Have a question or perspective on this episode? Send me a message.

AI governance is getting better at proving that controls exist. But proving a control exists isn't the same as proving it reduced the risk.

In this episode of Answerable AI, JM Wofford examines the gap between governance activity and actual risk reduction. What does a passing control really tell us? What happens when risk moves beyond the model into agents, tools, data, and business processes? And once the controls have operated, who is accountable for the risk that remains?

Because the question isn't just whether the control exists.

It's whether we can answer for what remains.

What does it actually mean when an AI system is described as compliant?

In this episode of Answerable AI, J.M. Wofford examines the limits of compliance as a measure of AI safety and performance.

An organization can complete required reviews, document controls, approve the use case, assess the vendor, and satisfy policy requirements... and still have an AI system that performs poorly or produces an unacceptable outcome.

That does not necessarily mean compliance failed.

It may mean compliance was asked to answer a question it was never designed to answer.

This episode explores:

• What the Compliance Lens actually measures
• Why compliance evidence is not the same as evidence of safety
• How organizations can mistake completed assessments for proof of acceptable performance
• Why “human in the loop” requirements do not automatically create effective oversight
• The difference between vendor compliance and responsible organizational use
• How confidence from one measurement domain can be incorrectly transferred into another
• Why compliance, capability, maturity, and risk must be measured separately
• What organizations should be able to explain when they say an AI system is “compliant”

The Compliance Lens is part of the Four Lenses of AI Measurement, a framework for separating different objects of measurement in AI governance.

Answerable AI is a Blue Narwhal podcast about AI governance, measurement, risk, compliance, and organizational accountability.

Hosted by JM Wofford

Better questions. Safer systems. Brighter tomorrows.

When AI influences a decision, who is actually accountable for the outcome?

In this episode of Answerable AI, J.M. Wofford examines the gap between involvement and accountability in AI governance. An AI system may involve vendors, technology teams, data owners, business leaders, legal, compliance, and human decision-makers, yet an organization can still struggle to answer a basic question: Who answers when the system produces an unexpected or unacceptable result?

This episode explores:

• Why distributed responsibility can become distributed ambiguity
• The difference between explainable AI and answerable AI
• Why “human in the loop” does not automatically create meaningful oversight
• Why accountability must come with real decision authority
• What organizations still own when AI is supplied by a third-party vendor
• Why governance requires current evidence, not just pre-deployment approvals
• The questions every organization should be able to answer about an AI-enabled process

Answerable AI is a Blue Narwhal podcast about AI governance, measurement, risk, compliance, and organizational accountability.

Hosted by JM Wofford

Better questions. Safer systems. Brighter tomorrows.

Enterprise AI failure is often treated as a technology problem. This episode asks whether the deeper problem is how organizations define success, measure readiness, assign accountability, and govern AI after deployment.

In the first full episode of Answerable AI, J.M. Wofford examines the research behind Why 80 Percent of Enterprise AI Fails and the organizational conditions that separate experimentation from durable enterprise use.

The discussion explores why technical capability alone is not enough, how governance and measurement gaps emerge, and why organizations can appear prepared for AI while still lacking the evidence needed to explain, defend, and take responsibility for the systems they deploy.

This episode is part of the research program behind Answerable AI: A Blue Narwhal Podcast, which examines AI through four distinct measurement lenses: capability, compliance, maturity, and risk.

Topics include: enterprise AI failure, AI governance, organizational readiness, measurement, accountability, deployment, and the difference between adopting AI and governing it effectively.

Research and related publications are available through The Blue Narwhal at thebluenarwhal.com.

Answerable AI is a research-driven podcast by J.M. Wofford exploring how organizations measure, govern, and defend artificial intelligence in practice.

Based on a four-paper research series, the show examines AI capability, compliance, maturity, risk, and accountability, and why those distinctions matter when organizations move AI from experimentation into real-world use.

Answerable AI asks a simple question: when an organization uses AI, who can explain what it does, why it was trusted, and who answers for the consequences?

A Blue Narwhal Podcast.