PEER · SOURCED PUBLIC FACTS

OpenAI

Sourced public facts, grouped by the twelve letters, and a machine-check FOUND score. Facts are what the source says; the score is what our code could see.

Just the domain. No email, no login, usually under a minute.

FOUND Score v1.0 · read 19 Oct 2026 · Low confidence
75of 100

Low confidence: 41% of the points rest on evidence we read ourselves (the site, listing pages, awarders' lists, filings, X data, answer samples); the rest was found by live search or estimated, and is labelled so.

No earlier read to compare with yet. Read from public evidence; not a judgement of the company.

Last refreshed 8 Nov 2026. Stale after 14 days.

Size: not read yet (no filer id held.)

Site: openai.com. Snapshots stored: 22. Nothing is overwritten; earlier snapshots stay.

FOUND machine-check score, letter by letter

Rank 12 of 21 · peers’ median 76.5 of 100

OpenAI ranks 12 of 21 scored companies. The peers' median is 76.5 and their range 61 to 87, all of 100. Not ranked, because no score is stored: Zoho. Peers are read at different times; each read date is on the peers table.

What OpenAI does well

No machine check fully passed in this read.

Gaps

No gap among the checks that ran.

Fix list for OpenAI: the 4 biggest gaps first

The parts of the score furthest below what the method lets them reach, worked out from the evidence shown on this page. Each says what to do and why it moves the score. “Up to” is the most the overall score could rise if that part reached full; it is not a forecast.

  1. Security and trust published · Distinctions

    12.5 of 25 points now · up to +3 overall · partly estimated, not read on this run

    Do this: Publish /.well-known/security.txt, a trust or security page, and name your certifications (for example SOC 2 or ISO 27001) on it.

    Why it moves the score: Security and trust published is worth 25 points of Distinctions, read on your site; each check that passes adds its share.

    Now: The site could not be read for security.txt, trust page and named certifications, so it counts as average (half), labelled.

    Draft it for me

    Draft: review and edit before you publish. Built from what we scanned, with each fact linked to where we read it; [square brackets] mark what only you know, and [needs proof] marks a claim you must back up before you publish it. Nothing is published or sent: you copy, edit and publish it yourself. We never write reviews or testimonials.

    • security.txt file
    • Trust and security page

    Loading the drafts…

  2. Content worth citing · Underdog Moves

    12.5 of 25 points now · up to +3 overall · partly estimated, not read on this run

    Do this: Add the kinds of pages others cite that you do not have yet: a blog or newsroom, customer stories, a partners page, docs or an academy, events or a community page.

    Why it moves the score: Content worth citing is worth 25 points of Underdog Moves: 5 points for each of the 5 kinds found on your site.

    Now: The site’s pages were closed to our reader, so content counts as average (half), labelled.

    Draft it for me

    Draft: review and edit before you publish. Built from what we scanned, with each fact linked to where we read it; [square brackets] mark what only you know, and [needs proof] marks a claim you must back up before you publish it. Nothing is published or sent: you copy, edit and publish it yourself. We never write reviews or testimonials.

    • Outlines for the kinds of pages others cite
    • Customer story outline

    Loading the drafts…

  3. Pages machines can read and quote · Underdog Moves

    7.5 of 15 points now · up to +2 overall · partly estimated, not read on this run

    Do this: Add Organization and Product structured data (JSON-LD), a clear title and meta description on each key page, and make sure the main text loads without JavaScript.

    Why it moves the score: Pages machines can read and quote are worth 15 points of Underdog Moves, read on your site; each check that passes adds its share.

    Now: The site could not be read for structured data, meta and readable text, so it counts as average (half), labelled.

    Draft it for me

    Draft: review and edit before you publish. Built from what we scanned, with each fact linked to where we read it; [square brackets] mark what only you know, and [needs proof] marks a claim you must back up before you publish it. Nothing is published or sent: you copy, edit and publish it yourself. We never write reviews or testimonials.

