Field notes from the front line of Fabric and agentic AI.
Research, playbooks, and perspectives drawn from more than a thousand engagements. Written by the people who deliver the work.
Agentic data engineering is not autocomplete
Code assistants made engineers faster at typing. Agents change what the job is. Here's what we've learned running agents on 400+ Fabric engagements.
Brian Ray, Founder & CEO. Sep 16, 2026, 9 min read
The Fabric capacity bill nobody explains
Smoothing, bursting, and throttling decide your invoice more than your SKU does. A CFO's plain-English guide to Fabric economics.
Direct Lake in production: nine lessons from 40 deployments
Direct Lake delivers import-mode speed without import-mode copies, if you design for it. What we'd tell every team before they start.
2026 State of Agentic Data Engineering
Survey of 612 data leaders on how agents are changing delivery speed, team shape, and cost.
The SSIS to Fabric migration playbook
Wave planning, conversion patterns, and parity testing for large SSIS estates.
Is your data ready for Copilot? A live readiness teardown
Our architects score a real (anonymized) semantic model for Copilot readiness, live.
Governance for the AI era: a Purview reference architecture
Domain design, labeling, and policy patterns that satisfy risk teams and still move fast.
Real-Time Intelligence for operations leaders
From historian to Activator: how manufacturers are cutting downtime with streaming analytics.
Field Notes podcast: measure AI before you scale it
Brian Ray on observational AI, token economics, and why the first metric should be cost per correct answer.
Fabric total cost of ownership benchmark
Three-year TCO across 85 client estates, compared with Synapse, Databricks, and Snowflake baselines.
One email a month. Zero fluff.
What we're seeing across Fabric programs: new features worth adopting, patterns that fail, and the economics behind both. Read by 38,000 data leaders.
Let's put agents to work on your Fabric estate.
Book a 45-minute working session with a Tulan architect. Bring your hardest data problem; leave with a plan.