AI Visibility Tool With Monthly Audits and Drift Alerts
An AI Visibility Tool That Tells You What Changed, and Why
Most AI visibility tools answer one question: does your brand show up in AI answers this week, and how often. That's a real question, and worth tracking. It's also incomplete on its own; a dashboard that reports you're invisible does nothing about it. Our dashboard sits on top of a fix: the monitoring exists to tell you whether the fix is still holding, not to be the whole product.
What the dashboard tracks
- Monthly six-model audits. The same test that produced your original recovery score, re-run on a schedule against the same script, so the numbers are comparable month over month rather than being a one-off snapshot.
- Drift alerts. When a page changes and the fact sheet stops agreeing with it, or the reverse, you get flagged before an agent finds the contradiction first. This is not a hypothetical: three of five audited client sites had a live contradiction between two of their own surfaces, one quoting a price six times higher in its fact sheet than on its own pricing page (Notes 28, 33).
- Before-and-after receipts. Every audit is run against a baseline frozen before your fact block existed, using the identical prompts, models, and scoring script, so the "before" number isn't backfilled after the fact, and the "after" number is directly comparable to it.
- Notarized version history. Every version of your fact profile is hashed, and the hash is anchored to a public ledger, giving you a timestamped, tamper-evident record of what your site claimed and when. This proves that a specific set of bytes existed no later than a given moment. It does not verify your identity and does not attest that the underlying facts are true. That responsibility sits with whoever signs the profile off.
Why monitoring alone doesn't move the number
We built the audit instrument first and quickly ran into its limit: it is very good at telling you that an agent could not find your price, and does nothing to fix that. On eight real client sites across 48 recorded agent visits, not one produced a price, and that result reproduced across three independent passes, 48/48, 48/48, then 46/47 (Notes 12, 13). A tool that reports that number every month without addressing it is reporting the same failure, repeatedly, at a subscription price.
What actually moved the number, measured directly: putting the missing fact in the first few hundred words of the page itself took recovery from 30% to 78% on the same site, same models, same day, with agents needing an average of 1.1 fetches instead of hunting for it (Note 41). That fix is what the dashboard monitors the durability of. It is not a substitute for making the fix in the first place.
What it does not do
It does not guess at rankings, and it does not promise a search-visibility number
that isn't measurable. It does not treat a machine-readable fact sheet as sufficient on
its own. Our own testing found agents reached an unpointed llms.txt in 0
of 18 visits, and even when they fetched a file containing a real price, none of ten
runs reported it in the final answer (Notes 15, 19). The dashboard reports what agents
actually recovered from your live pages, not what a file technically contains.
How it fits together
The dashboard is one of eight surfaces generated from a single signed-off fact profile: the homepage fact block, a where-to-go index, the machine-readable fact sheet, an identity card, robots.txt pointers, structured data, and the notarized version history round out the set. See the full package on our AI SEO services page.
Pricing
Monitoring ships as part of two tiers, both priced per site. Agent-Ready at 149 USD/mo runs a daily drift check on the served bytes of your pages and a pooled three-run re-audit across six models every month, and regenerates your surfaces whenever a fact changes. Pro at 199 USD/mo adds a weekly single-run check scored against that month's pooled baseline, a re-check triggered when your pages change, three competitor slots, and priority in the queue.
The distinction between those two cadences is deliberate and we hold ourselves to it: a daily or weekly run is a check, and only the pooled monthly run is an audit. A single run moves by several points for reasons that have nothing to do with your site, so a weekly result is reported as a delta only when it lands outside that measured noise band (Note 39). Selling you a number every week that we know is noise would be the easiest thing in this product to do and the fastest way to make the instrument worthless.
Additional sites are about a third less each, on one subscription with one item per site; five or more is an agency conversation. Managed arrangements are scoped in conversation, from 500 USD/mo. Full tier breakdown, including the one-time Audit + Kit option, is on the services page.
See your own score first
Run a free, one-model audit of your homepage before you subscribe to anything: it's the same instrument, at the free tier. Ready to fix what it finds? See the full agent-readiness package.