a CEO from catalysis and battery-materials work at Stanford and Sila Nanotechnologies, a CTO in quantum materials and generative models for chemical experimental data, a COO from robotics and biosignaling research); advisors from UC Santa Barbara (a CNSI director), MIT and a software company; no roles on the page itself, only a link to the Y Combinator jobs page.Source
    • HiringFounding AI Engineer (San Francisco, New York or remote US; $120K-$250K and 0.50%-5.00% equity; the only open role on the Y Combinator jobs page): "90% of experiments never make it to publication"; the company structures that data into a queryable "lab brain" and is working with its first cohort of university lab partners; the role builds data capture, structured records and querying, vision for handwritten notebooks and drawn structures, speech-to-text at the bench, lab-history reasoning agents and models that suggest process conditions, working directly with research partners.Source
  • 10

    The product's core skills are reading handwritten notebooks and drawn structures, transcribing bench speech and structuring instrument output. Asking a PI to hand over a whole lab history is a big step. A pilot on one notebook asks for little and shows the result on the lab's own data.

    High impactMedium effort First result in 1 month Founding AI Engineer

    First action

    Define the pilot's input limits and the one-week deliverable.

    How to measure it

    Pilots that become full partnerships

    Now
    –
    Target
    50%
    By
    1 quarter

    Steps

    1. 1

      Define scope

      One notebook or one folder, page limits.

      CTO and Co-Founder 1 day

    2. 2

      Build the pipeline run

      Repeatable scan-to-records process.

      Founding AI Engineer 2 weeks

    3. 3

      Write three questions

      Template queries that show value.

      CTO and Co-Founder 1 day

    4. 4

      Add to /partners

      "Start with one notebook".

      CEO and Co-Founder 1 day

    5. 5

      Run five pilots

      From the waitlist.

      Founding AI Engineer 1 month

    6. 6

      Review results

      Walk the PI through the records.

      CEO and Co-Founder 1 hour each

    7. 7

      Convert

      Full partnership proposal.

      CEO and Co-Founder 1 week

    First result
    1 month
    Cost
    Engineering time per pilot
    Owner
    Founding AI Engineer

    Smallest test

    Run one pilot on a founder's own old notebook.

    Why it works

    Package a fixed pilot: the lab scans one notebook (or shares one instrument folder), and within a week gets it back in Dalton as structured, queryable records, with three questions the lab brain can now answer about that work. End with a proposal for the full partnership. Put the offer on /partners.

    Seeing your own handwriting turned into a searchable record is more convincing than any demo, and the small scope fits an academic's caution about data.

    Risks and how to handle them

    • Messy notebooks give poor results

      Set expectations and show error rates honestly.

    Builds on

    Evidence

    • HiringFounding AI Engineer (San Francisco, New York or remote US; $120K-$250K and 0.50%-5.00% equity; the only open role on the Y Combinator jobs page): "90% of experiments never make it to publication"; the company structures that data into a queryable "lab brain" and is working with its first cohort of university lab partners; the role builds data capture, structured records and querying, vision for handwritten notebooks and drawn structures, speech-to-text at the bench, lab-history reasoning agents and models that suggest process conditions, working directly with research partners.Source
    • Your websitePlatform page: "Ask your lab brain anything through Dalton AI" and "Your historical corpus becomes one structured and queryable database"; one link to platform.83sciences.ai; no request-access form; not linked from the homepage.Source
  • 11

    A partner lab needs its members to record notes in the Companion app and to ask the lab brain questions; otherwise the data stays thin and the PI loses interest. The app requires lab membership and has no ratings yet, which suggests early usage. Without a clear onboarding path, adoption depends on the founders' personal follow-up.

    High impactMedium effort First result in 1 month COO and Co Founder

    First action

    Measure where each current partner lab stands on the three milestones.

    How to measure it

    Partner labs fully onboarded within 30 days

    Now
    –
    Target
    80%
    By
    1 quarter

    Steps

    1. 1

      Pick milestones

      Install, structure, first answer.

      COO and Co-Founder 1 day

    2. 2

      Measure current labs

      Baseline for the first cohort.

      COO and Co-Founder 2 days

    3. 3

      Build the dashboard

      Per-lab progress view.

      Founding AI Engineer 1 week

    4. 4

      Write check-ins

      Weekly email to the lab contact.

      COO and Co-Founder half a day

    5. 5

      Set triggers

      Call a lab that stalls for 10 days.

      COO and Co-Founder 1 day

    6. 6

      Review cohort

      What predicts success.

      COO and Co-Founder monthly

    First result
    1 month
    Cost
    Staff time only
    Owner
    COO and Co Founder

    Smallest test

    Track the three milestones in a spreadsheet for current partners.

    Why it works

    Define three first-month milestones: every lab member installs the app and records a session, the first 50 historical pages are structured, and the PI gets an answer to a real question from the lab's history. Show progress on a lab dashboard, send a short weekly check-in, and step in when a lab stalls.

    Clear early milestones predict which partners stay and produce discoveries, and they show the team where onboarding breaks.

    Risks and how to handle them

    • Labs feel monitored

      Share the dashboard with them as their own progress view.

    Evidence

    • App Store83 Sciences Companion (iOS, "Voice notes for your lab"): records experiment sessions of up to one hour plus photos and typed notes, with transcripts and suggested titles and notebook destinations; requires an existing account and lab membership; no ratings yet; updated November 4, 2026; the seller is shown as an individual developer rather than 83 Sciences, Inc.; its website link is 83sciences.ai/platform.Source
    • HiringFounding AI Engineer (San Francisco, New York or remote US; $120K-$250K and 0.50%-5.00% equity; the only open role on the Y Combinator jobs page): "90% of experiments never make it to publication"; the company structures that data into a queryable "lab brain" and is working with its first cohort of university lab partners; the role builds data capture, structured records and querying, vision for handwritten notebooks and drawn structures, speech-to-text at the bench, lab-history reasoning agents and models that suggest process conditions, working directly with research partners.Source
  • 12

    In most labs, students and postdocs record the experiments while the PI decides whether to keep a tool. Dalton keeps a record of every experiment, including failed runs, so it can tell the PI things they would not otherwise see: repeated failures, conditions that worked, results that echo older work.

