Pav Lertjitbanjong
Decision scientist. One question, worked on full time: who answers for a consequential decision once AI is involved in making it — and whether that person can defend it.
Speaking & keynotes · The 50-minute session · Write directly
- Most recentlyLeading strategic analytics at a Fortune 500 technology company
- Experience20+ years inside Fortune 500 & global organizations
- TrainingBBA Decision Science · MBA Kellogg · Stanford-trained in AI-driven leadership
For two decades, Pav Lertjitbanjong’s job was deciding what a Fortune 500 organization should measure, building the measurement systems around it, and analysing what happened when the numbers and the decisions drifted apart.
The pattern repeated in every high-stakes room: organizations move first and measure later. Most days, nothing happens. Then the stakes change, somebody asks who made the call, and the honest answer is that the decision happened and the decision-maker didn’t.
Why a decision scientist
Pressure Room begins with the human decision, not the technology. AI may be one force in the room, alongside conflicting evidence, authority, time and stakeholders. The question that survives every capability advance is whether the decision system still holds when humans carry the consequences.
Computer science, law, policy and compliance each ask important questions when AI is involved. Decision science asks what happens to the call itself.
- Is the model accurate?
- Is it lawful, and is it inventoried?
- Are the controls documented?
- Every one of these is answerable without a human being in the room
- Should a machine make this call at all?
- Whose name is on the ones it shouldn't?
- Can that person defend the call under real questioning?
- None of these can be answered on paper
A system can make a decision. It cannot be held to account for one — no legal personhood, no liability, nothing at stake. Accountability has to sit with a person. Giving that person and their organization a place to practice what they notice, decide, own and revisit is decision-science work, not tool training.
She is one of the few decision scientists working specifically on decision ownership: plainly, whose name is on the call when it goes wrong, and whether they can defend it.
The record, in the order it happened
What that means for the room you’re in
Someone who has sat where they sit.
Twenty years inside large organizations means the questions in the room come from operating reality, not from a framework — which is the difference between a leadership team engaging with a session and performing for it.
Credible in front of your own people.
Running a room full of senior reviewers only works if the person running it has held comparable accountability. A facilitator who has never signed anything consequential gets a polite hour and no change on Monday.
Nothing to sell you afterwards.
She doesn't build, implement or resell what she recommends, and there is no remediation revenue riding on what the work finds. That is the business model, not a courtesy.
On the record
Pav keynotes on the same question the firm is built around—speaking and keynotes—and the research she publishes is dated and bylined. Media coverage is listed on the press page only when a direct, checkable source link is available.
She is on stage at AI Rising in Columbus, 19–20 October 2026, and on a TEDx stage on 1 November 2026 under the title A Machine Will Never Be Brave.
The firm
PAVNESS keeps consequential decisions owned, explainable and defensible when the decision system changes — when a machine joins it, or when the person who carried it leaves. It sells one session, in two lengths, at a published price. It doesn’t build, implement or resell what it recommends, and it never certifies, approves or declares a system safe to deploy.
An AI has no legal personhood, no liability and nothing at stake. When an automated decision goes wrong, the exposure lands on a human name.
Write to us directly.
A deal, a department, an exit, or a question about the research. Anything that starts with a decision somebody has to answer for — write to us.
This goes straight to pav@pavness.com and you'll receive a response within 48 hours. Prefer the calendar? Book a 15-minute scope call.
Bring one real decision.
Bring a workflow where AI already participates and a person who signs off on it. Fifteen minutes is enough to tell whether this belongs in your organization. If it is not a fit, we will say so on the call rather than send a proposal.