From AI-Ready to AI-Reliable: Building Semantic Views and Debugging AI Analysts

In 2026, Datalere will host Denver Data Dialogues, an in-person networking event for data practitioners and analytics leaders featuring guest speakers, panels, catered food, and more. A reboot of our former Data Science in Colorado meetup, the series will explore real-world conversations around data strategy, architecture, AI readiness, and analytics leadership, with practical insights you can take back to work. Events are held on the second Tuesday of every month at the Washington Street Community Center.

November: From AI-Ready to AI-Reliable: Building Semantic Views and Debugging AI Analysts

Our November session features Datalere data engineers Gil Platt and Rob Browning. They’ll lead a hands-on, two-part look at what it takes to get trustworthy answers from AI analysts, from preparing your data to tracking down the problem when the AI gets it wrong.
AI analysts promise plain-English answers from your data, but those answers are only as good as the models underneath. When a business user gets a wrong number, the cause could be anywhere between the raw data and the SQL the AI wrote.

In Part 1, Gil will walk through a layered dbt project (staging, refined, and datamart) and show how to turn those models into a Snowflake semantic view that an AI analyst can query. He’ll do this in a live demo on synthetic data, going from source all the way to a working semantic view.

In Part 2, Rob will pick up where Gil leaves off by deliberately breaking the semantic view Gil just built. He’ll then work through a four-step framework for debugging AI output, checking the generated SQL, the semantic view, the dbt model, and the raw data in turn, and fix the problem live.
You’ll leave with a clear picture of what “AI-ready” means in practice and a repeatable method for diagnosing bad AI answers. No deep dbt or Snowflake expertise is required.

As always, there will be time for discussion and Q&A.

We can’t wait to see you there!

About the Speakers

Gil Platt is a data engineer at Datalere who builds ETL solutions and data models. He has worked across defense manufacturing, health insurance marketplaces, pharmaceuticals, fire districts, and banking. He starts every project by listening to business users and stakeholders, then turns their requirements into solutions that unlock the value of their data. His experience in regulated, high-stakes environments shows in his pipelines and models, which are accurate, well-governed, and clear enough for the people who rely on them to trust and use.

Rob Browning is a data engineer at Datalere with a strong background in business intelligence. He designs cloud data platforms and analytics solutions that help organizations turn complex data into practical insight. His work covers data modeling, reporting architecture, pipeline development, and AI-enabled applications. That includes helping build a chatbot that lets business users query their data in natural language through semantic modeling and interactive analytics.