Questions, answered.
What we are, what the workshops cost, what we'll never do. If it isn't answered here, ask us directly — a person reads every message.
What PAVNESS is
What is PAVNESS, in one sentence?
PAVNESS is a decision science company for founders, CEOs and corporate executive teams. It runs workshops powered by Pressure Room, a living-company simulation that makes consequential judgment visible under pressure so leaders can clarify ownership, evidence, escalation and AI authority.
Are you consultants? Software? A course?
PAVNESS is the company you engage for a facilitated workshop. Pressure Room is its proprietary simulation product—not a software subscription the client must install and not a slide-deck course. Leaders make consequential decisions inside it and see where ownership, evidence, escalation and AI authority diverge.
Who is it for?
Founders and CEOs whose company still depends on their constant presence, plus corporate executive and functional leadership teams that need stronger alignment on consequential decisions. AI authority is one application, not the whole category.
Who runs it?
Pav Lertjitbanjong — a decision scientist and trusted authority on decision ownership and AI accountability, with 20+ years inside Fortune 500 and global organizations, a Kellogg MBA, a BBA in Decision Science, and Stanford training in AI-driven leadership. A TEDx speaker ("A Machine Will Never Be Brave," November 2026), she designs and facilitates every PAVNESS workshop personally.
What is decision ownership?
Decision ownership is knowing, before an AI-shaped decision moves, whose name is on the call and whether that person can defend it. Plainly: not who approved the work — four people usually did — but who decided, what they verified, and what they would say if it went wrong. It is the thing PAVNESS trains.
What is AI accountability training?
AI accountability training teaches the people who sign off on AI-assisted work how to actually review it — instead of trusting confident output. In a PAVNESS decision-simulation workshop, executives make a realistic high-stakes call with an AI recommendation in front of them, watch what moves their judgment under pressure, and agree one shared review standard for their organization.
Who is accountable when AI makes a decision?
A human, always — but often nobody has been named. A model scores, a manager forwards it, a committee notes it, and no one in that chain experiences themselves as having decided. Accountability resolves to a person only if someone was named before the decision was made, not after it goes wrong.
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.
Can AI be held accountable?
No. AI systems can and do make decisions — approving transactions, routing claims, screening applications. What they cannot do is be held to account: no legal personhood, no liability, nothing to lose. In practice, accountability resolves to a human name. That is a fact about liability rather than capability, so no model release changes it.
This is why the claim is never that AI cannot decide. It demonstrably can. The claim is narrower: it cannot be held to account for what it decides.
The workshops
What actually happens in the workshop?
Four beats. First, everyone commits to a decision in writing before the AI weighs in. Then the pressure rises: the facts change, the AI sounds more confident, time runs short — decide again. Then the reveal: who changed their mind, what moved them, and why capable people split on the same problem. Finally, the takeaway: four plain questions your team agrees to ask before anyone signs off on an AI-assisted decision.
What are the three levels?
Every level is the same decision-simulation method. Level 1 — get your people to see the risk (50 minutes, any group from C-suite to a whole department). Level 2 — prove your team can do it (2 hours, hands-on, built around the kinds of decisions that team actually makes). Level 3 — make it your standard (half day, one shared verification language across a department). The difference is how far you take it, not how many people attend.
What does it cost?
The 50-minute workshop is USD 5,000 at the founding rate — held for the first three workshops booked before 31 October 2026 — and USD 7,500 after that. The 2-hour level is USD 12,000 and the half-day level is USD 17,000. We publish prices because most firms in this category don't, and we think you should be able to budget before you talk to anyone.
Do you need access to our systems or data?
No systems access. No integration. No proprietary company data required for the session. Every case we use is invented and clearly labeled as such — which is also why we can show you the exercise before you buy anything.
What do we walk away with?
A shared four-question review standard your team agreed to out loud, the practice of naming one owner per decision before a recommendation moves, and a short report of what happened in the room — in counts: how many reversed, how many could name the owner, how many escalated.
Do you score or assess our employees?
Never. We report what happened in aggregate counts only. We never assess, score, or name an individual, and the session is never called a study.
The method
What is the THINK Protocol?
The routine a reviewer runs before signing an AI-assisted decision — the practical answer to "what does reviewed-by-a-human actually mean here?" It is taught inside every workshop, and it is the reason two of your managers stop giving two different answers to that question.
What is the Pressure Room?
Pressure Room is PAVNESS's proprietary living-company simulation product. It sits inside workshops and also has playable Rooms. You decide; the company remembers; facts, people and AI change; and the choice returns. It makes judgment visible under pressure and never scores a named individual.
How do we set an AI sign-off policy for our company?
Start with four questions, answered out loud per decision type: what may AI do here — draft, recommend, or decide subject to review; who owns the call, by name, before the recommendation moves; what reasoning stays visible — which assumption was verified, by whom; and when to escalate. A policy written this way survives a bad quarter. A PAVNESS workshop is how a leadership team agrees those answers in one session instead of six months of memos.
Is this AI governance?
No — and the difference matters. AI governance governs models. Decision ownership governs decisions. A model can be fully compliant while the decision it drives has no owner. PAVNESS works on the decision side: the humans who sign, what they verify, and whether they can defend the call.
Can we try it before booking anything?
Yes. Play the eight-week Founder / CEO Pressure Room simulation in your browser with no account, install or email. You can also use the 30-second sample or bring one of your own decisions to the two-minute decision framer.
Working with us
How do we start?
Book a 15-minute scope call. It answers one question: is this right for your company? Come with a rough idea of which team it's for and how many people review AI-assisted work. If it isn't a fit, we'll say so on the call.
Media, press, or speaking enquiries?
Use the contact form and choose "Media & press" or "Speaking & keynotes" — those go to the front of the queue. Pav speaks on human judgment and accountability in AI-era decisions, including at AI Rising (Columbus, Oct 19–20) and TEDx (Nov 1).
Do you guarantee results?
No — and be wary of anyone in this category who does. We never promise a saving or a business outcome. What we commit to is what the workshop observably delivers: a shared review standard, named ownership, and a report of what happened in the room.
No matches for that — try a shorter word, or send us the question.
Fifteen minutes answers the question this page can't.
Is this right for your company? Bring a rough idea of which team it's for and how many people review AI-assisted work. If it isn't a fit, we'll say so on the call.