What will be — by the numbers.
Transparent clockwork. Not a black box.
Deterministic mathematical projection on signed inputs —
every gear visible, every digit reachable from the source data that produced it.
When the stakes are high enough that "the AI thinks..." is not acceptable, you need maths, not guesses. Board papers, investor decks, regulatory submissions, capacity planning -- these require numbers you can defend, from inputs you can point to, through a model you can explain.
Most predictive tooling on the market is built on generative AI. It guesses a plausible answer. It cannot tell you exactly why it answered the way it did. And under the EU AI Act, if your forecast informs a decision that affects people, you may need to explain it. Generative models create classification headaches. Pure maths does not.
wot.then is transparent clockwork. Not a black box. Every part of the modelling software is visible — every gear, every input, every step. Closed-form equations, simulations, solver-driven projections, all running over data that lives in a content-addressed graph. Every input is a signed thought. Every projection step is a because-line. Walk the forecast back to the assumptions that produced it. No generative AI means no AI Act classification headaches. Auditors love it.
The result: forecasts you can put in a board paper and defend under questioning, because every digit traces back to a signed input through a model where every gear is visible. The receipts go all the way down.
If you're running cash-flow forecasts, scenario planning, or sensitivity analysis — this is built for you. Operations leads projecting capacity, lead times, and inventory positions across multiple suppliers. Energy and sustainability teams modelling SECR and ESOS submissions, where every assumption must be an audited, signed input. Regulated industries where every projection that informs a decision must trace back to the data that produced it. The because-line from output to input is the deliverable.
One model, three scenarios, full audit trail. Board-ready in minutes.
Your financial model lives as signed thoughts in a WoT pool. Inputs -- revenue forecast, cost lines, hire schedule -- are signed by their owners. The model itself is a signed thought referencing the inputs through because-lines. Every assumption has a name attached.
You ask, by voice or by API:
"Run the cash forecast through year-end with the optimistic, base, and pessimistic revenue scenarios. Show me when each one runs out of runway."
wot.then runs three simulations in parallel. Each emits a signed result thought with because-lines pointing at the model definition, the input assumptions, the scenario parameters, and the runtime version of the solver. The output is a structured projection -- month-by-month cash position, with the deterministic model transparent from output to input.
When the board asks "how did you get this number?" you walk the because-line. Every digit is reachable from the originating data with a single graph traversal. No hand-waving. No "the spreadsheet said so." The receipt is the answer.
When an auditor asks for the assumptions behind a regulatory submission, you hand them the same graph. The because-lines are the audit trail. Nothing to reconstruct. Nothing to hope someone documented at the time.