    • Organization and Product structured data (JSON-LD, template only)
    • Brand facts sheet (template only, every fact sourced)
    • Title, description and sharing tags

    Loading the drafts…

  4. Evidence reachable · Distinctions

    5 of 10 points now · up to +1 overall · partly estimated, not read on this run

    Do this: Serve every page over HTTPS with redirects from plain HTTP, and make the home page answer quickly.

    Why it moves the score: Evidence reachable is worth 10 points of Distinctions: HTTPS and a quick response.

    Now: The site could not be read for HTTPS and speed, so it counts as average (half), labelled.

    Draft it for me

    Draft: review and edit before you publish. Built from what we scanned, with each fact linked to where we read it; [square brackets] mark what only you know, and [needs proof] marks a claim you must back up before you publish it. Nothing is published or sent: you copy, edit and publish it yourself. We never write reviews or testimonials.

    • HTTPS redirect and speed settings

    Loading the drafts…

Get your improvement plan Free. A person reads this result and replies with a plan built on this list.

30-day plan for OpenAI

Four weekly steps. Your Start here choice comes first when selected; other steps follow points versus effort. Each step names the FOUND letters it moves. Re-score on day 30 to see what changed. Built from the fix list with no AI model. Put a name next to each step.

  1. Week 1 · days 1 to 7

    • Security and trust published · moves D (Distinctions) · low effort · up to +3 overall

      Publish /.well-known/security.txt, a trust or security page, and name your certifications (for example SOC 2 or ISO 27001) on it.

      On your site: you or your web team publish it after review.

      Drafts ready: security.txt file, Trust and security page

  2. Week 2 · days 8 to 14

  3. Week 3 · days 15 to 21

    • Content worth citing · moves U (Underdog Moves) · medium effort · up to +3 overall

      Add the kinds of pages others cite that you do not have yet: a blog or newsroom, customer stories, a partners page, docs or an academy, events or a community page.

      On your site: you or your web team publish it after review.

      Drafts ready: Outlines for the kinds of pages others cite, Customer story outline

  4. Week 4 · days 22 to 29

    • Evidence reachable · moves D (Distinctions) · low effort · up to +1 overall

      Serve every page over HTTPS with redirects from plain HTTP, and make the home page answer quickly.

      On your site: you or your web team publish it after review.

      Drafts ready: HTTPS redirect and speed settings

    • Check what is live

      Open each page and file you changed and check it is live on your site. Write down the date and address of everything published on or off your site, so the re-score can be checked against it.

  5. Day 30: re-score

    Run the FOUND Score again for openai.com and compare each letter. Re-score openai.com

    A score can move for reasons outside this plan, and answer engines change. It is not a forecast.

Optional. The calendar file is the reliable reminder.

Before and after

Looking for an earlier stored score…

The 0 machine checks, as stored

Each check reads only the public home page and a few standard files, once, from one place, at the time shown. A check that could not run says so and is left out, not counted as zero. Read from https://openai.com/, score method found-score-v1.0.

Re-runs the machine checks live. One refresh per company per hour.

Evidence at a glance

Evidence at a glance

The first few sourced facts of each kind. Every one links to the page it came from and the date it was read.

FOUND · O · Opinions

What customers and reviewers say 4 sourced

  • First section headings (customers): Stories source, 4 Nov 2026
  • Description the page gives itself (customers): Explore how our network of customers is using our technology to advance their goals. source, 4 Nov 2026
  • Programme and proof terms in the main content (menus excluded) (customers): training, community source, 4 Nov 2026
Every O fact

FOUND · N · Nods

Analysts, awards and trusted names not collected yet

No analyst, award or partner nod was found on the pages we read.

Every N fact

Press

Press coverage 5 sourced

  • Moneycontrol: OpenAI names Uday Ruddarraju to lead data center, computing unit source, 14 Oct 2026
  • Mint: 'We will have a cool demo in 2027': What Sam Altman said about OpenAI's humanoid robot and the challenges source, 18 Oct 2026
  • CNBC (incl. CNBC-TV18): Eight US states sue OpenAI, claiming ChatGPT harms young users source, 15 Oct 2026
Press and coverage card

Social

Social presence 2 sourced

  • @OpenAI (official): 5,450,664 followers on X source, 19 Oct 2026
  • @sama (CEO): 6,390,564 followers on X source, 19 Oct 2026
Social standing card

Company facts

Size, market and filings not collected yet

No company facts or filings are stored for this company.

