The Certainty Engine.
A deterministic prediction system that anticipates a specific arrival, watches the prediction converge, and records the exact moment it becomes fact.
From probability to proof
Almost everything we know about the physical world is inferred. Sensors count bodies, models estimate attention, algorithms guess at attribution. The output is always a probability — useful, but never certain, and never auditable at the level of a single event.
The Certainty Engine takes a different path. Rather than sampling a crowd and estimating behaviour, it establishes who is expected, narrows that expectation through sequential confirmation, and produces a single deterministic event when arrival is certain. The result behaves like a confirmed digital action — except it happens in physical space.
The distinction matters commercially: a probability can be argued with. A confirmed, auditable event cannot. That is what changes the value of the measurement.
Four stages, one outcome: certainty.
A genuine sequence — each stage depends on the one before it.
Pre-assignment
Identity is established before the interaction begins, on a consent-based, token-only basis. No cameras, no personal data in the decision path.
Prediction
The system forms a time-bounded expectation of arrival and holds it as a live, narrowing window rather than a static guess.
Convergence
Sequential confirmations tighten the window. Accuracy compounds across venues and visits — the data flywheel at work.
Confirmation
A deterministic event fires the instant arrival is certain, producing the auditable record everything downstream relies on.
Certainty and privacy, at the same time.
Most measurement systems trade privacy for precision. The Certainty Engine is built so it never has to. It operates camera-free, on consent, with no personal data in the decision path — identity is handled as a token, not a profile.
That isn't a compliance afterthought; it's the architecture. As data-protection expectations tighten across the UK and EU, an approach that is private by design turns regulatory pressure into a durable advantage rather than a liability.
What the system will not do
- No cameras, no facial recognition, no biometric capture
- No personal data in the measurement decision path
- No estimation where confirmation is possible
- No lock-in of the evidence away from the partners who need it
An AI capability that gets sharper with scale.
Prediction accuracy improves with data volume across many locations. Each venue makes the next one better — which is why this is a compute-hungry AI system, not a fixed sensor product.
Data volume → accuracy
More venues and visits sharpen the prediction models, so measurement quality rises as the network grows rather than plateauing.
Built to train
The roadmap is bounded by model development and compute, making supercomputing access — not more hardware in venues — the accelerant to full readiness.
One core, many markets
The same deterministic capability applies across retail, measurement, access control and crowd safety — a platform, not a point solution.
The detail goes deeper.
This page describes the capability. The underlying method and patent estate are available to qualified partners and investors following IP-counsel review, under NDA.
Request a technical briefing →