Selected engagements
Bring me a bounded, expensive problem.
I work on AI visibility, agent infrastructure, and search-to-revenue systems. The engagement begins with evidence and ends with a usable artifact, not a deck explaining that AI is important.
- 01
AI search visibility
Find where a company disappears between search, AI answers, buyer questions, cited sources, and the page that should close the decision.
When this fits: Start here if buyers ask questions your company should answer, but you do not know which pages or sources AI systems find and cite.
- AI visibility diagnostic: what the evidence supports and what is still missing
- Map of buyer questions, search queries, sources, and AI citations
- Page plan: which buyer questions to answer and which page should be the source
- Prioritized implementation plan with agreed measures of success
Related work — Constellation (private evidence)
- 02
Agent-native product systems
Design and build APIs, MCP tools, RAG systems, control planes, and workflows that agents can use without inventing policy as they go.
When this fits: Start here if an agent needs to take real actions and you need the rules, permissions, and failure cases made explicit before expanding the build.
- Product and architecture specification
- Explicit rules for agent actions, permissions, and evidence
- A working end-to-end part of the product
- Handoff with tests and visibility into what the system is doing
Related work — InboxPolicy (public evidence) · Gauntlet (working evidence)
- 03
Search-to-revenue infrastructure
Turn real demand into useful pages, tools, proof, distribution, and measurable customer actions without feeding the content landfill.
When this fits: Start here if search demand exists, but the path from a useful answer to a customer action is scattered across pages, tools, and measurement.
- Review of real search demand and the constraints in the way
- Research record with sources
- Public tools, resources, or pages that help buyers decide
- Indexing, conversion, and outcome instrumentation
Related work — OnFarmCompost search system (public evidence)
Evidence → design → build → verify → measure.
Scope is explicit. Claims are sourced. Private data stays private. The handoff includes what exists, how it was tested, where it can fail, and what to measure next.
Not sure which lane fits?
Start with the bottleneck. Missing from AI answers? Visibility. An agent needs to act safely? Product systems. Buyers can find you but the next step is a mess? Search-to-revenue. A problem can cross lanes; bring the decision you need to make rather than a shopping list of tools.
What does the related work prove?
InboxPolicy shows an email API with explicit send decisions. Gauntlet shows a working verification foundation. OnFarmCompost shows pages and tools built around buyer decisions. Constellation is private AI-search research. These are examples of the systems behind the work, not promises of your results. The evidence labels explain what you can inspect and what remains restricted.
Before the first call
Bring the problem, the system or pages involved, and the evidence you already have. Use the call to discuss the constraint, the decision you need to make, and a useful first scope. A project starts with a bounded problem, not a bag of AI confetti.
Scope and measures are agreed for the engagement. Pricing and timing are not published here; discuss them before committing to work.