Opinions and Nods evidence

Opinions and Nods evidence

Review sites (in the score, Opinions)

Searched 19 Oct 2026, each line tied by Google Search to the review site's own listing page. Our crawler then opens each listing page where the site allows it (robots.txt obeyed, no login, no bot check worked around): a rating read on the page counts in full and replaces the search figure; one only from search counts 0.9 and is marked “from search, page not read”. Business-software sites count fully; consumer sites such as Trustpilot count 0.4. Capterra, GetApp and Software Advice share one pool of reviews and count once. In this read the review part scored 37 of 50.

Review-site listings for OpenAI
SiteListingRatingReviewsAs ofRead on the pageWeightSource
G2ChatGPT4.6 of 53,11019 Oct 2026From search, page not read: The site refused our crawler (HTTP 403); not worked around.Business site, 1G2 listing
Gartner Peer InsightsChat GPT4.4 of 56419 Oct 2026From search, page not read: The site refused our crawler (HTTP 403); not worked around.Business site, 1Gartner Peer Insights listing
CapterraChatGPT4.4 of 540519 Oct 2026From search, page not read: The site refused our crawler (HTTP 403); not worked around.Business site, 1 (Gartner Digital Markets, counts once)Capterra listing
Software AdviceChatGPT4.4 of 540419 Oct 2026From search, page not read: The site refused our crawler (HTTP 403); not worked around.Business site, 1 (Gartner Digital Markets, counts once)Software Advice listing
TrustRadiusChatGPT8.9 of 1074319 Oct 2026From search, page not read: robots.txt answered 403, so the site is treated as closedBusiness site, 1TrustRadius listing
SourceForgeChatGPT4.7 of 51019 Oct 2026Confirmed on the pageBusiness site, 1SourceForge listing
SoftwareReviewsChatGPT8.9 of 1074219 Oct 2026Confirmed on the pageBusiness site, 1SoftwareReviews listing
Trustpilotopenai.com1.3 of 51,26019 Oct 2026From search, page not read: robots.txt does not allow our crawler to read this page.Consumer site, 0.4Trustpilot listing

7 Opinions and 15 Nods evidence items, each with its source, tier and date: see them all on the evidence page.

Press, coverage and social standing

Press, coverage and social standing

Press and coverage

  • 0 press releases or wire copies seen; no release to measure pickup against.
  • 39 earned stories, 12 in tier 1 outlets, 5 executive interviews or op-eds.
  • Release topics: none seen. Sentiment of earned stories (keyword-based): 0 positive, 33 neutral, 6 negative.

Top stories by quality

  1. OpenAI's Altman says AI benefits warrant accepting some risks (CNBC (incl. CNBC-TV18), 5 Oct 2026; quality 8.8 of 10)
  2. Sam Altman says OpenAI won't go public until its AI models are safe (Business Standard, 30 Sept 2026; quality 8.8 of 10)
  3. 'We will have a cool demo in 2027': What Sam Altman said about OpenAI's humanoid robot and the challenges (Mint, 18 Oct 2026; quality 7.5 of 10)

From Bing News RSS, read 19 Oct 2026. Recent items only, as Bing News returns them (mostly the last few weeks); twelve-month counts not read yet. Wire copy, the company’s own releases and sponsored content score 0. Headline, outlet, date and address only; we do not get past paywalls. Method: coverage quality. Not in the score.

Social standing

  • @OpenAI (official account): 5,450,664 followers; no posts returned for the window.
  • @sama (Sam Altman, CEO; executive voice, counts toward O): 6,390,564 followers; no posts returned for the window.
  • Brand chatter (counts toward N): 120,617 posts naming the company in 7 days, 0% share of voice among the peers. Sentiment of 0 sampled third-party posts (keyword-based): 0 positive, 0 neutral, 0 negative.