    Medium impactMedium effort First result in 1 month Founding AI Engineer

    First action

    Hand-write one digest for one partner PI from their lab's records.

    How to measure it

    PIs clicking into Dalton from the digest

    Now
    –
    Target
    60% weekly
    By
    1 quarter

    Steps

    1. 1

      Write one by hand

      For one partner PI.

      CTO and Co-Founder 2 hours

    2. 2

      Ask for feedback

      What was useful.

      CEO and Co-Founder 1 week

    3. 3

      Template it

      Four sections, under 200 words.

      Founding AI Engineer 1 week

    4. 4

      Automate

      Generated from lab records, reviewed at first.

      Founding AI Engineer 2 weeks

    5. 5

      Roll out

      All partner PIs, opt-out link.

      COO and Co-Founder 1 day

    6. 6

      Measure

      Opens, clicks and renewals.

      COO and Co-Founder monthly

    First result
    1 month
    Cost
    Engineering time only
    Owner
    Founding AI Engineer

    Smallest test

    Send one hand-written digest to one PI.

    Why it works

    Every Monday, email each PI a short digest: experiments recorded, two notable results, one failed run that matches an older attempt, and one suggested question to ask the lab brain. Link each item into Dalton.

    It keeps the decision-maker seeing value without logging in, and it shows the lab-history reasoning at work.

    Risks and how to handle them

    • Errors in an automated summary

      Review digests by hand for the first month.

    Builds on

    Evidence

    • Careers listingThird-party careers listing: "More than 85%" of experimental data is discarded, worth over $100B in R&D value a year; under six weeks from a partner's data to the first new material discovered; tools capture voice notes and instrument output and keep a record of every experiment, including failed runs; discovery contracts with industry, publications and commercialization with academia; priority verticals: critical minerals and metals, energy storage, catalysis and chemicals, life sciences solid forms, semiconductors.Source
    • HiringFounding AI Engineer (San Francisco, New York or remote US; $120K-$250K and 0.50%-5.00% equity; the only open role on the Y Combinator jobs page): "90% of experiments never make it to publication"; the company structures that data into a queryable "lab brain" and is working with its first cohort of university lab partners; the role builds data capture, structured records and querying, vision for handwritten notebooks and drawn structures, speech-to-text at the bench, lab-history reasoning agents and models that suggest process conditions, working directly with research partners.Source
  • 13

    Academic labs turn over constantly: students graduate, postdocs become PIs or join industry, and the lab's knowledge leaves with them. That is the problem Dalton solves, and it is also a channel. Every partner-lab member who moves on carries the habit of voice notes and a queryable lab history into a new group.

    Medium impactLow effort First result in 2 quarters COO and Co Founder

    First action

    List partner-lab members expected to leave in the next year.

    How to measure it

    New partner labs referred by alumni

    Now
    –
    Target
    3
    By
    2 quarters

    Steps

    1. 1

      Design the seat

      One year, personal workspace, invite link.

      COO and Co-Founder 1 day

    2. 2

      Check data rules

      What the lab owns stays with the lab.

      CEO and Co-Founder 1 week

    3. 3

      Build the flow

      Offboarding to alumni workspace.

      Founding AI Engineer 1 week

    4. 4

      Track leavers

      Ask labs for expected departures.

      COO and Co-Founder ongoing

    5. 5

      Offer priority

      Fast onboarding for new PIs.

      COO and Co-Founder ongoing

    6. 6

      Measure

      Invites sent and partner starts.

      COO and Co-Founder quarterly

    First result
    2 quarters
    Cost
    Hosting and engineering time
    Owner
    COO and Co Founder

    Smallest test

    Offer the seat to the next partner-lab member who leaves.

    Why it works

    Create an alumni seat: when a partner-lab member leaves, they keep a personal Dalton workspace for a year and can invite their new lab with one link to the partner application. New PIs who started on Dalton get priority onboarding. Ask leaving members for a short quote about what the lab brain saved them.

    People who have used a tool daily recommend it from experience, and new PIs choose tools for their first lab exactly when they arrive.

    Risks and how to handle them

    • Lab data leaves with the student

      The alumni workspace starts empty; lab records stay with the lab.

    Builds on

    Evidence

    • HiringFounding AI Engineer (San Francisco, New York or remote US; $120K-$250K and 0.50%-5.00% equity; the only open role on the Y Combinator jobs page): "90% of experiments never make it to publication"; the company structures that data into a queryable "lab brain" and is working with its first cohort of university lab partners; the role builds data capture, structured records and querying, vision for handwritten notebooks and drawn structures, speech-to-text at the bench, lab-history reasoning agents and models that suggest process conditions, working directly with research partners.Source
    • App Store83 Sciences Companion (iOS, "Voice notes for your lab"): records experiment sessions of up to one hour plus photos and typed notes, with transcripts and suggested titles and notebook destinations; requires an existing account and lab membership; no ratings yet; updated November 4, 2026; the seller is shown as an individual developer rather than 83 Sciences, Inc.; its website link is 83sciences.ai/platform.Source
  • 14

    Industry revenue comes from discovery contracts, with examples such as PFAS replacements and a silicon anode binder compatible with existing battery manufacturing. Rivals like Citrine, Uncountable and Albert Invent sell platform software to industrial R&D teams. 83 Sciences can sell an outcome instead: candidate materials for a defined problem. The site has no page for industrial buyers.

