# Combot

> Combot measures what AI models actually say about your brand — visibility, citations, recommendation share and accuracy across Claude, ChatGPT, Gemini and Perplexity. By Joni Kautto / Accolade Partners Oy.

## Core pages
- [About](https://combot.ai/about/): Why Combot exists, who built it, and what we believe about analytics in the AI era.
- [The 7 layers of AI visibility](https://combot.ai/blog/7-layers/): The Combot AI visibility funnel: seven measurable stages from Trained-on to Acted-on, mapped to the three Knowledge Modes, aligned with how the rest of the industry talks about generative visibility.
- [AI Share of Recommendation: a new north-star metric](https://combot.ai/blog/ai-share-of-recommendation/): Why
- [Fact-checking the machines: building an automated accuracy center](https://combot.ai/blog/fact-accuracy/): Hallucination isn
- [Blog](https://combot.ai/blog/): Notes on building the AI visibility stack. KPI framework, technical architecture, common failure modes.
- [AI Knowledge Modes: Memory, Search, Fetch](https://combot.ai/blog/knowledge-modes/): AI answers are assembled from Memory, Search, and Fetch; each mode fails differently, rewards different work, and moves on a different timescale.
- [Lean Render: AI-Bot Rendering Without the JavaScript Trap](https://combot.ai/blog/lean-render/): Dynamic rendering died for Googlebot, but AI bots made the problem return. Lean Render is the clean, token-efficient HTML path for AI fetchability.
- [Measuring Fetch: per-vendor URL recall in practice](https://combot.ai/blog/measuring-fetch/): A practical method for measuring whether AI systems can fetch, read, and summarise exact URLs across vendors.
- [Measuring Memory: probing the parametric layer](https://combot.ai/blog/measuring-memory/): Measure AI Memory by disabling tools, using declarative top-3 prompts, and scoring parametric brand recall across models.
- [Measuring Search: instrumenting live AI retrieval](https://combot.ai/blog/measuring-search/): Search-mode measurement is not rank tracking; it measures which retrieval backend a model uses, which URLs survive synthesis, and which brands appear in the grounded answer.
- [Optimising for Fetch: surviving the single-URL pull](https://combot.ai/blog/optimising-fetch/): Fetch mode is the exact-URL test for AI visibility. Learn how to make JavaScript apps readable to Claude, ChatGPT, Perplexity and other AI fetchers.
- [Optimising for Memory: how brands earn durable presence in LLM training data](https://combot.ai/blog/optimising-memory/): Memory is the durable layer of AI visibility. Learn how brands earn parametric presence through independent sources, entity anchoring, and evergreen evidence.
- [Optimising for Search: the engineering side of AI retrieval](https://combot.ai/blog/optimising-search/): Search mode is retrieval-time candidate selection. Learn how to engineer your website for RAG pipelines, llms.txt, and AI query-rewriting.
- [Server logs for the AI era: what your access log already knows about your visibility](https://combot.ai/blog/server-logs-ai-visibility/): AI bots split by purpose — training, search, fetch. Parsed access logs are the most highly-correlated leading indicator of LLM visibility. The 2026 inventory + the Brave↔Claude correlation.
- [Source mapping: tracing an LLM's answer back to its roots](https://combot.ai/blog/source-mapping/): The citation is not the source. The source is whoever wrote the page the model paraphrased from.
- [From fetchability to trust: the technical SEO of language models](https://combot.ai/blog/technical-seo-of-llms/): AI bots crawl differently. Most don
- [Tools for Modes: how the major LLMs implement Memory, Search, and Fetch](https://combot.ai/blog/tools-for-modes/): A per-vendor tour of the tools that make Memory, Search, and Fetch real — what Claude, GPT, Gemini, Perplexity, and Grok actually ship, and how the timeline got here.
- [FAQ](https://combot.ai/faq/): Short, opinionated answers about Combot, AI Share of Recommendation, the 7-layer funnel, the three knowledge modes, and the pilot.
- [Features](https://combot.ai/features/): What Combot actually does: data ingestion, anomaly detection, multi-channel correlation, AI visibility tracking, chat-native command (Slack, Teams, Mattermost — or any platform via API integration).
- [Combot](https://combot.ai/): Combot is the BI platform for the AI era. It correlates business data with how language models see your brand — from training-data memory to final recommendation.
- [Combot knowledge base](https://combot.ai/knowledge/): The Combot knowledge base — the 7-layer AI visibility funnel, the three knowledge modes, AI Share of Recommendation, and the playbooks behind them.

## Articles