“Here’s how it will work.”
Launch our new platform.
Reach 100 paying customers in six months.
You know what you want your project to achieve. What might you be overlooking?
Talk it through with an AI thinking partner. Step into a future where it failed, and find what you can change today.
New here? Sign in with your email. We send a 6-digit code, no password. Looking around is free, and you buy a project inside when you are ready: $9.99, paid once.
Launch our new platform.
Reach 100 paying customers in six months.
People signed up, but never became active.
We built before testing whether anyone would pay.
Same project. A different perspective.
The same six steps you see in the app. One made-up project runs through all of them: a platform for small businesses that wants 100 paying customers in six months.
Describe the project in your own words. The AI assistant asks about what is still unclear, and you confirm the project details before anything else happens.
AI assistantWho decides to pay: the shop owner, or the person who uses it every day?
What must exist for the goal to work? What does each of those need? Keep going until you reach first steps you could start this week. The Recipe is drawn as you build it, and the AI assistant drafts levels when you ask.
100 paying customers who use it regularly
An ingredient is ready when everything under it is ready.
Write the failure as if it already happened. Add as many failed scenarios as you like, and the reasons under each. The AI assistant suggests more when you ask. What several scenarios share becomes a failure vector. Then do the same for the alternative, because waiting has a cost too.
300 sign-ups and 12 paying shops. Most never finished setup.
Failure vector · Shared by 2 scenariosEverything waited on the two founders.
With both futures in front of you, decide: go ahead, go ahead with changes, try a smaller version first, wait, or do not do it. Say why, in your words. Then pick a date to look at the decision again.
Prevent it, test it early, stop doing something, or decide to live with it. The page shows what still has nothing against it. You can also send a link so the people around you write their own pages, with no names attached.
Every step says what it is for and how you will know it is done. Sensors name the warning sign to watch for. When you report one, you get an email with what you said you would do.
Start with what you want to achieve.
A sample project, not a prediction or a live AI session.
It has not failed yet.
There is time to change the plan.
Not a blank page and some encouragement. A guided exercise that ends in a plan where every reason you found has something against it.
You write first. The AI assistant drafts only when you ask, and nothing enters your project until you accept it. Every suggestion you take stays marked as the AI's, so you always know which thoughts were yours.
You work through going ahead and it not working, then through taking your alternative instead. Deciding to wait is a real answer, and the exercise is built to let you reach it.
Send a link. Each person writes one page, about a scenario you choose or one of their own, and no page records who wrote it. You see where separate people raised the same thing, which is the part worth deciding on.
When you report a warning sign, you get an email with what you said you would do. On the date you chose, it asks what happened and whether the decision still holds.
No subscription, nothing to cancel, and no charge for thinking about it first.
$9.99
for one project, start to finish
A project you have begun is yours. Archiving is free. Delete one before you write anything in it and the payment stays with you for the next.
Start a projectOrganisations can be set up with their own price. Write to hello@premortemlab.com.
Every exercise ends in a record like this one: your decision, in your words, with the concerns that shaped it and what you said you would do. Print it, keep it, or come back and change your mind with the reasons still in front of you.
A sample. The project below is made up.
Proceed: run the repair table for one season, two evenings a week, starting in October.
A weekly repair table at the library, for people who live within a mile. Success looks like neighbours fixing things they would otherwise replace, and four volunteers who still want to be there in January.
Whether a second evening is worth it. Nobody knows how many people will come to the first one, and that answer is three weeks away.
Your own record is private to you, and you can download or delete it at any time.
When you’re invested in a plan, doubts can be difficult to articulate. A premortem gives you room to explain a different outcome.
Imagine the project has already failed. Working backward from that future can help bring overlooked concerns into the conversation.
Imagining it as settled is the part that works. Klein reports that prospective hindsight, assuming the outcome has already happened, increases how well people identify reasons for it by about 30 percent, citing research by Mitchell, Russo and Pennington published in 1989.
Based on the premortem method developed by Gary Klein. Read the original article ↗
The usual question“What might go wrong?”
The premortem question“It’s six months later.
The project failed.
What happened?”
Bring the idea, the uncertainty, or the decision you keep turning over.
Start a projectAlready have an account? Use the same email to sign in.