Notes on building AI that works on company data
How we approach gateways, data analysts, evaluation, and agents, with sources rather than slogans.
AI Agents for Business: Use Cases, Limits, and Human in the Loop
AI agents for business fit repetitive, rule-shaped, reversible work. Where they fit, where a human approves, which risks to plan for, and what to measure.
Read →AI Readiness Assessment: A Two-Week Checklist That Ends in a Decision
An AI readiness assessment works best when it ends in a go/no-go decision on one pilot. A two-week checklist for scoring use cases on value, feasibility, and risk.
Read →LLM Evaluation for Product Teams: The Test Set Comes Before the Prompt
How product teams evaluate LLM features before launch: a test set from real requests, rubrics, metrics, regression runs, and gates on cost and latency.
Read →Semantic Layer: Why Text-to-SQL Needs One Before You Can Trust It
A semantic layer defines metrics, joins, and owners once, so text-to-SQL answers match official numbers. What it is, how it's built, and how it's tested.
Read →What Is an AI Gateway? How It Keeps Sensitive Data Out of LLM Prompts
An AI gateway is one entry point between apps, staff, and LLM providers. How it routes, budgets, and logs requests, and masks PII before prompts leave.
Read →Tell us which process you want to change.
A 30-minute call is enough to tell whether AI fits, what it would cost, and where to start. We reply within one business day.