What are data agents? What is the commons package?

AI has made it possible to produce a plausible-looking analysis in seconds now. Although, what happens when a manager, board member, or auditor looks at that number and asks: Walk me through how you got this number. What exact data was used? Can you reproduce it? Large language models are designed to guess the most likely next word, which makes them unpredictable. Ask the same question twice and you're likely to get two different answers, with no way to know which one is right. That's the gap between an answer that's correct and one that's verifiable. In this demo on October 28th, Sara and Simon from Posit's Core AI Team showcased commons, a new R and Python package for building analysis agents that reach for your team's trusted code first, cite trusted documentation when they can't, and clearly flag it when they have to improvise, so every answer comes with proof of how it got there. Resources: commons GitHub: https://posit-dev.github.io/commons/ Slides: https://simonpcouch.github.io/commons-demo/#/title Clinical Trials Agent: https://github.com/posit-dev/clinical-trials-agent DevRel Agent:https://github.com/posit-dev/devrel-agent AI Newsletter: https://opensource.posit.co/blog/2026-10-09_ai-newsletter/ Talk to Posit: https://posit.co/schedule-a-call