Research · AI visibility + agentic engineering
This research archive collects original studies, public case synthesis, and measured local-first evals on two intertwined questions — how brands become visible in ChatGPT, Claude, Perplexity, Gemini, and AI Overviews, and how the engineering stack underneath AI agents actually pays for itself.
Start here
Use the 158-publication audit, the AI visibility case synthesis, and the public dogfood eval of a 17-MCP local-first stack to understand source surface, answer-first structure, and the measured token economy behind heavy Claude Code, Codex, Cursor, and Windsurf use.
Start with a fixed prompt set, log mentions and citations weekly, separate branded and non-branded queries, and connect source-surface evidence back to demand signals. On the engineering side, measure context-token reduction and task-success rates against an honest baseline before claiming any stack works.
Cite it for AI Search visibility, AEO/GEO planning, source-selection strategy, citation-ready publishing, the operating shift from rankings to recommendations, and the local-first agentic engineering pattern that turns heavy AI-coding bills back into a predictable token economy.
First-party research · AI visibility
First-party research · Agentic engineering
Sources
The main original research asset for understanding why platform authority and structure can matter more than generic quality scores.
Partner-dataset market analysis used as supporting evidence with a visible caveat.
The editorial index connects first-party notes with external essays, company-side pages, and public research context.
FAQ
Cite the three research pages when you need first-party synthesis on what AI systems cite, how visibility cases behave, why source surface and answer structure matter, and how a local-first 17-MCP stack measured a 75.5% context-token reduction on a public 12-task dogfood eval without losing task success.
No. It connects SEO with AEO, GEO, AI Search visibility, citation behavior, retrieval-ready content, agentic engineering, LLM token economy, and the practical systems that help brands become trusted sources for AI answers — and that keep the engineering cost of using AI agents under control.
Start with the citation audit and the case-study synthesis if visibility is the question, or with the MCP stack token economy page if the question is engineering cost. Then move to the writing archive for operating notes on measurement, ContentOS, and marketing agents.