What we don't yet publish, and why

Every vendor in this category publishes what its system can show. This page is the other list. It exists because a governance layer that reports only its strengths is asking to be taken on trust. That is the exact thing it claims to replace.

Not published

Live performance of the underlying decision engine

The platform generates investment suggestions, and their performance is measured. Those figures are not published, because the measurement work behind them is not finished to a standard we would defend in front of someone competent. Several known defects in the input data are documented internally and are being corrected in sequence. Publishing a number produced from data we know to be flawed would be worse than publishing nothing.

Not claimed

Any advance through the autonomy levels

The gate machinery has been running since May 2026 and has never returned a pass. As of 26 August 2026 that was 12,786 evaluations. More precisely, it has never yet computed a metric to compare against a threshold, because the trade history is far below the sample floors the conditions require. The system sits at the entry level it was enrolled at. That is what a gate is supposed to do when the evidence is not there, and it is stated here rather than omitted.

Not measurable

Cost-adjusted return of the signal set

We tested whether one class of signal survives transaction costs and concluded the question is not answerable at our current data volume. The daily dispersion is far larger than the effect being measured, and separating them would require thousands of trading days. The honest status is not measurable yet, not unproven, and certainly not positive.

Not offered

Certification of a customer's own metrics

The governance kernel is portable in the sense that another firm's measures could be written into it. It is not portable in the sense that we could certify results we have not defined and cannot verify. Those are different claims and we make only the first.

Not yet built

Integration with a system we did not build

The governance layer is designed to be separable from the decision engine, and the integration shapes that would make that real are scoped rather than built. Until they exist, this platform governs its own engine and nothing else. We would rather say that here than describe an integration path as though it were available.

Why this page exists

Two reasons.

The first is regulatory. The CSA notice spends real space on the opposite failure: inaccurate, embellished or misleading claims about AI capability, in marketing materials and client agreements, implying an advantage that is not there. A vendor selling an evidence layer is the last party that can afford an unsupportable claim about what its own system captures.

The second is simpler. If we will state this about ourselves in public, the record the platform produces about a decision is more likely to be worth reading.

Current status, plainly

In development. Implementation began April 2026; design and architecture since November 2025. Not incorporated. No customers, no revenue. Discovery conversations underway with securities-registered firms in Canada.