This is the last post in the build-log series, and it closes it into the present. So here is the plain version: the product is pre-launch. It is not finished. And we would rather tell you that than have you find it out.

Why say it at all

Most software is sold as though it were already complete. The marketing describes the finished vision; the roadmap items are written in the present tense; the gaps are left for the customer to discover after they've committed. It is a small dishonesty, repeated everywhere, and it is exactly the kind of thing that erodes trust with the sort of buyer who reads carefully.

Solicitors read carefully. Overstating what's built is not a growth tactic with this audience — it is a way to lose them the first time reality diverges from the pitch. So we don't do it.

The capabilities page tells the truth

There is a capabilities page that lists every concrete action in the product, each with its real status — what's shipped, what's partial, what's still on the roadmap. That status is checked against the code, not against our hopes. When something says shipped, it means it is in the software and behaves as described. When something is partial, it says partial and describes honestly which part is done and which isn't. When something is planned, it is not written up in the present tense as though you could use it today.

You saw an example of this in an earlier post: the self-learning layer's opt-in toggle and local journal are live, while the piece that would have carried pseudonymised behaviour to a central learner was built as a proof-of-concept and then declined — we are not shipping it. We said so there, and the capabilities page says so too. That is the standard for the whole list — no roadmap item dressed up as done.

Where the product is now

The core is real and running. Matter data lives in your browser, not on our servers. The privilege gate sits in front of every outbound AI call — soft-warn where the law allows judgement, hard-block where the duty is absolute. A human is required on every consequential step; the software does not act alone. UK and PL are native builds, walled apart on one shared engine. The populated demo lets you watch a fictional matter go through the whole thing, no card required.

Those are not aspirations. They are what the demo shows and what the code does.

What's next

More markets, built the same native, walled way — Germany, Spain and the Netherlands are live, each written natively rather than translated from the ones that came first. Depth on the parts marked partial. The cross-firm self-learning loop is NOT among them — we declined to build it, and the layer stays on the firm's own disk. And the ordinary work of hardening a product toward launch, done in the open, with the capabilities page kept honest as it goes.

The launch window matters here too: firms joining before the end of August 2026 get their monthly AI credit allowance doubled — the price stays the price, the room to work gets bigger. The pricing page has the figures, and the capabilities page has the status.

That is the whole of it. Built by solicitors, for solicitors — not a generic tool with a legal skin. Pre-launch, honest about it, and shown rather than told. Thank you for reading the series.

← All posts