Google finally measures AI-search visibility. Here is what it still cannot tell you
July 17, 2026
Google now reports generative-AI search impressions. Learn what the new data proves, what it omits, and how businesses should measure AI visibility.
Blog
Notes on shipping AI features into production software. Infrequent, specific, written to be read once.
AI agents need approval limits, not better prompts
July 10, 2026
Prompts express intent. Production systems still need to enforce what an agent may do, for whom, under which limit, and with what approval. Procurement already has the operating model.
How I got my own sites cited by ChatGPT and Perplexity
June 30, 2026
My sites get named by AI engines when people ask for recommendations in my categories. The method, what the published evidence says actually mattered (Bing indexing, direct answers, third-party mentions), and the popular tactics that did nothing.
Extracting structured procurement fields from RFA emails with Claude
June 30, 2026
Build notes from Clotho, a live AI procurement intake tool: Claude Sonnet extraction with a human review step, a pg_trgm pre-filter for supplier matching, and an LLM rerank on the shortlist. Why the database searches and the model judges.
Why the WordPress AI-readiness audit is the right first step for mid-market companies
May 27, 2026
What AI-ready actually means for a 50-200 employee company's WordPress install, and why a flat-fee audit gets better results than hourly engagements.
VPAT 2.5 in 3 days for $2,500: what it covers and who it is right for
May 27, 2026
Big accessibility firms charge $25k for a VPAT 2.5 because their cost structure demands it. Mine does not. Here is what the $2,500 deliverable covers and who it is right for.
How to evaluate an AI consultant in one phone call: 5 questions that filter for actual builders
May 20, 2026
Five questions that separate AI consultants who have shipped production systems from those who have built demos. Ask these before you sign anything.
Why your first AI use case shouldn't touch sensitive data
May 7, 2026
Regulated-industry AI buyers are stuck in security reviews instead of shipping tools. The fix is to skip sensitive data entirely on the first build, using procurement as the wedge. A 90-day expansion path that mirrors how regulated industries adopted cloud between 2014 and 2019.
Why most enterprise AI pilots stall in month two
May 6, 2026
A 30-day pilot that ships is the easy part. Three specific failures in month two quietly kill more enterprise AI projects than bad models or bad vendors.
Indianapolis enterprise AI: first pilot in 30 days
April 29, 2026
A 4-stage playbook for getting your first internal AI pilot live in under 30 days, built for the mid-market companies that anchor the Indianapolis economy.
What overnight AI research can't see
April 21, 2026
Autonomous AI research produces confident plans on public facts. It can't see the decisions you've already made in private. A short piece on the load-bearing assumption that has to be dumped into the prompt before any overnight session starts.