AI recommendation website audit
This demo HVAC site has some AI recommendation strength — but key high-intent jobs can still go to competitors first
This action plan shows where recommendation opportunities are being lost, what on-site trust and service-area signals may be weakening recommendation confidence, and which fixes should come first.
This report turns the earlier AI recommendation benchmark into a prioritized website action plan.
It separates sampled prompt evidence from page-type, source-type, measurement, proof, and machine-readable support checks.
Opportunity snapshot
What the report is showing first
The business is not invisible.
Sample HVAC Demo Company has enough entity and local context to earn some AI visibility.
The strongest buyer moments are not consistently covered.
Urgent repair, service-specific, and commercial ventilation questions are where competitors can still be named first.
Turn the misses into signal repairs.
Prioritize clearer service/location proof, stronger corroboration, and machine-readable support.
https://example.com/sample-hvac-demo/
Sample URLs are shown for demonstration only and are intentionally rendered as non-clickable text in the public sample report.
Sample benchmark report
Source benchmark reference for this action plan.
2026-07-01T17:53:25Z
97c3aa41
From benchmark to fixes
Where this report fits
This page is the middle step between seeing where the business was missed and implementing the fixes that make future recommendations easier to earn.
Free benchmark
Shows where AI visibility was won or missed in sampled local recommendation questions.
Website audit
Explains why those misses are likely happening across pages, proof, schema, and service-area signals.
Fix Sprint
Implements the approved priority fixes and verifies what changed after the work is live.
Overall AI readiness score
Level 8 of 10 — Good (74/100)
10-level pain scale: Level 8 of 10 — Good
The review found a mix of strengths and gaps. The action plan below prioritizes fixes most likely to help AI systems understand the business, service area, and trust signals more clearly.
What 74/100 means on the 10-level smiley/frowny scale
74/100 maps to Level 8 (Good).
- The demo business is visible in some AI recommendations, so the problem is not total invisibility.
- The score still leaves meaningful leakage in high-intent HVAC clusters where urgency, service specificity, and trust proof matter most.
- The missed opportunities are not random; they point to repairable gaps in local proof, service-page coverage, metadata, and structured support.
Report scope
What the score is based on
The 74/100 score is the quick read on how much signal repair is needed. The details below keep the fix inventory and review scope available after the score is clear.
3 Priority 1 issues, 3 Priority 2 issues, 1 Priority 3 issue; no Priority 0 issues
Sample HVAC Demo Company was mentioned in 2 of 20 logged AI referral customer questions and recommended or shortlisted in 2 of 20 customer questions in this bounded Example City HVAC demo benchmark sample.
Focused review scope
Optional Fix Sprint after this action plan
Action summary
- Sample HVAC Demo Company is not invisible, but it is not yet consistently recommendation-ready in the highest-intent HVAC moments.
- The largest leaks appear where urgency, service specificity, local proof, and structured support need to reinforce each other.
- The fastest path is to fix the priority signal gaps, verify the changes live, and then rerun the same benchmark question families.
Bottom line: The audit is useful because it converts missed AI recommendation opportunities into a prioritized implementation path, with the $500 Fix Sprint as the clearest next step for approved fixes.
System-specific diagnosis map
Different AI surfaces use different evidence patterns. This table keeps ChatGPT-first measurement from turning into ChatGPT-only advice.
| System/surface | What matters | Paid audit check |
|---|---|---|
| ChatGPT | Strongest standalone referral surface in current industry data; may use broad web evidence and page clarity. | Can a clear service, location, proof, and cited-source story be found for the questions tested? |
| Google AI / Gemini | Local discovery remains tied to Search, Google Business Profiles, crawlable pages, reviews, and local details. | Check GBP/local detail consistency, indexability, local pages, reviews, and Search Console availability when provided. |
| Claude | More relevant for professional, technical, B2B, legal, or content-rich buyers. | For relevant niches, review About/trust depth, credentials, detailed service explanations, and citation-worthy guides. |
| Perplexity | Citation and source behavior can matter for content-forward or research-heavy decisions. | Review third-party citations, factual source quality, public proof, and helpful long-form pages when the niche warrants it. |
Measured AI referral evidence, if available
Analytics access was not available, so this report does not claim measured AI referral traffic. Findings are based on benchmark outputs, public page review, and public proof signals only.
