Can AI answers get company facts wrong?
Open Dental's row in our Wrong About You study labels 12 of 77 classified statements wrong or stale. All twelve flagged quotes were captured on 22 September 2026, with vendor references checked on 8 October. One published example is the assertion that Open Dental required a local server, which the dated record marked wrong.
What we measured
Measured. The published study and method retain 57 statements labelled true, eight wrong, four stale and eight unverifiable in this vendor's 77-statement row. The flagged records preserve their engine, question, run, quotation and vendor-reference details. Their quote dates are 22 September 2026; the reference-page read date is 8 October. The source keeps Team review: in progress. These historical labels do not establish a business loss, a current product assessment or a trend in error rates.
Reasoned. The deployment assertion and the outdated pricing assertions require different reference checks. Keeping the wrong and stale classes separate tells you whether the retained problem concerns the claim itself or its dated pricing wording.
What this means for you
Read the deployment claim. The local-server assertion appears in retained items for Claude, Google AI Mode and Google AI Overviews. A further Overviews item specifies a Windows server. The study links the local-server items to its Open Dental Cloud reference and the Windows-server item to a different computer-requirements reference. Reading those linked records keeps two distinct deployment assertions from being treated as one.
For your own company, a change in billing terms and a statement about deployment call for different corrections. The Open Dental row makes that distinction visible without combining every problem into a single negative score. Its four stale items are all pricing claims. The eight wrong items divide into four pricing claims and four platform claims, with the historical quote and reference dates retained for each item.
Also asked as
Are you guys starting to see AI backfiring on companies?
Related
- Wrong About You study and data
- The published company-fact study
- The Manual on vendor facts
- Requesting a correction with its source receipt
- The Manual on valid measurement
Next step
Use the Free 10-question check (one engine) for a separate dated company-answer sample. No email is needed; runs are limited per network and per day, and by the site's daily allowance.
For one documented company-fact error, Fix one answer, $490. Keep the retained answer and its supporting public source with the correction.
Our answer-study method, v1.1.
Prepared by the Broadcastwell team.
Page owner: Vikram Patel, AI Architect. Corrections: hello@broadcastwell.com