Proof/AI visibility analysis
What answer engines say about you, and why.
A full analysis for Northlane, a fictional B2B SaaS company, across a frozen 40-prompt set.
The short answer
An AI visibility analysis measures how often a brand is cited across a fixed prompt set on each major answer engine, diagnoses why the gaps exist, and returns a prioritized fix list tied to the prompt classes each fix should move.
- The prompt set is frozen at baseline so progress cannot be manufactured by swapping prompts
- Results are reported per engine and per prompt class, not as one vanity score
- Every gap names the mechanism, not just the symptom
- Fixes are sequenced and mapped to the prompt class they should move
Example data. Northlane is a fictional B2B SaaS company we use to show the exact format of our deliverables. Nothing on this page is a client result.
Step one
Freeze the prompt set.
Forty prompts across five intent classes, written down and locked before any measurement happens.
Prompt set
Example account: Northlane
| Intent class | Example prompt | Count |
|---|---|---|
| Problem | how do compliance teams reduce manual review time | 8 prompts |
| Solution | what is compliance workflow automation | 8 prompts |
| Category | best compliance workflow automation tools | 8 prompts |
| Comparison | alternatives to the incumbent case management tool | 8 prompts |
| Qualification | compliance automation that keeps an audit trail for regulated firms | 8 prompts |
Step two
Measure citations, engine by engine.
Citation results, baseline
| Prompt class | Engine A | Engine B | Engine C | Engine D |
|---|---|---|---|---|
| Category prompts | 2 of 8 | 1 of 8 | 0 of 8 | 3 of 8 |
| Comparison prompts | 1 of 8 | 0 of 8 | 0 of 8 | 1 of 8 |
| Solution prompts | 4 of 8 | 3 of 8 | 2 of 8 | 5 of 8 |
| Problem prompts | 1 of 8 | 1 of 8 | 0 of 8 | 2 of 8 |
| Qualification prompts | 3 of 8 | 2 of 8 | 1 of 8 | 4 of 8 |
Cited means the brand is named in the answer. Being linked in a source list without a mention is tracked separately, because it converts at a far lower rate.
Step three
Diagnose the gaps.
Gap diagnosis
| Gap | Mechanism | Severity |
|---|---|---|
| No comparison or alternatives pages exist | Comparison prompts are answered almost entirely from third-party review sites, so the brand is absent from the shortlist stage. | Critical |
| Category description is inconsistent across sources | The site, the review profiles, and the press coverage each describe the product differently, so engines have no stable entity to cite. | Critical |
| No answer-first definitions on money pages | Pages open with context rather than a quotable definition, which is what engines lift. | High |
| Thin structured data | No FAQPage, no Product schema, no llms.txt, so machine readers work harder than they need to. | High |
| Regulated-firm qualification is buried | The strongest differentiator appears only in a PDF, not on an indexable page. | Medium |
| No original data | Nothing on the domain is worth citing on its own, which caps ceiling regardless of on-page work. | Medium |
Step four
The prioritized fix list.
Fix list
| Rank | Fix | Timing | Prompt class it should move |
|---|---|---|---|
| 1 | Publish four alternatives and comparison pages | Weeks 1 to 4 | Comparison prompt citations |
| 2 | Normalize the category description everywhere the brand appears | Weeks 1 to 3 | Entity consistency and category prompts |
| 3 | Rewrite the top six pages answer-first with definitions | Weeks 2 to 6 | Solution prompt citations |
| 4 | Add FAQPage, Product schema, and llms.txt | Week 2 | Machine readability across all prompt classes |
| 5 | Move the regulated-firm qualification onto an indexable page | Week 5 | Qualification prompt citations |
| 6 | Ship one original data asset | Weeks 6 to 12 | Third-party citations and ceiling |
The benchmark data behind the severity ratings comes from our own study of 160 software websites, published in full on the AI Discovery Technical Hygiene Benchmark. The channel detail sits on our AEO service page and GEO service page.
Keep reading the operating system
Every artifact lives in the proof library. The engagement itself is documented in our methodology, the commercial terms are on pricing, and the fastest paid entry point is the Growth System Diagnostic.
Want a real analysis of your own domain?
The free AI discovery technical hygiene snapshot scores your domain against ten technical hygiene signals in about a minute, no call required.