    High impactMedium effort First result in 1 quarter CEO and Co Founder

    First action

    Write the standard scope and IP terms with counsel.

    How to measure it

    Industry discovery inquiries

    Now
    –
    Target
    6 in the quarter
    By
    1 quarter

    Steps

    1. 1

      Define the offer

      Scope, deliverables and timeline.

      CEO and Co-Founder 1 week

    2. 2

      Set terms

      IP, data use and fees, with counsel.

      CEO and Co-Founder 2 weeks

    3. 3

      Write the page

      Offer, example problems, process.

      CEO and Co-Founder 3 days

    4. 4

      Add the form

      Problem, field, timeline, contact.

      CTO and Co-Founder 1 day

    5. 5

      Outreach

      R&D leads in the top two verticals.

      CEO and Co-Founder ongoing

    6. 6

      Review pricing

      After the first three projects.

      COO and Co-Founder quarterly

    First result
    1 quarter
    Cost
    Legal review
    Owner
    CEO and Co Founder

    Smallest test

    Send the one-page offer to five R&D leads.

    Why it works

    Write a "Discovery projects" page with one standard offer: the problem brief, the data used (partner-lab history plus the client's own), deliverables (ranked candidates, synthesis conditions, validation data), a 6-12 week timeline, a fixed fee with a success component, and plain IP terms. Show the two example problems. Add a short project inquiry form.

    Industrial R&D managers buy defined projects more easily than open-ended collaborations, and outcomes set the company apart from software rivals.

    Risks and how to handle them

    • Partner labs worry about their data in industry work

      Agree data-use and revenue-share terms with partner labs first.

    Builds on

    Evidence

    • Y Combinator profile"AI-native materials discovery powered by unpublished experimental data"; Summer 2026 batch, team of 3, San Francisco; target users are industrial R&D teams and academic labs in batteries, semiconductors, membranes, catalysts and pharma; example applications include PFAS replacements and a silicon anode binder compatible with existing battery manufacturing; says it has already discovered a novel material and drafted a manuscript; the CEO was previously an MIT fellow focused on commercializing IP.Source
    • Careers listingThird-party careers listing: "More than 85%" of experimental data is discarded, worth over $100B in R&D value a year; under six weeks from a partner's data to the first new material discovered; tools capture voice notes and instrument output and keep a record of every experiment, including failed runs; discovery contracts with industry, publications and commercialization with academia; priority verticals: critical minerals and metals, energy storage, catalysis and chemicals, life sciences solid forms, semiconductors.Source
    • Competitor researchRivals' own homepages: Citrine Informatics, "AI platform for materials and chemicals development"; Uncountable, "The AI Platform for end-to-end product development", with customers such as Dow and Syngenta; Albert Invent, "Invent faster with AI that knows your chemistry"; Benchling, "Your infinite program team for R&D", citing 1,300+ biotechs and with a For Academics page. All sell software to industrial R&D teams.Source
  • 15

    The homepage names focus areas in one line each and has no pages behind them. Industrial researchers search for problems ("silicon anode binder", "PFAS-free alternatives", "rare-earth-free magnets"), not for "novel materials". The team already has relevant work: a permanent-magnet paper and examples in battery binders and PFAS replacements.

    Medium impactMedium effort First result in 2 months CTO and Co Founder

    First action

    Choose the two verticals from the agreed list.

    How to measure it

    Inquiries from vertical pages

    Now
    –
    Target
    5 in the quarter
    By
    2 quarters

    Steps

    1. 1

      Choose two

      From the agreed vertical list.

      CEO and Co-Founder 1 hour

    2. 2

      List problems

      Three per vertical, in buyer language.

      CTO and Co-Founder 1 day

    3. 3

      Write pages

      Problem, data, proof, offer.

      CTO and Co-Founder 1 week

    4. 4

      Link papers

      Relevant publications on each page.

      CTO and Co-Founder 1 hour

    5. 5

      Publish

      With titles and descriptions.

      CTO and Co-Founder 1 day

    6. 6

      Review

      Search queries in Search Console.

      COO and Co-Founder monthly

    First result
    2 months
    Cost
    Founder time only
    Owner
    CTO and Co Founder

    Smallest test

    Publish one vertical page and watch queries.

    Why it works

    Build two pages, for example critical minerals and magnets, and energy storage. Each covers the industry problem, the kinds of unpublished lab data that help, the team's relevant papers, example targets, and links to both the partner and discovery-project offers. Add more pages only after these show traffic.

    Problem-specific pages match what R&D buyers search for, and they give partner labs in those fields a reason to apply.

    Risks and how to handle them

    • Revealing research direction

      Describe problem classes, not specific compositions.

    Builds on

    Evidence

    • Your websiteHomepage (checked November 10, 2026): headline "We are creating the materials of the future"; four focus areas: magnetic materials, critical mineral extraction materials, semiconductor packaging materials, and "Materials science RL environments" "With real experimental data"; navigation is Focus, Team & Open Roles and Publications; the only contact options are two email links; "Backed by Y Combinator"; no link to the /platform page, no customer or partner logos and no request-access button.Source
    • Your websitePublications page: seven papers, six in the Journal of the American Chemical Society (2022-2026), including "MnBi2 Is a Permanent Magnet" (2025) and "Crystal Structure Determination from Powder Diffraction Patterns with Generative Machine Learning" (2024), plus a 2026 ChemRxiv preprint on an agentic workflow for high-throughput synthesis; titles and DOI links only, no summaries.Source
    • Y Combinator profile"AI-native materials discovery powered by unpublished experimental data"; Summer 2026 batch, team of 3, San Francisco; target users are industrial R&D teams and academic labs in batteries, semiconductors, membranes, catalysts and pharma; example applications include PFAS replacements and a silicon anode binder compatible with existing battery manufacturing; says it has already discovered a novel material and drafted a manuscript; the CEO was previously an MIT fellow focused on commercializing IP.Source
  • 16

    83 Sciences offers academic labs publications and commercialization, and its CEO previously worked on commercializing IP as an MIT fellow. University tech-transfer offices are paid to find licensable inventions, and unpublished results are exactly what they never see. One tech-transfer office can introduce the company to many labs and handle the agreements that slow single-lab deals.