Category scorecard
| Category | Checks passed | Score and pain scale |
|---|---|---|
| Crawlability | 6/6 | 100/100Level 10/10 — Excellent |
| Entity Clarity | 5/6 | 83/100Level 9/10 — Strong |
| Content Structure | 5/7 | 71/100Level 8/10 — Good |
| Schema | 2/4 | 50/100Level 6/10 — Mixed |
| Local SEO | 5/6 | 83/100Level 9/10 — Strong |
| AI Answer Readiness | 4/6 | 67/100Level 7/10 — Fair |
| Trust Signals | 4/6 | 67/100Level 7/10 — Fair |
| Technical Performance | 3/5 | 60/100Level 7/10 — Fair |
Missed recommendation opportunities
These are customer-style questions we tested where AI tools did not recommend the business, which may point to missed chances to win attention or leads.
| Opportunity | Count |
|---|---|
| Best / recommended provider | 4 |
| Emergency / urgent local recommendation | 3 |
| Service-specific local recommendation | 3 |
| Near-me local intent | 1 |
| Comparison local intent | 1 |
| Audience/problem local intent | 2 |
| Nearby-market local intent | 2 |
| Problem-specific local intent | 2 |
Question-to-page/source gap map
Emergency / urgent local recommendation
What we saw: Sample HVAC Demo Company missed urgent furnace repair and urgent AC repair customer questions in the benchmark.
Likely causes:
- Competitors appear to have stronger third-party urgency/review proof.
- Sample HVAC Demo Company has broad emergency messaging, but the reviewed site does not show extra machine-readable structure that reinforces that claim.
Recommended fix types: strengthen trusted-source corroboration, add schema, surface emergency proof and response expectations
Commercial ventilation
What we saw: Sample HVAC Demo Company was absent from the commercial ventilation customer question cluster.
Likely causes:
- No dedicated ventilation-first owned page was identified in the selected-page review.
- Commercial hub language is broad instead of tightly mapped to ventilation intent.
Recommended fix types: new/expanded ventilation page, stronger internal linking from commercial hub, add service-specific proof
General non-branded shortlist
What we saw: Competitors such as Example Comfort Company, Demo Service Group, Sample Mechanical, and Example Air Services appeared repeatedly in broad shortlist customer questions.
Likely causes:
- Stronger market corroboration and review presence for competitors.
- Sample HVAC Demo Company entity/service signals are present but not strongly structured for machines.
Recommended fix types: entity/schema improvements, local proof reinforcement, comparison/decision-support content
Pages reviewed
| # | Page URL |
|---|---|
| 1 | https://example.com/sample-hvac-demo |
| 2 | https://example.com/sample-hvac-demo/about |
| 3 | https://example.com/sample-hvac-demo/contact |
| 4 | https://example.com/sample-hvac-demo/service-area/heating-cooling-southside |
| 5 | https://example.com/sample-hvac-demo/service-area/heating-cooling-example-city |
| 6 | https://example.com/sample-hvac-demo/service-area |
| 7 | https://example.com/sample-hvac-demo/service-area/heating-cooling-north-ridge |
Page-type diagnosis map
Previsible-style measurement is more useful when the report separates homepage, service, location, trust, conversion, resource, and internal-search surfaces instead of treating the whole site as one average.
| Page type | Pages or sources | Paid audit check | Evidence note |
|---|---|---|---|
| service page | https://example.com/sample-hvac-demo | Service-specific proof, buyer-intent match, local relevance, and clear internal links. | Derived from selected pages when no explicit page-type map is provided. |
| about/trust page | https://example.com/sample-hvac-demo/about | Credibility signals, team/owner proof, public background, and third-party corroboration. | Derived from selected pages when no explicit page-type map is provided. |
| contact/conversion page | https://example.com/sample-hvac-demo/contact | Phone/form visibility, action path clarity, and conversion readiness for AI-referred visitors. | Derived from selected pages when no explicit page-type map is provided. |
| location/service-area page | https://example.com/sample-hvac-demo/service-area/heating-cooling-southside, https://example.com/sample-hvac-demo/service-area/heating-cooling-example-city, https://example.com/sample-hvac-demo/service-area, https://example.com/sample-hvac-demo/service-area/heating-cooling-north-ridge | City/state consistency, service-area proof, local reviews/listings, and duplicated/cross-city copy risks. | Derived from selected pages when no explicit page-type map is provided. |
Internal search and navigation check
Do not recommend internal search by default for this scope.
For small local service sites, review navigation labels, service hierarchy, local hubs, and internal links instead of adding a search feature just because AI traffic studies mention internal search.
High-priority issues
1. Schema coverage is incomplete on reviewed pages
Where: Homepage, About, Contact, Example City service-area page, Southside service-area page, North Ridge service-area page
What we found: Live browser checks found JSON-LD and Organization-level entity markup on sampled reviewed pages, but the selected-page review did not find Service schema or FAQPage schema, so schema coverage is incomplete.
