Evidence-Based AI Governance: Beyond Marketing Claims with Petar Tsankov

We're not ready for high-risk agentic systems.

Petar Tsankov, CEO of LatticeFlow AI, delivers the hard truth about AI governance, and explains what evidence-based governance actually means.
In this episode of The Digital Analyst, John Santaferraro explores why 600 billion in global AI investment only returned 100 billion in production value. Petar explains the critical distinction: data governance ensures trustworthy inputs, AI governance ensures trustworthy outcomes.

Discover why even perfect data quality doesn't guarantee trustworthy AI, and what evidence-based governance means beyond marketing claims. Actual benchmarks, measurable results, and technical assessments instead of opinions. Learn why AI governance must be implemented in the tech stack, not on paper, and why it's closer to cybersecurity than compliance documents.

Explore the challenge of governing multi-agent systems where failures cascade and multiply, and why autonomous decision-making dramatically increases risk profiles. Petar shares why organizations struggle to translate principles like "performance" into measurable technical controls, and how automation will make governance seamless rather than burdensome.

**************
Petar Tsankov
https://www.linkedin.com/in/petartsankov/
https://latticeflow.ai/

**************
John Santaferraro
https://www.linkedin.com/in/johnsantaferraro/
https://www.ferraroconsulting.com/
http://www.thedigitalanalyst.ai/

**************
The Digital Analyst is powered by Insight Jam
Join the human conversation around AI at https://insightjam.com/

**************
0:00 Cold open and Introduction
3:10 Meet Petar Tsankov: From AI Safety to AI Governance
4:06 Data Governance vs AI Governance: Trustworthy Inputs vs Outputs
6:12 The $600 Billion Investment Gap: Why AI Isn't Reaching Production
10:13 Why Perfect Data Quality Doesn't Guarantee Trustworthy AI
12:19 What Is Evidence-Based AI Governance?
17:37 Technical Assessments: Measuring What You Can Actually Trust
20:31 Operationalizing AI Governance: Beyond Pre-Production Assessment
23:05 The Agentic AI Challenge: Autonomous Decisions and Cascading Failures
27:02 Why Orchestration and Governance Must Work Together
29:43 The Vision: Implementing Governance in the Tech Stack, Not on Paper
34:00 Where to Start: Building Your AI Governance Foundation
38:02 Connect and Subscribe