    High impactMedium effort First result in 2 quarters CEO and Co Founder

    First action

    Draft the one-page briefing and a standard disclosure flow.

    How to measure it

    Master agreements signed

    Now
    –
    Target
    2
    By
    2 quarters

    Steps

    1. 1

      Write the briefing

      Data, disclosure, IP and revenue share.

      CEO and Co-Founder 1 week

    2. 2

      Draft the agreement

      A master template with counsel.

      CEO and Co-Founder 3 weeks

    3. 3

      Pick universities

      Where partner labs and advisors are.

      COO and Co-Founder 1 day

    4. 4

      Get introductions

      Through PIs and advisors.

      CEO and Co-Founder 1 month

    5. 5

      Meet

      Present to each office.

      CEO and Co-Founder 1 month

    6. 6

      Sign and announce

      Tell that university's labs.

      CEO and Co-Founder ongoing

    First result
    2 quarters
    Cost
    Legal fees
    Owner
    CEO and Co Founder

    Smallest test

    Brief one tech-transfer office informally.

    Why it works

    Prepare a tech-transfer briefing: how partner-lab data is structured, how discoveries are disclosed to the university first, how IP and revenue are shared, and the standard data agreement. Approach the tech-transfer offices at the universities where partner labs and advisors already sit, and offer a master agreement that new labs at that university can join.

    A master agreement removes weeks of contract work per lab, and the tech-transfer office becomes a channel instead of a hurdle.

    Risks and how to handle them

    • Slow university processes

      Run single-lab agreements in parallel.

    Builds on

    Evidence

    • Y Combinator profile"AI-native materials discovery powered by unpublished experimental data"; Summer 2026 batch, team of 3, San Francisco; target users are industrial R&D teams and academic labs in batteries, semiconductors, membranes, catalysts and pharma; example applications include PFAS replacements and a silicon anode binder compatible with existing battery manufacturing; says it has already discovered a novel material and drafted a manuscript; the CEO was previously an MIT fellow focused on commercializing IP.Source
    • Careers listingThird-party careers listing: "More than 85%" of experimental data is discarded, worth over $100B in R&D value a year; under six weeks from a partner's data to the first new material discovered; tools capture voice notes and instrument output and keep a record of every experiment, including failed runs; discovery contracts with industry, publications and commercialization with academia; priority verticals: critical minerals and metals, energy storage, catalysis and chemicals, life sciences solid forms, semiconductors.Source
    • Your websiteTeam page: three co-founders (a CEO from catalysis and battery-materials work at Stanford and Sila Nanotechnologies, a CTO in quantum materials and generative models for chemical experimental data, a COO from robotics and biosignaling research); advisors from UC Santa Barbara (a CNSI director), MIT and a software company; no roles on the page itself, only a link to the Y Combinator jobs page.Source
  • 17

    The CTO has published repeatedly in JACS, including a generative method for crystal structures from powder diffraction and a 2026 preprint on agentic high-throughput synthesis. The domain is months old, with no blog and no social links. Conference talks reach PIs and industrial scientists where they already gather, from a speaker with standing.

    Medium impactMedium effort First result in 2 quarters CTO and Co Founder

    First action

    List upcoming meeting abstract deadlines in the priority fields.

    How to measure it

    Partner calls booked per talk

    Now
    –
    Target
    5
    By
    after each talk

    Steps

    1. 1

      List meetings

      Abstract deadlines in priority fields.

      COO and Co-Founder half a day

    2. 2

      Write abstracts

      Failed-experiment learning.

      CTO and Co-Founder 2 days

    3. 3

      Build the talk

      Results and a partner QR code.

      CTO and Co-Founder 1 week

    4. 4

      Book meetings

      Ask PIs in the audience for a call.

      CEO and Co-Founder at the event

    5. 5

      Follow up

      Within a week.

      CEO and Co-Founder 1 week

    6. 6

      Post the slides

      On the site with a summary.

      CTO and Co-Founder 1 day

    First result
    2 quarters
    Cost
    Travel and registration
    Owner
    CTO and Co Founder

    Smallest test

    Give one department seminar this month.

    Why it works

    Submit talks or posters to the next materials and chemistry society meetings on learning from unpublished and failed experiments, using partner-lab results where allowed. End every talk with a QR code to /partners. Hold short follow-up calls with attendees in the following week.

    Scientists trust peers presenting data. One talk can produce several partner conversations.

    Risks and how to handle them

    • Long abstract lead times

      Start with department seminars and webinars meanwhile.

    Evidence

    • Your websitePublications page: seven papers, six in the Journal of the American Chemical Society (2022-2026), including "MnBi2 Is a Permanent Magnet" (2025) and "Crystal Structure Determination from Powder Diffraction Patterns with Generative Machine Learning" (2024), plus a 2026 ChemRxiv preprint on an agentic workflow for high-throughput synthesis; titles and DOI links only, no summaries.Source
    • Your websiteHomepage metadata: title "83 Sciences | Novel Materials"; meta description "83 Sciences develops magnetic materials, critical mineral extraction materials, semiconductor packaging materials, and materials science reinforcement learning environments"; no social share image; no blog, pricing page or social profile links; domain registered June 24, 2026.Source
  • 18

    Partner labs hand over notebooks, audio recordings and photos; the Companion app's privacy label lists audio, photos and other user content linked to the user. The open engineering role works directly with research partners on data capture and querying. The site has a privacy policy but nothing that tells a PI, a university IT team or an industrial R&D lead who owns the data, how it is used and how to get it back.