Recommended next step: Keep the existing entity schema, validate it, and expand visible-content-aligned markup where appropriate: Organization/LocalBusiness on the homepage plus Service schema on core service pages and FAQPage only where FAQ content is actually visible.
Why it matters: The free benchmark missed Sample HVAC Demo Company in 18 of 20 logged AI recommendation customer questions. Existing entity markup can support consistency, but visible service, location, and proof content should lead. Add Service or FAQ markup only where it matches visible page content.
Effort: Medium (2-4 hours)
Supporting detail
- Live browser inspection found JSON-LD including Organization-level entity markup on sampled reviewed pages.
- Example City service-area page exposed JSON-LD in the live browser check, but Service and FAQPage types were not detected in the sampled markup.
2. Geo/service-area template QA error on North Ridge page metadata
Where: https://example.com/sample-hvac-demo/service-area/heating-cooling-north-ridge
What we found: The North Ridge page title/H1 target North Ridge, but the meta description says “Your local Westfield, MI air conditioning company.”
Recommended next step: Correct the North Ridge page meta description so the city matches the page target, then scan all city/service templates for cross-city copy leaks before publishing more geo pages.
Why it matters: Wrong-city metadata weakens local relevance and can make AI systems less confident about which market the page actually serves.
Effort: Low (15-30 minutes)
Supporting detail
- Inspection output captured title “North Ridge, MI Air Conditioning Repair and Service” and meta description “Your local Westfield, MI air conditioning company.”
3. About and Contact pages are missing meta descriptions
Where: https://example.com/sample-hvac-demo/about and https://example.com/sample-hvac-demo/contact
What we found: Both pages returned empty meta descriptions in the focused website review.
Recommended next step: Write concise, non-duplicative meta descriptions that reinforce business identity, market, and trust signals.
Why it matters: These are key corroboration pages for branded/entity understanding. Empty metadata leaves easy context signals unused.
Effort: Low (20 minutes)
Supporting detail
- About page meta_description=""
- Contact page meta_description=""
Medium-priority issues
1. Commercial ventilation intent is weakly matched by owned-page coverage
Where: Commercial section and customer question cluster
What we found: The free benchmark missed Sample HVAC Demo Company on the commercial ventilation customer question cluster, and the reviewed commercial pages did not show a dedicated ventilation-specific landing page or clear ventilation-first page targeting.
Recommended next step: Create or expand a commercial ventilation page that explicitly covers ventilation design/service/repair, building types served, emergency response, and Example City service area language. Link it from commercial hub pages.
Why it matters: One of the missed customer question families asked who handles commercial ventilation systems in Example City. If Sample HVAC Demo Company lacks a clearly attributable owned page for that service, AI systems have less page-level evidence to cite.
Effort: Medium (3-5 hours)
Supporting detail
- Missed customer question family: “Which HVAC companies in Example City install or service commercial ventilation systems?”
- Commercial page check showed mentions_ventilation=false for https://example.com/sample-hvac-demo/commercial
2. Visible FAQ content exists, but no FAQ schema was detected
Where: Homepage and service-area pages
What we found: The reviewed pages appear to include FAQ sections, but the selected-page check found no FAQPage schema.
Recommended next step: If the FAQ sections are stable and visible in HTML, add FAQPage schema that matches the on-page questions and answers exactly.
Why it matters: This can improve machine-readable consistency and rich-result eligibility without changing the core copy, as long as the markup matches visible content.
Effort: Low (30-60 minutes)
Supporting detail
- Homepage has_faq_section=true, has_faq_schema=false
- Example City service-area page has_faq_section=true, has_faq_schema=false
3. Third-party proof exists but is not strongly evidenced in the audit surface
Where: Homepage/About plus off-site listings
What we found: Search results show review-directory and accreditation profiles, and the demo site claims published accreditation and membership claims, but the benchmark still missed Sample HVAC Demo Company in most non-branded customer questions. That suggests the off-site proof either is not strong enough yet or is not being reinforced clearly enough by the owned site.
Recommended next step: Verify the primary GBP/Yelp/Facebook/BBB listings, standardize NAP/categories, and surface named proof on-site (review snippets, memberships, service awards, or project proof) where it is honest and current.
Why it matters: The benchmark repeatedly surfaced competitors instead of Sample HVAC Demo Company in high-intent buyer customer questions, which often correlates with stronger third-party corroboration for those competitors.
