Make the Method Scale
VerdeWealth captures the firm's method, so software can become personal without the method fragmenting.
A distributor in Mumbai described its book to us: roughly a thousand clients. Its advisers have time to prepare individually for perhaps the top two hundred.
The other eight hundred are not neglected. They get statements, market notes and calls when something is being launched. What they do not get is why this, why now, and why for them.
Preparing for a client conversation takes more work than having it. The holdings are already in the system, the goals were captured at onboarding, the house view came out on Monday. The work is in joining those things up: what matters to this client today, and what to do about it. That is why, at this firm, the number is around two hundred.
Cheap personalised software changes the economics because the application no longer has to be one screen serving everybody.
But preparation is where the firm's judgment enters. Suppose a fund sells off and the house view is to reduce. Four clients hold it, but one is completing on a property purchase in three weeks, another would crystallise a tax charge by selling now, and a third has already told his adviser that he cannot sit through another drawdown; for the fourth, a straightforward rebalance may be right. The source and the house view are the same. What changes is the client, and therefore the decision.
Scale that with generated software and you are not just scaling a workflow. You are scaling the method, and wherever the method is unstated, the software has room to invent.
The obvious answers do not solve it.
Hire more advisers. That adds capacity, but it does not change the work required to prepare for each client.
Buy another platform. Most wealth software helps firms win clients, execute transactions or report on portfolios. Much less helps advisers work out what matters for a client between those moments. Whatever you buy also comes with its own way of doing the work.
Build it yourself. Part one showed how far a capable analyst can get with a frontier model and a text file. It can work. But it is still one analyst's interpretation of the firm's method. The firm has to decide what is authoritative, which version other software should follow, and whether what runs actually matches what was written.
Getting the firm's method into a form software can use is the work. Put its decisions side by side, surface where the sources disagree or say nothing, and put those questions back to the firm. What the firm approves becomes the specification. The adviser still decides what to send, what to change and what to discard.
Part one's twenty-day lookup is the test. When an index did not publish on the day a client bought, the firm walked backwards up to twenty days and took the last close it found. Nobody designed a product feature for that. The machinery has to find it, show it against anything that conflicts with it, put undecided parts back to the firm, and hold the software to what comes back approved.
A rules engine starts once the rules are known. The harder problem comes before that: finding the firm's actual method when it is still scattered across files, code and people, and where some decisions have never been made at all.
Once the method and its record live outside the application, the application can change without taking the method with it. Reopen a comparison produced months ago and it still renders from what was frozen at the time: the values used, the rows selected, the disclosure, and the versions of the method and data it ran against. If a figure has since changed, the original artefact remains intact, with any correction shown alongside it. The record proves what was shown then, not that the figure is still right now.
Part one described a report where ten of nineteen rows could not be reproduced from the supplied data. VerdeWealth is designed so that does not happen again: the values and context used to produce the result are preserved with it, rather than left to be reconstructed later.
That is enough to make one workflow reliable. But a firm's method spans many workflows, and the harder question is how that method carries across the rest of the firm.
The way in is a workflow, not a transformation programme. Start with something consequential the firm already does, and work through the files with the person responsible for it. Then do the same with the next workflow. Its existing portfolio, client and data systems stay where they are. Early deployments still need us in the room, but that is the deployment motion rather than the product.
Now take a case the firm has never decided. A client asks for a comparison the firm has never established as valid. VerdeWealth turns the unresolved case into structured work: what was asked, the data involved, the missing rule and the evidence around it.
Their research team resolves it. The inputs, calculations and decision logic are captured in a form the firm can review. Once approved, that decision becomes part of the specification. The next time the same kind of case appears, the firm does not start again.
A missing method decision only has to be resolved once.
That decision can now serve every relevant client in the book. The second workflow starts showing which decisions recur, and which parts of the method can carry across workflows.
The method will vary from firm to firm. VerdeWealth is designed so those differences live in the firm's specification while the machinery underneath stays reusable. The second firm tests whether that separation holds: different method, same machinery. If it does, each deployment adds customer-specific method without turning VerdeWealth into customer-specific software.
Frontier models will keep getting better at finding candidate method. That makes the first step cheaper for everyone, but it does not create the product economics by itself. Different firms will have different comparability rules, cost bases and freshness thresholds. That variation is the point: it is part of the method the firm owns. What VerdeWealth reuses is the machinery for finding where method is incomplete or contradictory, turning the firm's decisions into something the software can use consistently, and carrying them into the next client and workflow.
Each deployment should therefore make the next one easier without making the firms more alike.
Traditional platforms reuse the application and let each firm vary inside the shapes the vendor has built. What the application cannot represent does not disappear. It stays outside, in spreadsheets, prompts, code and people's heads. Part one found the firm's method distributed across exactly those places.
VerdeWealth makes that method a thing in its own right. Once captured separately, the firm can change the software around it without rebuilding the method inside each application. An incumbent can add generation and policy to an existing product. But following this all the way means letting the firm's method, rather than the vendor's application model, become authoritative.
A capable practitioner with a frontier model can already generate the software.
That does not shrink the opportunity. It is why this market is opening now.
Once firms can generate applications faster than they can state what those applications should do, the constraint moves from code to method.
The window is before today's improvised method becomes tomorrow's operating model.