    Medium impactMedium effort First result in 1 month COO and Co Founder

    First action

    Write down the current answers to the six questions.

    How to measure it

    Median days to signed agreement

    Now
    –
    Target
    -30%
    By
    2 quarters

    Steps

    1. 1

      Answer six questions

      Ownership, storage, access, use, credit, deletion.

      COO and Co-Founder 1 day

    2. 2

      Check with counsel

      Match the agreements.

      CEO and Co-Founder 1 week

    3. 3

      Write the page

      Plain language with a short FAQ.

      COO and Co-Founder 2 days

    4. 4

      Link it

      /partners, /platform and the app listing.

      CTO and Co-Founder 1 hour

    5. 5

      Prepare a questionnaire pack

      Answers for university and company security reviews.

      CTO and Co-Founder 1 week

    6. 6

      Update

      When practices change.

      COO and Co-Founder quarterly

    First result
    1 month
    Cost
    Legal review
    Owner
    COO and Co Founder

    Smallest test

    Add a five-question FAQ to /partners.

    Why it works

    Write a plain-language page: the lab owns its data; how it is stored and encrypted; who at 83 Sciences can see it; whether it is used beyond the lab's own workspace (and only with consent); how discoveries and credit are shared; export and deletion. Link it from /partners, the platform page and the app listing.

    Data questions are the first objection from labs and the first procurement step for companies. Answering them up front shortens every deal.

    Risks and how to handle them

    • Promises the team cannot keep yet

      Describe only current practice and date the page.

    Builds on

    Evidence

    • App Store83 Sciences Companion (iOS, "Voice notes for your lab"): records experiment sessions of up to one hour plus photos and typed notes, with transcripts and suggested titles and notebook destinations; requires an existing account and lab membership; no ratings yet; updated November 4, 2026; the seller is shown as an individual developer rather than 83 Sciences, Inc.; its website link is 83sciences.ai/platform.Source
    • HiringFounding AI Engineer (San Francisco, New York or remote US; $120K-$250K and 0.50%-5.00% equity; the only open role on the Y Combinator jobs page): "90% of experiments never make it to publication"; the company structures that data into a queryable "lab brain" and is working with its first cohort of university lab partners; the role builds data capture, structured records and querying, vision for handwritten notebooks and drawn structures, speech-to-text at the bench, lab-history reasoning agents and models that suggest process conditions, working directly with research partners.Source

Big bets 4

Larger plays that can change the trajectory.

  • 19

    Public materials datasets are built from published, successful results. 83 Sciences' edge is the opposite: unpublished experiments, including failed runs. The team has already shown it can publish methods (the powder-diffraction paper) and lists "real experimental data" as a focus. A small public benchmark of real failed and successful syntheses, released with partner-lab consent and credit, would make the company the reference for this kind of data.

    High impactHigh effort First result in 2 quarters CTO and Co Founder

    First action

    Ask the first partner labs which subset they would release.

    How to measure it

    Downloads in the first six months

    Now
    –
    Target
    1,000
    By
    2 quarters after release

    Steps

    1. 1

      Get consent

      Two or three labs, with co-authorship.

      CEO and Co-Founder 1 month

    2. 2

      Choose the subset

      Non-sensitive syntheses with outcomes.

      CTO and Co-Founder 2 weeks

    3. 3

      Curate

      Clean, label and document.

      Founding AI Engineer 1 month

    4. 4

      Define the task

      Prediction target and baseline.

      CTO and Co-Founder 2 weeks

    5. 5

      Write the data paper

      With partner co-authors.

      CTO and Co-Founder 1 month

    6. 6

      Release

      Public repository with a license and a partner link.

      Founding AI Engineer 1 week

    7. 7

      Track

      Downloads, citations and inbound.

      COO and Co-Founder monthly

    First result
    2 quarters
    Cost
    Two engineer-months
    Owner
    CTO and Co Founder

    Smallest test

    Release 200 failed syntheses with outcomes from one lab.

    Why it works

    With two or three partner labs, release a curated dataset of a few thousand synthesis attempts with outcomes, including failures, plus a defined prediction task and baseline. Publish it with a data paper, credit the labs as co-authors, and keep the full corpus private. Use citations and downloads as a lead source for labs and R&D teams.

    A benchmark puts the company's name in every paper that uses it, gives partner labs citations, and shows buyers how different this data is.

    Risks and how to handle them

    • Giving away the moat

      Release a small subset; keep the full corpus and the live data private.

    Builds on

    Evidence

    • Your websiteHomepage (checked November 10, 2026): headline "We are creating the materials of the future"; four focus areas: magnetic materials, critical mineral extraction materials, semiconductor packaging materials, and "Materials science RL environments" "With real experimental data"; navigation is Focus, Team & Open Roles and Publications; the only contact options are two email links; "Backed by Y Combinator"; no link to the /platform page, no customer or partner logos and no request-access button.Source
    • Your websitePublications page: seven papers, six in the Journal of the American Chemical Society (2022-2026), including "MnBi2 Is a Permanent Magnet" (2025) and "Crystal Structure Determination from Powder Diffraction Patterns with Generative Machine Learning" (2024), plus a 2026 ChemRxiv preprint on an agentic workflow for high-throughput synthesis; titles and DOI links only, no summaries.Source
    • Careers listingThird-party careers listing: "More than 85%" of experimental data is discarded, worth over $100B in R&D value a year; under six weeks from a partner's data to the first new material discovered; tools capture voice notes and instrument output and keep a record of every experiment, including failed runs; discovery contracts with industry, publications and commercialization with academia; priority verticals: critical minerals and metals, energy storage, catalysis and chemicals, life sciences solid forms, semiconductors.Source
  • 20

    "Materials science RL environments, with real experimental data" is one of four homepage focus areas and appears in the meta description, but the company never explains it: who buys it, what an environment contains or how to get one. The buyers (teams building models for science) are different from PIs and industrial R&D leads and need a different page and offer.

