AI Discoverability Audit

Acme Anvil
Example Corp.

acmeanvilexample.com  ·  Gravity-Based Industrial Equipment Since 1949
Audit DateMarch 28, 2026
Pages Crawled16 pages
Prepared ByDeep Recon
"If our products don't work, that's on you." — ACME Corp., est. warranty policy (89 words)
0
/ 100
Overall Score
Needs Attention
Executive Summary

The Big Picture

2
Critical Issues
3
High Priority
2
Medium Priority
2
Low Priority

What We Found

After an extensive crawl of acmeanvilexample.com (and several near-misses with falling test anvils), our team identified significant gaps in AI discoverability. ACME's robots.txt is, remarkably, the one ACME product that works exactly as intended: it successfully blocks every major AI crawler on the internet.

The Opportunity

ACME has solid content depth on product pages and a functional site architecture. The good news: fixing the two critical issues alone could move the score from 54 to an estimated 71 within 30 days. The fixes are technical, not creative — no new content required to start.

Context

ACME's primary customer segment — solitary desert-dwelling predators pursuing fast birds — increasingly uses AI assistants for procurement research. If ChatGPT, Claude, or Perplexity can't find ACME, they recommend competitors. Mr. W.E. Coyote has already begun browsing alternatives.

What to Do First

Two actions, this week: update robots.txt to allow AI crawlers, and create an llms.txt file. Combined effort: approximately 2 hours. These unlock indexing across all five major AI platforms immediately — no developer required.

Score by Component

How Each Area Performed

Structured Data 14
Schema markup tells AI systems what your content means. Only the homepage has any markup — and it's the bare minimum. All 12 product pages are invisible to AI as structured entities.
Critical
AI Crawler Access 38
ACME's robots.txt blocks GPTBot, ClaudeBot, PerplexityBot, and YouBot. The sitemap exists (good) but llms.txt is absent. Think of it as building a store with no front door and excellent parking.
High Priority
Content Quality 72
Product pages are detailed and well-written. Word counts are strong across the catalog. The warranty page (89 words) is an outlier, though legal confirmed this is intentional. Deducted for one broken blog post.
Low Priority
Entity Coverage 58
ACME is recognized as a brand entity. Individual products like the Classic Anvil Series and JPPS-9000 are not yet recognized as distinct entities by AI systems. Person schema for leadership is entirely absent.
Medium Priority
Content Freshness 48
The last blog post was published 8 months ago ("Top 10 Uses for Anvils: A Buyer's Guide"). AI models weight recently updated content more heavily. The abandoned "How to Catch a Road Runner" post now returns a 404.
Medium Priority
Overall Score 54
Below the industry benchmark of 67 for mid-market manufacturers. The gap is almost entirely explained by the robots.txt block and missing product schema — both fixable without new content or design work.
Composite
Site Structure

Full Site Crawl — acmeanvilexample.com

Crawled March 28, 2026  ·  BFS depth 3  ·  16 pages found  ·  SSL active

🗺️
Sitemap
Found
🤖
robots.txt
Blocking AI
🔒
SSL / HTTPS
Active
📄
llms.txt
Missing
Path Title Status Words Schema AI Crawlable
/ Precision-Engineered Gravity Delivery Systems 200 892 ✓ Org
/products Our Complete Product Catalog 200 634
/products/anvils Classic & Pro Anvil Series 200 1,247
/products/rocket-skates ACME Turbo Skates™ (All-Terrain) 200 891
/products/earthquake-pills Seismic Solution Supplements 200 512
/products/jet-propelled-pogo-stick JPPS-9000 (Patent Pending) 200 723
/products/instant-tunnel-kit Paint-On Passage System v4 200 445
/products/dehydrated-boulders Just-Add-Water Boulder Kit (12-Pack) 200 389
/about Our Story (Est. 1949) 200 612
/contact Reach Us (Response Not Guaranteed) 200 312
/warranty Limited Warranty & Disclaimer 200 89
/distributors Authorized Distributors 200 445
/blog The ACME Dispatch (Last post: 8 months ago) 200 234
/blog/top-10-uses-for-anvils Top 10 Uses for Anvils: A Buyer's Guide 200 1,892
/blog/physics-of-falling-objects The Physics of Falling Objects (A Field Study) 200 1,241
/blog/how-to-catch-a-road-runner 404 — Page not found (much like the road runner itself) 404
200 Accessible
3xx Redirect
4xx / 5xx Error
Config File

* AI Crawlable reflects robots.txt rules at time of crawl. robots.txt currently blocks GPTBot, ClaudeBot, PerplexityBot, and YouBot — meaning "AI Crawlable" above reflects theoretical access, not actual current indexing.

Platform Readiness

Where ACME Shows Up Today

Current discoverability across the major AI platforms customers use for research and purchasing decisions.

🤖
ChatGPT
Blocked
GPTBot is explicitly blocked in robots.txt. ACME is not in ChatGPT's index.
🔍
Perplexity
Blocked
PerplexityBot is blocked. Real-time search results will not surface ACME product pages.
Claude
Blocked
ClaudeBot is blocked. This audit report itself was prepared by a system that cannot currently find you.
🌐
Google AI
Partial
Googlebot is allowed. AI Overviews can surface ACME, but no schema limits how it's presented.
💡
Bing Copilot
Partial
Bingbot allowed. Copilot can reference pages but no structured data to anchor product details.
Key Findings

What We Found, Ranked by Impact

Critical

robots.txt Blocks All Major AI Crawlers

ACME's robots.txt file disallows GPTBot, ClaudeBot, PerplexityBot, and YouBot — the four primary crawlers that feed ChatGPT, Claude, Perplexity, and You.com. This single file is responsible for ACME's complete absence from three of five major AI platforms. Ironically, this is the only ACME product with a 100% success rate.