Calibration anchor: followers (X users lookup) and 7-day mention counts (X recent counts, 2026-10-12 10:00 UTC to 2026-10-19 09:58 UTC) read 19 Oct 2026 15:28 IST. Own posts and third-party sample not read. Share of voice is among the peers only, so it is not given for anchors. From X API v2 through the X connector (public reads, no login), read 19 Oct 2026. Windows: own posts 2026-09-19 to 2026-10-19 (posts returned by X search; may be incomplete where the search index is shorter); mentions 2026-10-12 08:00 UTC to 2026-10-19 07:04 UTC (X recent counts). LinkedIn: Not read yet: LinkedIn follower counts need a login. YouTube: Not read yet: no YouTube channel address held per company. Not in the score.

Awards evidence in the score

Awards evidence in the score

  • Nods, Authority: The award-list check did not run for this read.
  • Opinions, Authenticity: No review-based award list applies.
  • The company’s own recognition pages: Not read for this score.
Every award list read for this company, newest read of each, with its source and date
Award listHow winners are pickedIn the scoreWhat the awarder’s page showsRead on
CRN Partner Program Guide 5-Star, security vendors 2026application-basedCountsNot found on the page read source19 Oct 2026
CRN Partner Program Guide 5-Star, software vendors 2026application-basedCountsNot found on the page read source19 Oct 2026
CRN Partner Program Guide 5-Star, networking vendors 2026application-basedCountsNot found on the page read source19 Oct 2026
CRN Tech Innovator Awards 2026application-basedCountsNot found on the page read source19 Oct 2026
CRN Channel Chiefs 2026editorial selectionNot counted: an award for a personNot found on the page read source19 Oct 2026
G2 Best Software Awards 2026review-basedCountsNot found on the page read source19 Oct 2026
ANA B2 Awards winners 2026entry-basedCountsNot found on the page read source19 Oct 2026
Stevie Awards for Sales & Customer Service winners 2026entry-basedCountsNot found on the page read source19 Oct 2026
CMI Content Marketing Awards winners 2026entry-basedCountsNot found on the page read source19 Oct 2026
SIIA CODiE Awards winners 2026entry-basedCountsNot found on the page read source19 Oct 2026

A listing is read on the awarder’s own page and is not verified by a person, so it counts at most 1 of 5 for Authority (2 for Authenticity, review-based only). “Not found” may mean not seen: most programmes need an entry, and some pages are built by script. Not finding a company never lowers its score. How each programme works is on the awards page; the method is on the score page.

Facts by letter

Facts by letter

Taken from official pages and filings, each with its source and the date it was read. The full tables are on the evidence page.

What one answer engine said

What one answer engine said

Perplexity Sonar, asked category questions that do not name the company. One sample per question, at one time. A peer not named here was not named in this one answer; it is not a ranking and not a verified result. File: 2026-10-19-perplexity.json.

QuestionNamed in the answer
What are the best endpoint security platforms for a mid-size company? Name up to 8 vendors.no
Which vendors lead in zero trust and secure access service edge (SASE) for enterprises? Name up to 8.no
What are the best identity and access management vendors for a growing SaaS company? Name up to 8.no
Which vulnerability and exposure management tools should a security team shortlist? Name up to 8.no
What are the leading cloud observability and monitoring platforms? Name up to 8.no
Which cloud data platforms should a company consider for analytics? Name up to 8.no
What are the best CRM and marketing automation platforms for B2B marketing teams? Name up to 8.no
Which vendors lead in network security firewalls for enterprises? Name up to 8.no
What are the best software team collaboration and work management tools? Name up to 8.no
Which cloud platforms are best for startups and small development teams? Name up to 8.no

Named in 0 of 10 answers. Not every question is about this company’s category, so “no” is expected for most.

Freshness and gaps

Freshness and gaps

Source groupLast refreshedStale afterState
Home page, robots.txt, llms.txt, sitemap, trust page and the FOUND machine checks8 Nov 202614 daysfresh
Winner and list pages of B2B and channel awards, checked for each company's name19 Oct 202660 daysfresh
Partner, marketplace, academy, events, community, analyst, customer and trust pages4 Nov 202614 daysfresh

Sources we could not read at last attempt

See recent changes across all peers.

See how your own site compares with OpenAI.

The same machine checks, on your site, free and with no login.