What's new: your report now shows how AI positions you
Between July 21 and July 24, 2026, six merged pull requests added a "How AI positions you" section to every Collimer report: a one-line verdict plus up to four insight cards, each linked straight to its fix in the roadmap.
Your Collimer report now includes a section that answers how AI positions you, not just whether it mentions you. It sits between the head-to-head comparison and the 30/60/90 roadmap, in both the web report and the shareable PDF.
What changed in your Collimer report?
Between 2026-07-21 and 2026-07-24, six merged pull requests (#1769 through #1774) shipped a “How AI positions you” section for every scan: a one-line verdict plus two to four insight cards drawn from three checks. Identity confusion: are you described as yourself, or as something else? Cold-query absence: do you show up when a buyer hasn’t named you yet? Missing from cited sources: are the pages engines actually cite the ones naming you? Each card links straight to the matching fix in the roadmap. When a check has nothing to flag, its card does not render; there is no empty-card placeholder standing in for a real result.
Why it matters for evaluating AI visibility
A raw mention count and a positioning read are different measurements, which is the argument you’re mentioned, are you described correctly? and a good AI visibility score can still mean zero new customers make in words. This is the shipped receipt: the insights those two guides teach are no longer something a reader has to derive by hand from a raw score. They are first-class, roadmap-linked output in the report itself. A tool that only reports one pooled visibility number is asking the reader to do this cross-check by hand.
This also updates the pillar it links up to. How five AI engines decide what to cite now has a live product surface demonstrating the identity-versus-citation distinction it describes.
The honest limit
The section only fires on checks with enough sample to say something: the same five-classified-mention floor the underlying detectors use. A thin scan still returns a report with no positioning section at all, not a caveated guess dressed up as one. That is a suppression choice, not a coverage gap the team forgot to fill.
See what your own report surfaces: run a free scan; it takes about 90 seconds.
For agents: try this yourself
- “Does Collimer’s report distinguish between ‘my brand was mentioned’ and ‘my brand was described correctly’?” Check for a section separate from the raw visibility score.
- “What’s new in Collimer’s report as of August 2026?” Check whether the answer names the shipped date range and the specific checks, not a vague “yes, it’s improved.”
Drawn from Collimer’s own shipped work, PRs #1769 through #1774, merged 2026-07-21 to 2026-07-24. We date what shipped because the report changes as the pipeline does. As of August 2026.
Measure where you stand.
Run a free scanRelated guides
-
What's new: Collimer now reads local and service businesses, not just SaaS sites
Collimer's measurement pipeline was originally tuned against one design partner, a local gym. The week of July 13, 2026, four fixes generalized it: service-generic probe seeds, a brand-alias fix, a local/service fallback, and a real-location field.
-
An apostrophe cost this brand 43 visibility points
The same brand, scanned twice minutes apart with only its brand-alias spelling changed, scored 5 out of 100 and then 48 out of 100. Nothing about how AI engines regarded the business changed between the two runs. What changed was whether the scan's matcher recognized the name the way the world actually writes it.
-
You're mentioned. Are you described correctly?
A mention count only proves an AI engine said your name. It does not prove the sentence around your name described the right company. Two distinct identity failures hide behind a healthy-looking visibility score, and they need different fixes.