Features
Load-bearing assumptions: the few things your idea depends on
An idea rests on dozens of assumptions. A handful decide the outcome. The report finds them and says how well each one is evidenced.
The problem
Ask a founder what their idea depends on and the answer is usually a list of features. Ask what has to be true about the world for the idea to work, and the list gets longer and less comfortable: customers have to feel the problem strongly enough to switch, the price has to cover the cost of reaching them, the channel has to keep working after the first hundred sales. Most of those statements are untested, and they are not equally important. Spending the same effort on all of them wastes the time you have.
How it works
The engine first breaks your idea into nine claim links: who the customer is, the job they are trying to get done, what they do today, why your product would cause a better outcome, the value it creates, the price, the path to the customer, the cost to serve and what persists. From those links it writes down the assumptions each one depends on.
It then simulates the outcome you asked about and measures how much each assumption moves it. The five to eight assumptions that carry most of the uncertainty are marked as load-bearing. Each one carries a grade for its evidence, so the report can separate two kinds of risk: an assumption that matters a great deal and is well supported, and one that matters a great deal and rests on nothing yet.
A chart in the report places every assumption by how much it carries against how well it is evidenced. The corner where an assumption is both heavy and weakly evidenced is where the report suggests you spend first. Assumptions that barely move the outcome are shown too, so you can see they were considered and set aside.
Who uses it
Founders use the list to decide what to test before building. Incubators and consultants use it to see, for each idea they review, which claims carry the plan and how much is known about them.
Limits
The assumptions come from what you wrote. If your description leaves out part of the plan, the report marks those claim links as empty rather than inventing them, and an assumption that was never stated cannot be ranked. Today the base figures behind each assumption come from the AI model's own knowledge, and the report files them on the lowest rungs of its evidence ladder. Linking assumptions to published data you can open and check is planned.
Related features
Questions
How many assumptions does the report list?
It lists everything the engine extracted from your idea, and marks the 5 to 8 that the outcome depends on most. Each one carries a grade for its evidence.
Can I see why an assumption counts as load-bearing?
Yes. The report shows how much of the uncertainty in the outcome each assumption carries, so the ranking is something you can check, not a label you have to accept.