    High impactHigh effort First result in 2 quarters CEO and Co Founder

    First action

    Decide: launch it with a page, or remove it from the homepage.

    How to measure it

    Paid pilots signed

    Now
    –
    Target
    2
    By
    2 quarters

    Steps

    1. 1

      Decide

      Launch now or park it.

      CEO and Co-Founder 1 week

    2. 2

      Check data rights

      Partner-lab consent for this use.

      CEO and Co-Founder 1 month

    3. 3

      Define the product

      Tasks, scoring and delivery.

      CTO and Co-Founder 2 weeks

    4. 4

      Write the page

      What, who, rights, request access.

      CEO and Co-Founder 1 week

    5. 5

      Run pilots

      Two paid pilots.

      Founding AI Engineer 1 quarter

    6. 6

      Price it

      From pilot results.

      CEO and Co-Founder 1 month

    First result
    2 quarters
    Cost
    Engineering and legal time
    Owner
    CEO and Co Founder

    Smallest test

    Add a one-paragraph explanation and a request link under the focus area.

    Why it works

    Decide whether this is a real product line now. If yes, write a page that explains what an environment is (tasks, real outcomes from lab data, scoring), who it is for, how data rights are handled with partner labs and how to request access, and run two paid pilots. If not, take it off the homepage until it is ready.

    A clear offer to a distinct buyer can create a second revenue stream from the same data, and a focus area with no explanation confuses the other audiences.

    Risks and how to handle them

    • Partner labs object to their data being sold

      Opt-in only, with revenue share and clear terms.

    Builds on

    Evidence

    • Your websiteHomepage (checked November 10, 2026): headline "We are creating the materials of the future"; four focus areas: magnetic materials, critical mineral extraction materials, semiconductor packaging materials, and "Materials science RL environments" "With real experimental data"; navigation is Focus, Team & Open Roles and Publications; the only contact options are two email links; "Backed by Y Combinator"; no link to the /platform page, no customer or partner logos and no request-access button.Source
    • Your websiteHomepage metadata: title "83 Sciences | Novel Materials"; meta description "83 Sciences develops magnetic materials, critical mineral extraction materials, semiconductor packaging materials, and materials science reinforcement learning environments"; no social share image; no blog, pricing page or social profile links; domain registered June 24, 2026.Source
  • 21

    Each lab brain holds a record of failed runs. Labs working on similar chemistry repeat each other's failures because negative results are rarely published. If partner labs opt in to share anonymized failures (conditions and outcome only), every new lab joins a growing body of "this didn't work" knowledge that no single lab or rival has.

    High impactHigh effort First result in 3 quarters CTO and Co Founder

    First action

    Ask two partner labs in the same field whether they would share failures on these terms.

    How to measure it

    Contributing partner labs

    Now
    –
    Target
    5
    By
    3 quarters

    Steps

    1. 1

      Test the appetite

      Interview PIs on sharing terms.

      CEO and Co-Founder 1 month

    2. 2

      Design privacy

      What is shared and what never is.

      CTO and Co-Founder 1 month

    3. 3

      Agree terms

      Consent, credit and withdrawal.

      CEO and Co-Founder 1 month

    4. 4

      Build the match

      Similar-attempt search across contributed records.

      Founding AI Engineer 1 quarter

    5. 5

      Pilot with two labs

      Same field.

      COO and Co-Founder 1 quarter

    6. 6

      Measure

      Experiments changed or avoided.

      COO and Co-Founder monthly

    First result
    3 quarters
    Cost
    Engineering quarter plus legal
    Owner
    CTO and Co Founder

    Smallest test

    Manually compare failures between two consenting labs and report overlaps.

    Why it works

    Design an opt-in layer: when a researcher plans an experiment, Dalton can say "three similar attempts in partner labs failed under these conditions" without revealing the lab or unpublished results. Labs that contribute get access; contributions are credited. Start with two labs in the same field.

    It creates a network effect: each new lab makes the product more useful for every other lab, which is hard for a single-tenant platform to copy.

    Risks and how to handle them

    • Labs fear being scooped

      Share only failed conditions, never successes or compositions under study, and allow withdrawal.

    Builds on

    Evidence

    • Careers listingThird-party careers listing: "More than 85%" of experimental data is discarded, worth over $100B in R&D value a year; under six weeks from a partner's data to the first new material discovered; tools capture voice notes and instrument output and keep a record of every experiment, including failed runs; discovery contracts with industry, publications and commercialization with academia; priority verticals: critical minerals and metals, energy storage, catalysis and chemicals, life sciences solid forms, semiconductors.Source
    • Competitor researchRivals' own homepages: Citrine Informatics, "AI platform for materials and chemicals development"; Uncountable, "The AI Platform for end-to-end product development", with customers such as Dow and Syngenta; Albert Invent, "Invent faster with AI that knows your chemistry"; Benchling, "Your infinite program team for R&D", citing 1,300+ biotechs and with a For Academics page. All sell software to industrial R&D teams.Source
  • 22

    The tools already capture voice notes and instrument output, and the open engineering role builds the data capture and structuring. If instrument files are still moved by hand, records go missing. The most valuable records (diffraction, spectroscopy) would be captured more completely if the instrument's output landed in Dalton automatically, linked to the voice note from the same session.

