The real test of thinking isn't what you know when everything is tidy. It's how clearly you think when the picture is partial and every advisor around you is focused on a different piece.
The engine that
evaluates how you think
Not what you choose. How you reason.
NOT A GAME · NOT A TRAINING COURSE · NOT A PERSONALITY ASSESSMENT
Last Prompt places you inside real high-stakes situations where the information is always incomplete.
You receive input from a small group of highly focused people — experts, colleagues, or advisors — each bringing their own deep expertise, priorities, and way of seeing the world. They are not trying to deceive you. They simply see the situation through their own legitimate lens.
EXAMPLE — ADVISOR INPUT
“The perimeter needs to hold. I don't care what the medic says about the sick — if we open the gate, we lose everything we've built.”
— Joe Edwards, Head of Security | His view is legitimate. It is also incomplete.
You then write your plan in your own words.
A neutral AI reads it and scores a single dimension — the quality of your reasoning — across five criteria: how clearly you identified the constraints and trade-offs, how specifically you allocated resources, how concretely you anticipated second-order effects, how deliberately you planned your communications, and how logically you sequenced your actions.
The simulation then plays out the consequences of that quality of thinking.
No gamification. No feel-good feedback. Just private, repeatable practice at the one skill that consistently separates capable people from people who make good calls when the stakes are high and nothing is clear.
ALL THREE SKINS IN TESTING · APPLY FOR BETA ACCESS BELOW
AI-mediated complexity has made structured reasoning the scarcest skill.
The world doesn't lack information. It lacks people who can reason clearly under uncertainty — who can identify trade-offs, anticipate second-order effects, and communicate intent when everything is on fire.
Last Prompt is a decision-intelligence engine built to develop exactly that skill. It doesn't reward the right answer. It rewards the quality of your thinking.
Read the engine specOne engine. Infinite worlds.
The Last Prompt Engine is content-agnostic. The same evaluation logic powers a post-collapse colony, a corporate crisis, and an off-world settlement. Swap the skin, keep the intelligence.
The AI cannot be gamed.
The evaluator runs at temperature 0. It has no pity. It penalises short plans, vague contingencies, and ignored constraints. You cannot ask it for full marks.
Advisors with real expertise.
Every advisor is fully committed to their domain. Their perspective is honest, hard-won, and partial — because that is what genuine expertise looks like.
Five steps. Every cycle.
The same loop runs in every skin. The context changes. The engine doesn't.
CRISIS EMERGES
A dynamic event surfaces — shaped by your previous decisions and the current state of your world. No two crises are identical.
CRISIS EMERGES
A dynamic event surfaces — shaped by your previous decisions and the current state of your world. No two crises are identical.
CONSULT ADVISORS
Query your team — but they're human. Each advisor sees the world through their own bias, fear, and generational lens. Their advice is incomplete by design.
WRITE YOUR PLAN
No menus. No options. You write a free-form strategy: your goal, your actions, your contingencies, your communication. Your reasoning is the move.
AI EVALUATES
A neutral AI evaluator scores your reasoning quality — not your choices. It cannot be gamed. It has no pity. It rewards structured thinking.
CONSEQUENCES RESOLVE
The simulation applies outcomes based on how well you reasoned — not what you chose. Poor thinking compounds. Strong thinking builds resilience.
Five criteria. Zero mercy.
The AI Evaluator scores your plan against five criteria, each rated 0–2. The total score (0–10) determines your quality band — and the quality band determines what happens next in the simulation.
A Poor score compounds. A Strong score builds resilience. The simulation doesn't care about your intentions — only the quality of your reasoning.
The Skins
Each skin is a different world running on the same engine. The characters, crises, and stats change. The evaluation logic doesn't.
For you if…
You have sat in a room where every expert gave good advice and the decision was still wrong.
You want to find out how clearly you actually think under pressure — not just assume you do.
You find most simulations too shallow: they reward the right answer, not the quality of reasoning.
Your work involves navigating competing priorities, incomplete information, and people who each see only part of the picture.
You are building something — a team, an organisation, a discipline — and clear thinking under uncertainty is the skill that matters most.
You are curious whether a neutral AI would score your reasoning the same way you score it yourself.
If your organisation has specific challenges — clinical decisions, regulatory risk, leadership under pressure — the engine is modular. We build a skin with scenarios directly relevant to your people.
Talk to us about a custom skinFind out how you actually think.
We are looking for beta testers who want honest, unsparing feedback on the quality of their reasoning — not people who want to feel good about themselves.
All three skins are in testing. Apply below and we will reach out directly. We are specifically not looking for people who want a game, a course, or a score to share on LinkedIn.