Prompt results
Which prompts exposed the gap?
The useful evidence is not the score by itself. It is the row where a buyer's question, a competitor recommendation, a cited source, and a fixable error meet.
| Prompt | Brand | Competitors | Reason | Sources | Error |
|---|---|---|---|---|---|
| best dental implant dentist near Phoenix for anxious patients | No | Competitor H, Competitor J | The answer preferred practices with implant-specific pages, patient-anxiety copy, and corroborating reviews. | Practice service page, health directory, review snippets, FAQ page | The prospect's sedation option was mentioned in reviews but not clearly stated on the site. |
| which law firm should a startup use for contract review in Denver | Yes, weak mention | Competitor K, Competitor L | Competitors had more specific pages for startups and contract review while the prospect's page stayed general. | Firm practice pages, local business directory, attorney profile | One directory listed an outdated attorney profile. |
| top Webflow agencies for SaaS redesign under 30k | No | Competitor M | The answer used partner-directory criteria and portfolio pages with explicit scope and budget signals. | Webflow partner profile, portfolio case study, agency service page | The prospect did not publish any budget or minimum-scope signal. |
Missing Source Map
Which cited sources did not support the brand?
01
Prospect screenshot: 20-prompt scorecard
A concise screenshot made the invisible problem visible before the sales call.
Package the free audit summary with raw prompts, competitor counts, and two proof rows instead of vague AI fear.
02
Vertical proof page: Practice-area or service-fit page
The prospect lacked the exact page needed to answer the buyer's constraint.
Recommend one evidence-led page, not a bulk content calendar.
03
White-label report template: Agency-branded PDF report
The client needed a deliverable they could understand without learning answer-engine mechanics.
Use a fixed structure: score, share of voice, prompt rows, missing source map, errors, and 30-day fix plan.
Fix Plan
What should be fixed first?
- 01Narrow the prospect listTarget businesses where competitors already appear in AI answers and the fix is plausible within 30 to 60 days.
- 02Send one screenshot, not a lectureUse the competitor share-of-voice line and one prompt row as the outreach hook.
- 03Sell diagnosis before retainersOffer a paid full audit before promising monthly implementation. The audit proves whether the source gap is real.
- 04Turn fixes into existing SEO workMost tasks map to content, schema, directory cleanup, comparison pages, and reporting the agency already knows how to ship.
How should the retest be judged?
After full audit
The client accepts a fix plan tied to named sources and prompts, not a generic content calendar.
If the client wants only rankings, keep AI visibility as a reporting add-on.
After first implementation sprint
At least one source, error, or recommendation reason changes in the target prompt family.
If nothing changes, inspect whether the cited source was outside the agency's influence.
FAQ
Questions this teardown answers
Why does a white-label report work for outreach?
It shows a specific loss: the prospect is absent while competitors are named. That is easier to understand than an abstract explanation of answer engines.
Should agencies promise AI rankings?
No. Agencies should promise measurement, evidence repair, and retesting. Recommendation behavior is observable but not fully controllable.
What should be hidden in a white-label sample?
Hide private prospects, account data, unpublished strategy, and credentials. Keep prompts, source types, and anonymized findings visible enough to teach the method.
Can agencies use the free audit as a lead magnet?
Yes, if the free audit is specific. A screenshot with prompt count and competitor mentions is more persuasive than a generic AI-search checklist.