    Medium impactHigh effort First result in 2 quarters Founding AI Engineer

    First action

    Count which instruments generate most of the current partner data.

    How to measure it

    Automatically captured instrument files

    Now
    –
    Target
    70%
    By
    3 quarters

    Steps

    1. 1

      Rank instruments

      By volume in partner labs.

      Founding AI Engineer 1 week

    2. 2

      Build the agent

      Watched folder to Dalton.

      Founding AI Engineer 1 month

    3. 3

      Link sessions

      Match files to voice notes by time and user.

      Founding AI Engineer 1 month

    4. 4

      Pilot

      Two partner labs.

      COO and Co-Founder 1 month

    5. 5

      Approach vendors

      Instrument and lab-data software makers.

      CEO and Co-Founder 1 quarter

    6. 6

      Publish the list

      Supported instruments on /platform.

      CTO and Co-Founder 1 day

    First result
    2 quarters
    Cost
    Engineering quarter
    Owner
    Founding AI Engineer

    Smallest test

    Run the folder agent on one diffractometer PC for a month.

    Why it works

    Start with the instrument types partner labs use most, such as X-ray diffraction. Build a watched-folder agent that uploads new output and links it to the session, then approach instrument makers and lab-data software vendors about listed integrations. Publish the supported instruments on /platform.

    Automatic capture removes the main reason lab data goes missing, and listed integrations make the platform discoverable to those vendors' users.

    Risks and how to handle them

    • University IT blocks agents on instrument PCs

      Offer a manual-upload fallback and an IT information sheet.

    Builds on

    Evidence

    • Careers listingThird-party careers listing: "More than 85%" of experimental data is discarded, worth over $100B in R&D value a year; under six weeks from a partner's data to the first new material discovered; tools capture voice notes and instrument output and keep a record of every experiment, including failed runs; discovery contracts with industry, publications and commercialization with academia; priority verticals: critical minerals and metals, energy storage, catalysis and chemicals, life sciences solid forms, semiconductors.Source
    • HiringFounding AI Engineer (San Francisco, New York or remote US; $120K-$250K and 0.50%-5.00% equity; the only open role on the Y Combinator jobs page): "90% of experiments never make it to publication"; the company structures that data into a queryable "lab brain" and is working with its first cohort of university lab partners; the role builds data capture, structured records and querying, vision for handwritten notebooks and drawn structures, speech-to-text at the bench, lab-history reasoning agents and models that suggest process conditions, working directly with research partners.Source

Wildcards 2

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

  • 23

    The company's whole idea is that failed experiments are valuable, yet it has no blog and no social profile links on its site. The founders have years of their own lab work behind six JACS papers. Short posts on what went wrong, and what it taught, would show the thesis instead of stating it, and they are the kind of honest science content researchers share.

    Medium impactLow effort First result in 1 month CEO and Co Founder

    First action

    Write the first post from a founder's own failed synthesis.

    How to measure it

    Applications attributed to the series

    Now
    –
    Target
    5
    By
    2 quarters

    Steps

    1. 1

      List ten failures

      From founders' own past work.

      CTO and Co-Founder half a day

    2. 2

      Check permissions

      Former labs where needed.

      CEO and Co-Founder 1 week

    3. 3

      Set up profiles

      Company LinkedIn page and a site section.

      COO and Co-Founder 1 day

    4. 4

      Write a template

      Tried, happened, learned, how to catch it.

      CEO and Co-Founder 1 hour

    5. 5

      Post weekly

      Rotate founders.

      CEO and Co-Founder weekly

    6. 6

      Invite partners

      Guest posts from partner labs.

      COO and Co-Founder after 8 weeks

    First result
    1 month
    Cost
    $0
    Owner
    CEO and Co Founder

    Smallest test

    Post one failure this week and watch the response.

    Why it works

    Each week, one founder posts a real failed experiment from their own past work (with permission from former labs where needed): what was tried, what happened, what was learned and how a lab brain would have flagged it. Post on LinkedIn and the site, and end with the /partners link. Invite partner-lab members to contribute their own later.

    It is cheap, specific and memorable, it builds a following among the exact audience, and it makes "unpublished data" concrete.

    Risks and how to handle them

    • Disclosing others' unpublished work

      Use only the founders' own work, with permission.

    Evidence

    • Your websiteHomepage metadata: title "83 Sciences | Novel Materials"; meta description "83 Sciences develops magnetic materials, critical mineral extraction materials, semiconductor packaging materials, and materials science reinforcement learning environments"; no social share image; no blog, pricing page or social profile links; domain registered June 24, 2026.Source
    • Your websitePublications page: seven papers, six in the Journal of the American Chemical Society (2022-2026), including "MnBi2 Is a Permanent Magnet" (2025) and "Crystal Structure Determination from Powder Diffraction Patterns with Generative Machine Learning" (2024), plus a 2026 ChemRxiv preprint on an agentic workflow for high-throughput synthesis; titles and DOI links only, no summaries.Source
  • 24

    The CTO co-led a JACS paper on determining crystal structures from powder diffraction patterns with generative machine learning. Nearly every materials lab has old diffraction files it never fully analysed. A free web demo that takes one pattern and suggests candidate structures would bring exactly the right researchers to the site with their own data in hand.

    Medium impactMedium effort First result in 2 months Founding AI Engineer

    First action

    Check whether the published method can run as a lightweight hosted demo.

    How to measure it

    Applications from demo users

    Now
    –
    Target
    5%
    By
    2 quarters

    Steps

    1. 1

      Check feasibility

      Runtime and hosting cost per pattern.

