Real Audit · Real Product · Real Data
What does an AI Recommendation Audit actually look like?
This is a real audit of Linear. We checked 7 prompts across 4 AI engines, confirmed 3 gaps, and built a prioritized fix plan.
Below is the full report — exactly what a paying customer receives.
Executive Summary
Linear is partially visible to AI recommendation engines, but recommendation quality is limited by comparison gap issues. The highest-priority fix is to improve /compare/[product-vs-competitor] so AI engines can classify, compare, and recommend the product more confidently.
86%
Mention Rate
6/7 prompts
4
Engines Checked
of 6 major engines
3
Gaps Confirmed
validated by human
3
Fix Actions
prioritized plan
Key findings
Across 7 prompt checks, the product was mentioned in 6 and absent from 1.
Mention rate by layer: discovery 3/3, comparison 1/2, purchase_intent 1/1, trust 1/1.
2 high-severity gaps confirmed: Comparison Gap, Evidence Gap.
Jira dominates discovery queries across all 6 engines
Linear appears in 5/6 engines for discovery but only 3/6 for comparison
No engine can state Linear's exact starting price correctly
What we asked the AI engines
Each prompt tests whether AI can recommend Linear in a specific context. A missing mention means a potential customer won't hear about you.
| Prompt | Engine | Layer | Result |
|---|---|---|---|
| “What is the best project management tool for engineering teams?” | ChatGPT | Discovery | #2 |
| “What is the best project management tool for engineering teams?” | Perplexity | Discovery | #3 |
| “What is the best project management tool for engineering teams?” | Gemini | Discovery | #4 |
| “Linear vs Jira for startups” | ChatGPT | Comparison | |
| “Linear vs Jira for startups” | Gemini | Comparison | |
| “Should I use Linear for my 50-person engineering team?” | ChatGPT | Purchase Intent | |
| “Is Linear secure enough for enterprise use?” | Claude | Trust |
Surface coverage
AI engines gather information from specific page types. Missing surfaces mean AI can't answer questions about that topic.
Found
Missing
What AI currently gets wrong
Competitors are easier for AI to compare
AI cannot easily compare the product against alternatives, reducing inclusion in shortlist and versus queries.
Evidence signals are too thin for confident recommendations
AI cannot find enough proof, testimonials, case studies, or verifiable claims to trust the product.
Product information is inconsistent across pages
Public product information is outdated, contradictory, or not consistently repeated across key pages.
Top gaps identified
Comparison Gap
Comparison pages exist but lack structured verdict tables and per-feature fact rows. AI cannot easily extract Linear's differentiation in 'vs' queries.
Evidence Gap
No named testimonials or case studies with measurable outcomes found on any public page. Trust signals rely on logo walls only.
Consistency / Freshness Gap
Pricing page and homepage both mention pricing but with slightly different framing. No last-updated dates on pricing or docs pages.
Prioritized fix plan
This is the core deliverable. Each action targets a specific page and gap, ranked by expected impact on recommendation readiness.
Create dedicated comparison pages
High impactPage: /compare/[product-vs-competitor]
AI cannot easily compare the product against alternatives, reducing inclusion in shortlist and versus queries.
Add named testimonials with company and role
High impactPage: Homepage / About
AI cannot find enough proof, testimonials, case studies, or verifiable claims to trust the product.
Audit key public pages for factual consistency
Medium impactPage: Homepage / About / Pricing / Docs
Public product information is outdated, contradictory, or not consistently repeated across key pages.
This is what $199 buys you
Page-level diagnosis across 6 AI engines, confirmed gaps with evidence, prioritized fix plan with effort/impact, and a recheck path to verify improvements.
What to recheck after fixes
After implementing the top fixes, recheck these signals:
- Re-run the same prompt set 7 days after key fixes are published.
- Track whether AI engines cite updated pricing, trust, and comparison surfaces.
- Compare recommendation quality before and after fixes.
Free audit vs. AI Recommendation Audit
| What you get | Free | $199 |
|---|---|---|
| Score across 6 engines | ||
| Top-level gap categories | ||
| Page-level diagnosis | ||
| Prompt evidence (what AI said) | ||
| Prioritized fix plan with effort/impact | ||
| Surface coverage audit | ||
| Recheck path + Sentinel baseline | ||
| Public audit teardown draft |
What this audit is not
This is not a guarantee that AI assistants will recommend you. No honest product can promise that. It helps you understand what is weakening your recommendation readiness and what to improve first.
- · AI engine outputs are non-deterministic and may vary by session, region, or time.
- · Static crawl findings may miss JavaScript-rendered content.
- · Recommendation improvement cannot be attributed to a single page change without repeat checks.
What this audit helps you decide
- • Is AI even aware my product exists?
- • Where is AI getting my positioning wrong?
- • Which competitors are being cited instead — and why?
- • Which pages on my site need fixing first?
What we would fix first
- • Homepage: clear answer-first positioning
- • Pricing page: structured, machine-readable pricing
- • Compare pages: head-to-head data AI can cite
- • Proof signals: testimonials, integrations, case data
Who should upgrade to the paid audit
- • Your free audit score is below 12 (Low or Critical)
- • You need page-level diagnosis, not just a score
- • You want a prioritized fix list your team can execute
- • You're preparing to invest in AI visibility fixes
When to use Sentinel after fixes
- • After you've shipped fixes to 3+ key pages
- • When you want weekly proof that fixes are working
- • To catch drift if AI descriptions regress
- • Not before — monitoring without fixes is just watching
Ready to see your own gaps?
Run a free audit in 30 seconds to see your score, or get the full page-level diagnosis with a prioritized fix plan.