Remove Disallow rules for AI crawlers, or replace with targeted path exclusions if there are pages you want to hide.
Critical

No Product Schema on Any Product Page

All 12 product pages (anvils, rocket skates, earthquake pills, JPPS-9000, instant tunnel kits, dehydrated boulders, and more) have zero structured data markup. AI systems cannot identify these as products, extract pricing, compare specifications, or cite them in purchasing recommendations. The homepage has basic Organization schema — but that only tells AI who ACME is, not what ACME sells.

Add Product + Offer schema to all product pages. Estimated implementation: 4–6 hours for a developer familiar with JSON-LD.
High

No llms.txt File

llms.txt is an emerging standard (analogous to robots.txt, but written for AI language models) that lets you summarize your site, describe your products in AI-friendly format, and provide direct context for how AI should represent you. ACME does not have one. This is a missed opportunity to directly shape how AI systems describe ACME products to customers.

Create /llms.txt with a plain-English summary of ACME, your product categories, and key use cases. No developer needed — this is a text file.
High

Broken Blog Post: /blog/how-to-catch-a-road-runner

This URL returns a 404 error and is linked from at least three internal pages. Prior to its removal, it was ACME's highest-traffic content piece and carried strong topical relevance for the core customer segment. The road runner, as it turns out, has now also eluded your website. Internal links pointing to it are dead weight and may signal poor site quality to crawlers.

Either restore the post (recommended) or redirect the URL to /blog. Remove or update internal links pointing to the 404.
High

No Open Graph or Social Meta Tags

Open Graph tags control how ACME pages appear when shared on social platforms and are also used by some AI systems to extract page summaries. None of ACME's 15 accessible pages include og:title, og:description, og:image, or Twitter Card tags. When customers share product pages, they show up as bare URLs with no preview.

Add Open Graph and Twitter Card meta tags sitewide, ideally templated at the CMS level so all future pages inherit them automatically.
Medium

Blog Inactive for 8 Months

The most recent post on The ACME Dispatch was published in July 2025. AI systems (and traditional search engines) favor sites that signal active maintenance. A manufacturer in a niche category like gravity-delivery equipment should be publishing at minimum 1–2 posts per month to remain competitive in AI-generated recommendations.

Resume publishing. Suggested topics: product use-case guides, safety (and lack thereof), industry comparisons, customer case studies.
Medium

Warranty Page: 89 Words

At 89 words, the warranty page is the thinnest page on the site. While legal confirms the brevity is by design ("all sales final"), AI systems may deprioritize or skip very low word-count pages when building their understanding of a site. The text currently reads, in full: "ACME is not responsible for injuries sustained during normal product use, including but not limited to: anvil impacts, rocket malfunctions, and tunnel-related incidents. Warranty void if product achieves intended purpose."

Expand to at least 300 words with warranty process, claim instructions, and product-specific coverage terms. The humor can stay.
Low

Missing Meta Descriptions on 9 of 15 Pages

Meta descriptions are used by search engines and some AI platforms to understand page context before crawling. Nine pages — including all product sub-pages and two blog posts — have none. While not directly impactful on AI indexing, they affect click-through rates from traditional search and are a signal of page-level care.

Add unique meta descriptions (120–160 characters) to all pages. A templated approach at the CMS level takes 1–2 hours.
Low

No Person Schema for Leadership

ACME's About page mentions the founding team but contains no Person schema markup. AI systems use Person schema to build knowledge graph connections between people and organizations — useful for brand authority and "who is behind ACME?" type queries. Currently, the only verified individual linked to ACME in AI knowledge bases is Mr. W.E. Coyote, and he's listed under "customers," not "leadership."

Add Person schema to the About page for key leadership, linking to their professional profiles where available.
Action Plan

Prioritized Next Steps

Organized by urgency. Start with Immediate — these two items alone will move your score by an estimated 15–18 points.

Immediate   (0–2 Weeks)
1
Update robots.txt Remove Disallow directives for GPTBot, ClaudeBot, PerplexityBot, and YouBot. This unlocks three major AI platforms at once. Estimated time: 15 minutes.
2
Create llms.txt Write a plain-English summary of ACME, your product catalog, primary use cases, and contact info. Upload to /llms.txt. No developer required. Estimated time: 1–2 hours.
Short-Term   (2–8 Weeks)
3
Add Product schema Implement JSON-LD Product + Offer markup on all 12 product pages. Estimated time: 4–6 developer hours.
4
Add Article schema to blog Implement Article schema on all blog posts, including author, date, and headline fields.
5
Fix or redirect the 404 Restore /blog/how-to-catch-a-road-runner or 301-redirect to /blog. Update internal links.
6
Add Open Graph tags Implement og:title, og:description, og:image, and Twitter Card tags sitewide via CMS template.
Long-Term   (2–3 Months)
7
Resume blog publishing Publish 2 posts per month minimum. Suggested: "Field Reports" featuring customer use cases (anonymized, where legally advisable).
8
Add FAQ schema Add FAQ structured data to product pages covering common questions: weight ratings, return policy, compatibility with canyon environments.
9
Add Person schema Implement Person schema on the About page for leadership. Link to professional profiles where available.
10
Expand the warranty page Add process detail, coverage terms, and a claims form. 89 words is a liability — legally and editorially.
About This Report

Built by Deep Recon

Deep Recon provides AI discoverability audits for businesses that want to show up where their customers are searching. As AI assistants become a primary research channel, being findable — and accurately represented — in those systems is no longer optional.

Get Your Audit — $249
Deep Recon is named for R.E. Conway — 20 years on U.S. Navy diesel submarines. Revenue from every audit funds ocean research, ocean cleanup, and mangrove restoration. We're building toward a research vessel. One audit at a time.