How Does AI Describe Your Business?

Your AI Perception Score combines four dimensions — Crawlability, Machine Readability, AI Interpretation, and Alignment — into a single number that tells you how well AI understands your business. Each dimension is weighted by its impact on how models represent you.

Weight: 20% of your AI Perception Score.

Can AI actually access your content? We check robots.txt rules for AI-specific agents — GPTBot, PerplexityBot, ClaudeBot, CCBot — along with sitemap coverage, SSL status, canonical URLs, and the emerging llms.txt and ai.json manifests. Many sites accidentally block the AI crawlers that feed the models people actually use. It's worth checking — many sites are blocked without anyone realizing.

Weight: 25% of your AI Perception Score.

Is your structured data present and correct? We parse JSON-LD schemas, Open Graph metadata, Twitter Cards, hreflang tags, and semantic HTML structure. Missing schemas, malformed markup, and metadata gaps mean AI models may ignore your content entirely — or misrepresent it — even when it reads perfectly to humans.

Weight: 35% of your AI Perception Score.

Does AI actually understand your content correctly? We run your pages through multiple LLMs and analyze their 5W1H interpretation — who you are, what you do, when and where you operate, and how you differ from competitors. We check entity extraction, topic coverage, and potential hallucination triggers. Traditional SEO tools don't measure this — they focus on ranking signals, not on whether AI understands what your business actually does.

Weight: 20% of your AI Perception Score.

Does AI's understanding match reality? We compare what models extract from your content against what your site actually says. Misalignment here means AI may be making claims about your business that aren't accurate — wrong locations, discontinued services, invented specialties. Understanding where AI gets your business wrong gives you something concrete to fix.

Every deep analysis produces a six-section executive summary you can hand to anyone — from your developer to your CEO. It covers Identity (how AI sees your brand), Confusion (where models get it wrong), Barriers (technical issues blocking AI access), Gaps (missing content AI expects), Strengths (what you're doing right), and Priorities (what to fix first).

Written so anyone on your team can read it and know what to do next.

Cockpit detects over 100 issue types across technical AI readiness and content understanding. Examples: AI_AGENTS_BLOCKED, JSONLD_MISSING, LLM_QUALITY_POOR, TITLE_DUPLICATE, CANONICAL_MISMATCH, SITEMAP_STALE. Every issue comes with a severity level, the affected pages, and specific remediation guidance.

If you're using aystos Client, many of these issues can be fixed automatically — right from your CMS.

Every Cockpit audit follows a six-stage process: Discovery identifies your pages and structure. Crawl fetches content the way AI models do. Metadata analyzes your structured data and technical signals. Preselection uses cost-optimized models to intelligently filter pages, saving 70%+ of AI processing costs compared to analyzing every page at full depth. AI Analysis runs deep multi-model interpretation. Finalization compiles your score, briefing, and issue list.

The result: a focused audit that spends its analysis budget where it counts — on the pages AI models are most likely to surface.