      CTO and Co-Founder 1 week

    2. 2

      Build the page

      Upload, result, paper link.

      Founding AI Engineer 3 weeks

    3. 3

      Set data rules

      Delete by default, opt-in to keep.

      COO and Co-Founder 1 day

    4. 4

      Add the next step

      "Bring your whole archive" link to /partners.

      CEO and Co-Founder 1 hour

    5. 5

      Launch

      Share in crystallography and materials communities.

      CTO and Co-Founder 1 week

    6. 6

      Measure

      Uploads and applications.

      COO and Co-Founder monthly

    First result
    2 months
    Cost
    Hosting and one engineer-month
    Owner
    Founding AI Engineer

    Smallest test

    Run ten patterns from volunteers by email before building the page.

    Why it works

    Build a simple page: upload one powder diffraction file, get candidate structures with a confidence score and a link to the paper. Files are processed and not kept unless the user agrees. After the result, invite them to put the lab's whole diffraction archive into Dalton through the partner path.

    A useful free tool on the user's own data is the strongest lead magnet for scientists, and it shows the team's published capability.

    Risks and how to handle them

    • Weak results on messy patterns

      Show confidence and say when a pattern is out of scope.

    Builds on

    Evidence

    • Your websitePublications page: seven papers, six in the Journal of the American Chemical Society (2022-2026), including "MnBi2 Is a Permanent Magnet" (2025) and "Crystal Structure Determination from Powder Diffraction Patterns with Generative Machine Learning" (2024), plus a 2026 ChemRxiv preprint on an agentic workflow for high-throughput synthesis; titles and DOI links only, no summaries.Source
    • HiringFounding AI Engineer (San Francisco, New York or remote US; $120K-$250K and 0.50%-5.00% equity; the only open role on the Y Combinator jobs page): "90% of experiments never make it to publication"; the company structures that data into a queryable "lab brain" and is working with its first cohort of university lab partners; the role builds data capture, structured records and querying, vision for handwritten notebooks and drawn structures, speech-to-text at the bench, lab-history reasoning agents and models that suggest process conditions, working directly with research partners.Source
More example playbooks and reviews
ProgrammableDigital

Show 83 Sciences’s product working

Turn screenshots or the website into an animated product demo in its own brand, for the homepage, launch posts and outreach in this playbook.

Make a demo (opens programmable.digital in a new tab)

More growth playbooks

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

  • Revin growth strategyAI voice, SMS and email agents for home-service call centres: inbound answering, speed-to-lead, scheduling and confirmations…21 growth ideas
  • Glia growth strategyA banking AI platform for voice and digital service: AI agents (Glia Banker, CoPilot, Analyst, Outreach), contact center and digital…21 growth ideas
  • RevPack growth strategyA Warsaw-based RevOps and GTM engineering agency that connects a company's CRM, sales and marketing into one revenue system: HubSpot…23 growth ideas
  • Maze growth strategyA user and market research platform: participant recruitment from its own panel, unmoderated tests, surveys, moderated and AI-moderated…21 growth ideas
  • WorkStore growth strategyWork Store Limited, a Mumbai B2B supplier (since 2007) of workplace and facility supplies online: cleaning equipment, Dyson hand dryers…21 growth ideas
  • i-Genie.ai growth strategyAn AI consumer intelligence platform that turns searches, social and video posts and eCommerce reviews into brand equity tracking (Brand…22 growth ideas
  • AMAROK growth strategyCommercial perimeter security as a subscription: The Electric Guard Dog solar-powered electric fence plus video surveillance, monitoring…20 growth ideas
  • Asendia AI growth strategyAI recruiter agents for staffing agencies. "Sarah" sources, calls, texts, emails and WhatsApps candidates, runs phone and video screens…23 growth ideas

See every playbook and website review

Independent research. Programmable wrote this playbook from public sources, as they stood on November 10, 2026, and cites them throughout. 83 Sciences did not commission, review or endorse it, and Programmable is not affiliated with 83 Sciences. 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 83 Sciences runs its business.

Something wrong, or do you represent 83 Sciences? Email [email protected] and we will correct it or take it down.

30/60/90 roadmap

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

  1. Days 1 to 307 ideas

    Make the site match the product: the lab-partner path, metadata and share image, a native request-access form, the app under the company name, one stat and one vertical list, publications as proof, and small speed and accessibility fixes.

  2. Days 31 to 6013 ideas

    Build the partner engine: the first discovery story, advisor and tech-transfer channels, the one-notebook pilot, onboarding milestones, the PI digest, alumni referrals, the discovery-contract page, vertical pages, conference talks and a data-ownership page; start the two wildcards.

  3. Days 61 to 904 ideas

    Commit to the big bets: a public dark-data benchmark, the RL-environment product line, a cross-lab failed-experiment network and instrument integrations.

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 getting value from years of unpublished and failed experiments. The most cited problem: lab knowledge lost when students and postdocs leave.

  • Capturing notes by voice at the bench instead of typing
  • Handwritten notebooks and scattered instrument files no one can search
  • In their words: “lab brain”

Market

Competitive landscape

4 competitors compared: Citrine Informatics, Uncountable, Albert Invent and Benchling.

  • Gap to exploit: Unpublished academic lab data as a discovery source
  • Citrine Informatics: Software for industrial teams, not built around academic lab history
  • Uncountable: Broad scope across QC, regulatory and manufacturing rather than discovery

Hiring signals

83 Sciences has one open role, a Founding AI Engineer (San Francisco, New York or remote US).

  • Founding AI Engineer: Lab brain capture and query for a first cohort of university lab partners. Ideas 1, 10, 12, 22, 24.

From the team that wrote this playbook

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

© 2026 AI-Powered Growth Intelligence