Why most director self-evaluations are bad for good governance — and how to fix it

Director self-evaluation is a governance best practice. Boards do it annually. Committees sign off. Boxes get ticked.

Unfortunately - to the detriment of good governance - most of it is theatre.

It's not because the people involved aren't serious. Not because the intent isn't genuine. But because there are three structural problems embedded in the process that standard tools have never been built to solve. After 25 years working in governance, here's what they are — and why a genuine alternative is now possible for the first time.

Problem one: most directors are not governance experts

This sounds harsh. It's meant to. Most directors are excellent at what they built their careers doing — finance, operations, law, medicine, engineering. Governance is a discipline in its own right. It has its own body of knowledge, its own failure modes, its own nuances.

Asking a director to self-assess their governance effectiveness without that foundation is like asking a brilliant surgeon to evaluate their own courtroom performance.

The subject-matter expertise doesn't transfer. You end up with confident ratings that measure almost nothing.

"The problem isn't that directors are unqualified people. It's that governance is a separate qualification — and most self-evaluation tools assume it isn't."

Problem two: self-evaluation is structurally biased

Even genuine experts struggle to evaluate themselves accurately. We are biased toward our strengths. We rate ourselves on intent, not impact. We don't know what we don't know — which means the gaps that matter most are often invisible to us.

Standard director self-evaluation tools do nothing to address this. They ask you to score yourself. They take your answer at face value. They aggregate the responses and call that insight.

It isn't insight. It's a summary of self-perception — which is a very different thing.

"Bad boards don't usually know they're bad boards. The self-evaluation process, as currently designed, is not built to tell them."

Problem three: the technology couldn't do better — until now

Here's the problem that rarely gets named: the industry's reliance on scoring scales wasn't just a design choice. It was a technical ceiling. Narrative responses — the ones that actually reveal how a director thinks — couldn't be interpreted consistently at scale. So the industry built around the limitation and called it methodology.

Think about what a number can't tell you. A director who scores themselves four out of five on strategic oversight but can't describe a single moment they pushed back on management has revealed something critical. A director who rates their relationship with the CEO as strong but hedges every answer about accountability has shown you something important. Scoring buries both. It has no mechanism to notice.

AI changes that. Not because AI understands governance — it doesn't. But because it can interpret how someone reasons through a governance question: where the thinking is thin, where confidence is overclaimed, where the gap between stated position and demonstrated reasoning is widest.

That capability didn't exist at scale before. The technology ceiling is gone. What replaces it depends entirely on the quality of the governance expertise behind the tool.

What a meaningful director self-evaluation actually requires

Genuine governance self-evaluation needs to do three things that standard tools don't:

First, it needs to ask questions grounded in what governance actually requires — not generic leadership competencies dressed up in board language.

Second, it needs to interpret narrative responses, not just scores. How a director explains their thinking often reveals more than any number on a scale.

Third, it needs to surface what the director isn't seeing — the gaps, contradictions, and blind spots that are invisible from the inside.

The first two requirements could always be met in theory, by a skilled practitioner in the room. The third — at scale, consistently, without a consulting relationship to protect — is what's genuinely new.

AI can handle all three - this is a breakthrough

I've written frameworks from the past. Consulted on hundreds of board evaluations. Watched the same comfortable process produce the same comfortable results for decades.

At some point, writing about the problem stopped feeling useful.

So I encoded 25 years of governance practice into a director self-evaluation tool — and used AI to give it something no human consultant can offer at scale: the ability to interpret open-ended responses the way a senior practitioner would, with no incentive to be kind about what it finds. The governance expertise is mine. The AI is the delivery mechanism — one that works at any hour, for any director, without a consulting relationship to protect.

It's not necessarily a comfortable tool. It's not designed to be. It will call you on your BS. If you want to know how you're actually doing as a director — not how you feel about how you're doing — that's what it's for.

Try it and see the difference AI makes

Try this AI-enabled Director Self Evaluation. It's free and takes 20–30 minutes if you engage with it seriously.

I'm particularly interested in hearing from governance professionals, institutional investors, and directors who've been through conventional evaluation processes and found them wanting. The goal isn't to sell a product. It's to raise the floor on what director self-evaluation can actually do.

Jim Crocker is an AI governance consultant and board director. He writes about what boards and senior executives need to know about AI at jimcrockerai.com. Here is Jim's LinkedIn profile.