Posit’s Blueprint for Responsible GenAI

Posit has been on its own journey with GenAI, starting from a position of total skepticism and moving toward a disciplined approach we now use with customers. Our CEO, Tareef Kawaf, recently shared that journey in a talk at a life sciences industry conference. While the talk was built for that audience, the questions it raises apply no matter what industry you're in:  

• How can organizations use GenAI tools responsibly?
• What has to be true for that to actually work?
• How do you build the right checks and balances for quality, validation, and reproducibility?

The Prime Directive: Why Code-First Data Science Matters

Tareef Kawaf traces Posit's philosophy back to John Chambers' "prime directive" — the idea that trustworthy analysis isn't enough; it has to be demonstrably trustworthy.

The Two Strands: Technical and Human Dimensions of GenAI

How the organizational challenges of AI adoption and the human experience of change are inseparable.

The Five Objections to GenAI in Regulated Data Science

A direct response to the most common concerns teams raise when considering AI in regulated workflows.

The Four Principles of Responsible AI

The non-negotiables for using GenAI in a way that satisfies the prime directive: code first, human in the loop, governed environments, and data privacy.

Mindsets, Not Skills for Navigating AI Change

Tareef Kawaf on the three dispositions (curiosity, customer centricity, and courage) that determine whether you adapt to AI or freeze.

Live Q&A: Tareef on Governance, Open Models, and What Comes Next

Tareef fields audience questions on Positron vs. generic AI tools, closing the skills gap, open-source workflows, and what new roles look like as work gets reshuffled.

• 0:36 — Positron's advantage over generic AI tools (VS Code + Copilot, Claude Code), plus governance/audit logs/GRC
• 3:52 — What skill gap is blocking companies from embracing AI
• 6:04 — The pivotal moment: engineers experimenting with LLMs over Christmas break
• 7:53 — On protecting focus time and avoiding distraction
• 8:54 — From early AI adopter to writing code by hand — where should this person start?
• 11:52 — Setting up adversarial agents in a workflow
• 14:02 — Where AI hesitancy comes from, and how to overcome it (data retention, security)
• 18:01 — Common open-source skills/workflows across domains vs. each org building its own
• 19:02 — Where open models are headed (cost, capability, frontier vs. open)
• 22:39 — Why coding skills still matter even as LLMs iterate and improve
• 23:43 — Building a business case for reproducible, code-first workflows
• 25:22 — New roles as GenAI reshuffles work (the Reshuffle book, airline automation analogy)

Resources Mentioned

📖 Reshuffle by Sangeet Paul Chaudhary — the book Tareef references throughout the Q&A on how AI unbundles and rebundles work
→ reshufflebook.com

🔗 Posit AI
→ posit.ai

Want to keep the conversation going?
Schedule a call with our team →