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 customer onboarding software for a 40-person SaaS company | No | Competitor A, Competitor C | The answer cited category lists that described onboarding workflows, implementation speed, and integrations in buyer language. | G2 category page, integration marketplace, two vendor comparison pages | The brand's current integration was absent from one marketplace profile. |
| alternatives to Competitor A for startup onboarding | Yes, position 4 | Competitor A, Competitor B, Competitor D | The brand appeared only after the model shifted from category evidence to broader software alternatives. | Vendor alternatives page, product docs, third-party roundup | The answer described an old free-plan limit that no longer matched the pricing page. |
| onboarding software with SOC 2 and HubSpot integration | No | Competitor B | Competitor B had a security page, integration page, and review-site profile that all repeated the same claim. | SOC 2 trust page, HubSpot marketplace, review profile | No false answer about the brand; the issue was missing extractable evidence. |
Missing Source Map
Which cited sources did not support the brand?
01
Category list: Best onboarding software roundup
The list supplied the initial shortlist for broad category prompts and gave competitors a reusable description.
Create a source packet with ideal customer, proof points, limitations, and current pricing caveats before outreach.
02
Integration marketplace: HubSpot ecosystem profile
The marketplace corroborated a technical claim that first-party copy stated only on a buried docs page.
Update the profile, add screenshots, and link the canonical integration page from the product navigation.
03
Comparison page: Competitor A alternatives article
The page framed the substitution decision while the brand's site had no honest alternative page of its own.
Publish a fair alternatives page with implementation fit, weak-fit boundaries, and dated pricing facts.
Fix Plan
What should be fixed first?
- 01Normalize product truthCreate one approved source for integrations, plan limits, SOC 2 status, implementation time, and ideal customer size.
- 02Publish one buyer-grade alternatives pageDo not copy competitor tables. Explain when the brand is a better fit and when another tool remains better.
- 03Repair third-party profilesUpdate marketplace and review-site descriptions so the same scoped claims appear outside the company domain.
- 04Retest constrained promptsRetest only the prompt families touched by the fix so movement can be tied to evidence changes.
How should the retest be judged?
After recrawl
The same cited sources include the corrected integration and pricing language.
If cited sources remain stale, repair distribution before publishing more pages.
After two answer runs
Brand appears in at least 6 of 20 prompts or receives a clearer recommendation reason.
If mentions rise but citations stay weak, improve first-party evidence density.
FAQ
Questions this teardown answers
Was the SaaS site technically blocked?
No. The sample diagnosis assumes the site could be crawled. The gap came from weak corroboration, not a robots or sitemap failure.
Should the first fix be more blog content?
Not here. The first fix is product-truth cleanup and evidence distribution because the losing prompts depended on current capabilities and third-party validation.
Can one alternatives page change every prompt?
No. It can improve a specific substitution family. Category, security, integration, and pricing prompts still need their own evidence.
Why keep this case anonymized?
AI visibility audits can reveal competitor gaps and factual errors. Public examples should teach the method without exposing private strategy.