Blog

Notes on data, dbt and AI.

Field notes on data warehouses, dbt, governance and AI — written with the same discipline I use to write SQL: clear, tested, and free of fluff.

May 29, 2026· AI consulting· AI

AI consulting for Brazilian mid-market companies: where to start without burning the budget

AI consulting for mid-market companies is rarely about training a model. It's judgment plus a data foundation plus one narrow, clear-ROI case before you spend.

Read post

May 13, 2026· AI· AI agents

AI agents for small and medium businesses: where to start, and why most pilots stall

Only 23% of organizations are scaling agentic AI, and ~95% of GenAI pilots show no P&L impact. For an SMB, the realistic starting point is one narrow agent on cleaner data, not transformation.

Read post

May 1, 2026· dbt· analytics-engineering

Analytics engineering for small teams: the six stages most teams skip past

A staged maturity model for analytics engineering for small teams, with the operational signal that tells you it's time to move and the failure mode at each step.

Read post

April 29, 2026· AI· SaaS

AI adoption in small SaaS product teams: what Gartner, McKinsey, and DORA actually measured in 2025

Eighty-eight percent of organizations report using AI; only 5.5% see real returns. For small SaaS product teams the gap is sharper, the failure mode faster, and a non-trivial fraction net out positive. A walk through the 2025 surveys on adoption, productivity, and what determines which side of the curve a team lands on.

Read post

April 29, 2026· AI· Brazil

Brazil's AI adoption boom in public numbers: what IBGE, Bain, Gartner, and ANPD actually say

Brazilian industrial firms went from 16.9% to 41.9% AI adoption in two years. The same public datasets say almost nothing about the foundation those models are reading. A walk through what is measured, what is missing, and what it implies for 2026.

Read post

April 28, 2026· AI· Governance

Why your company needs to fix its data before adopting generative AI

Generative AI is only as good as the data feeding it. If your KPIs disagree with each other, your copilot will too. Here's how to think about the order of operations.

Read post