geo/aeoplaybooks

We tried to review Scrunch AI. Its onboarding blocked us at the last step.

Our hands-on Scrunch AI test ended at step 6 of 6: "This brand name violates our company brand policy." What we saw of the onboarding before the door closed — limits, auto-scraping, UX details — and t

GEO & AEO Playbooks We tried to review Scrunch AI. Its onboarding blocked us at the last step. geo/aeo playbooks · independent GEO lab

We tried to review Scrunch AI. Its onboarding blocked us at the last step.

This was meant to be the fourth hands-on review in this week's series. Instead it's a short field note, because our test ended on the final click of a six-step onboarding with a toast message: "This brand name violates our company brand policy." The brand we'd entered operates in the AI-visibility space — Scrunch's own category — and Scrunch, it turns out, doesn't onboard competitors.

How we tested: real signup on July 12, 2026, all six onboarding steps completed, blocked at Finish. Screenshots throughout. Disclosure — we build GEO tools ourselves (the free checker and Monitor). Given how this test ended, that disclosure is apparently more than a formality.
TL;DR
  • Blocked at step 6 of 6 with "This brand name violates our company brand policy" — after all setup work was done.
  • What we saw before that: onboarding capped at 3 personas, 5 competitors, 5 topics; brand details auto-scraped from your site.
  • No pricing, no engine list, no dashboard reached — we can't review what we couldn't enter.
  • We'll retest with a neutral brand and update this post.

What the onboarding revealed before the door closed

Six steps with a live preview pane on the right that assembles your brand profile as you type — a genuinely nice pattern. Step one asks "For my brand" or "For my clients," so agency mode exists from the first screen. Step two auto-fills your website URL, brand name and primary country, with the note "We gathered this information from your website" — scraping instead of a blank form (company size it left for us to pick).

Then the limits appear as counters: up to 3 buyer personas, 5 competitors, 5 topics. The copy around them is the clearest articulation of the product we saw: competitors — "We'll track when you're mentioned together and estimate your share of visibility"; topics — "so you can see prompt visibility, citations, and rank by theme." The signup screen backs this with an enterprise logo wall: Clerk, Runpod, Lenovo, SKIMS, hims, NatWest.

The UX details a paying user would hit

  • The brand-name policy check runs only at Finish — step six — not at step two where the name is entered. All six steps of work are lost at the door. Validate early; this is the whole lesson of the post.
  • Duplicate topics aren't deduplicated: the same topic added twice consumed two of the five slots.
  • Persona normalization over-trims: "IT buyers" became a chip reading just "It."
  • One auto-generated persona description came back in Russian inside an otherwise English UI.

None of these are fatal. All of them are the kind of thing you only learn from actually walking the flow — which is why we publish these notes.

On the policy itself

To be fair about it: gating competitors out of your product is a legitimate choice, especially in a category where everyone is scraping everyone. But the execution — a generic "violates our company brand policy" after six completed steps — treats a rejection Scrunch could have made instantly as the user's sunk cost. And it leaves us unable to tell you anything about pricing, tracked engines, data quality or exports, because none of those appear before the wall.

What happens next

We'll rerun the test with a neutral, non-competing brand and either expand this into a full review or note what changed. Until then, Scrunch is the only tool of this week's four about which we have exactly zero product data — not because it's gated by price, but because it's gated by identity.

Meanwhile the part of AI visibility that needs no vendor's permission is still free: run the 5-signal check on your domain, see what AI crawlers actually get, and fix that first.

Also in this series: LLMrefs review , Otterly review , Profound pricing teardown .

No comments yet