Effort: Medium (2-4 hours plus review acquisition follow-up)
Supporting detail
- Web search found review-directory and accreditation listings for Sample HVAC Demo Company.
- Free benchmark recommended competitors such as Example Comfort Company, Demo Service Group, Sample Mechanical, and Example Air Services in multiple non-branded customer questions.
Lower-priority issues
1. Automated browser checks timed out with a JS parsing error
Where: Homepage and Contact page
What we found: Headless browser checks returned title data but also logged page error “Unexpected string” and a blank-body timeout result. Requests-based fetches still returned full HTML.
Recommended next step: Reproduce manually in a normal browser and inspect console/network to confirm whether this is a bot-only quirk or a real JS/rendering issue. If real, fix the offending script so users and renderers do not hit blank-body states.
Why it matters: This is not yet a confirmed customer-facing defect, but render instability can reduce trust in automated renderers and QA tooling.
Effort: Medium (1-2 hours to verify and fix if real)
Supporting detail
- page check on homepage: status TIMEOUT, body_appears_blank=true, page_errors=["Unexpected string"]
- page check on contact page showed the same error signature
Competitors that appeared in missed recommendation opportunities
| Competitor or source | Appearances |
|---|---|
| Example Comfort Company | 16 |
| Sample Mechanical | 15 |
| Demo Service Group | 13 |
| Example Air Services | 8 |
| Fictional Climate Pros | 6 |
Quick wins in the next 7 days
- Validate and keep the existing entity schema, then add Service schema and FAQPage markup only where the FAQ is already visible.
- Fix the North Ridge city-page metadata mismatch and scan every geo page for similar cross-city copy leakage.
- Write unique meta descriptions for About and Contact so branded/entity pages carry stronger context.
- Add a ventilation-first commercial page or strengthen the commercial hub so Example City ventilation intent has an owned landing page to cite.
30-day action plan
- Week 1: validate the existing schema, add missing service/FAQ coverage where appropriate, and fix the geo-page metadata mismatch plus empty meta descriptions.
- Week 2: tighten commercial/service-area coverage for ventilation, refrigeration, and emergency buyer-intent clusters shown in the benchmark.
- Week 3: verify GBP/Yelp/BBB/Facebook consistency, then add truthful proof elements and review snippets to the site.
- Week 4: rerun the AI referral visibility benchmark on the same customer question families to see whether the missed-cluster pattern improves.
Recommended next step
Your clearest path from here
You still have control over how the fixes get done. The recommended path is first because it keeps the audit momentum pointed at implementation instead of turning the report into another document to decode later.
Have Local AI Referrals implement the approved fixes ($500 flat)
Use the Fix Sprint if you want the highest-priority signal repairs handled for you and verified after they are live.
Give this action plan to your developer
Use the prioritized issues above as the implementation brief. Each item names the page, the business reason, the recommended next step, and the effort level.
Fix it yourself
If you have CMS or code access, work through the high-priority items first, validate each change, and avoid invisible or unsupported claims.
Fix it for you
$500 flat-rate Fix Sprint
If you want Sample HVAC Demo Company’s priority fixes turned into clearer service/location signals, stronger trust corroboration, cleaner machine-readable support, and easier recommendation confidence, the next step is the $500 flat-rate Fix Sprint.
We’ll handle the fixes we can implement for you, and if a recommendation depends on a third-party platform, account access, or factors outside our control, we’ll clearly tell you what’s needed next.
What is included
- Implementation of approved P1 and P2 fixes from the paid audit
- Schema markup cleanup and validation where it matches visible content
- Service-page, metadata, internal-linking, and local-signal improvements tied to the audit findings
- Post-implementation verification and a rerun of the visibility benchmark after fixes are live
Expected improvements
- Clearer service and location signals
- Stronger trust corroboration
- Cleaner machine-readable support
- Easier recommendation confidence for future benchmark checks
What to expect
- Timeline: 5-7 business days after approval and access
- Price: $500 flat for the scoped Fix Sprint
- Boundary: Implementation support only; no guarantee of rankings, leads, revenue, or AI recommendations.
Why this recommendation is bounded
- The report is based on a completed AI recommendation benchmark plus public-page review, not a generic SEO checklist.
- The action items are prioritized so implementation starts with the fixes most tied to missed high-intent recommendation opportunities.
- Local AI Referrals is led by Matt LaClear, with technical implementation and QA support from Marc LaClear when approved fixes move into production.
- The Fix Sprint is scoped implementation support, not a promise of rankings, leads, revenue, or AI recommendations.