Ten heavily promoted AI products, and the customer evidence behind them stops at the press release

Asked (summary):

Find examples of goods or services that AI companies have hyped over and over, yet where we never directly hear from customers, or read articles or reviews about them from industry insiders.

This report covers a shortlist of 10 promoted AI products or product families, drawn from 33 candidate source mentions gathered across the web; the 10 primary citations span 9 independent sites, and no URL backs more than one row. For each product the matrix below marks what public proof exists today — a named customer, a verified public deployment, and an independent outside voice — against the vendor's own promoted claim.

Fairness caveat: absence of public customer testimony is not proof that no customers exist, and lack of public validation is not proof a product fails. These are evidence gaps, not fraud findings — the matrix distinguishes vendor metrics, customer statements, independent tests, regulatory findings and internal deployments.

The evidence ladder: claim first, proof later, if ever

Rows grouped by evidence stage, weakest public proof at top. Hover or tap a cell for the exact wording from the record.
vendor claim (always present) public evidence present present but contested or vendor-reported nothing public found

What the matrix cannot say alone

Tesla Optimus is the loudest case — 21 mentions across 14 hosts promote million-robot capacity, yet zero robots have been sold outside Tesla and JPMorgan analysts found the Fremont Optimus area tarped off. teslanewswire.com
Pieces Technologies shows adoption is no validation: at least four Texas hospitals deployed it live, while the Texas Attorney General found its advertised one-in-100,000 hallucination rate was likely inaccurate. aiplusinfo.com
In the AI SDR category, TechCrunch reporting found 11x listed ZoomInfo and Airtable as customers when they were never paying customers, and counted trial usage as ARR. rapidevelopers.com
BastionGPT's 99.3% accuracy claim rests on a 14-parent, single-center feasibility study of a GPT-4o chatbot — a study that does not validate the current multi-model commercial platform. hugoscore.org
Microsoft's Sales Agent is a "Customer Zero" deployment inside Microsoft's own sales force: usage is tracked monthly, weekly and daily, but the company concedes business outcomes are not yet measured. microsoft.com

All 10 records

ProductEvidence stageNamed customerEvidence gapSource

Method: 10-row shortlist of promoted AI products, compiled from 33 candidate source mentions; each row carries one primary citation (10 URLs across 9 sites). source_mentions and source_hosts count candidate evidence gathered per product, so several rows are single-source evidence reviews even though the shortlist as a whole spans independent hosts. The matrix marks whether a named customer, a public deployment account, or an independent outside voice appears in the record; amber marks evidence that is vendor-reported, internal-only, or contested by regulators or journalists. Long gap descriptions are truncated in the table; full wording is in the cited pages. Compiled from the reviewed material only; absence of evidence here is not proof of absence.

This report was generated automatically by Keenable SELECT at a user's request, from publicly available web sources linked herein. Keenable does not review, verify, or endorse its contents and makes no representation as to accuracy, completeness, or timeliness; AI-based extraction may contain errors. Nothing in this report is investment, legal, financial, or other professional advice. All trademarks and referenced content remain the property of their respective owners; no affiliation or endorsement is implied. To report an error, rights concern, or request removal: legal@keenable.ai.

Keenable SELECTAsk your own question
Made with Keenable SELECT