
From Banking Transformation to Wealth-Tech Product Systems
A practical translation layer from large-bank transformation programs to product systems that work in modern wealth-tech environments.
Read essayWriting
Notes from building inside banks, startups and product teams — mostly around wealth technology, AI, personal intelligence, data and the messy reality between strategy and execution.
Start here
Institutional systems, client-owned context and private intelligence — the shortest route into how I think.

A practical translation layer from large-bank transformation programs to product systems that work in modern wealth-tech environments.
Read essay
Wealth firms are building AI-ready data layers. The missing question is whether the client's own wealth context should be portable and controlled by the client.
Read essay
Private AI should not be a trust-me label. It should be an architecture claim that can be tested, inspected, and challenged.
Read essayTopic tracks
Four durable themes. Enough structure to navigate without turning the site into a taxonomy project.
Wealth technology, client engagement, open finance and the data beneath advice.
Once a personal agent can compare institutions continuously, the incumbent bank has to earn its place in the client's shortlist.
Read essayWhen two agents remember an exchange differently, both sides need evidence of what crossed the boundary and what happened next.
Read essayCRM, BI, product adoption and what transformation looks like after the slide deck ends.
Once a personal agent can compare institutions continuously, the incumbent bank has to earn its place in the client's shortlist.
Read essayAgents can reduce the searching and chasing around a banking relationship. The harder question is what should reach the relationship manager, and why.
Read essayMemory, context, relationships and the architecture required for useful personal AI.
Once a personal agent can compare institutions continuously, the incumbent bank has to earn its place in the client's shortlist.
Read essayWhen two agents remember an exchange differently, both sides need evidence of what crossed the boundary and what happened next.
Read essayPrivacy, permissions, portability and the difference between controlling data and merely deleting it.
When two agents remember an exchange differently, both sides need evidence of what crossed the boundary and what happened next.
Read essayCalifornia’s DROP is real progress, but deletion is the defensive version of control. Ownership has to mean more than removal.
Read essayFurther afield
Occasional notes on ordinary objects, places and habits that keep becoming more interesting the longer I look at them.

The leaderboard shows a name and a time. It leaves out the person who spent part of an evening deciding that the run counts.
Read essay
The family CIO is not the person who knows the most about technology. It is the person who knows how technology fits this particular family.
Read essay
The ugly spreadsheet does not look objective. That can be its most honest feature.
Read essayArchive
Older and more specific pieces remain available at their original URLs. Dates are useful here; they are not the editorial hierarchy.
Selected from Baxnet Technologies
Three current pieces that extend the territory: persistent context across changing AI systems, a practical market for personal intelligence and the privacy boundary around what an assistant learns.
Selected lead · From Baxnet Technologies
AI models are becoming easier to replace. The difficult thing to recreate is an accurate history of the person they are meant to represent.
From Baxnet Technologies
Before Personal Intelligence becomes useful to outside researchers or businesses, Mimoto can test the model itself: ask for one defined contribution, offer a fixed reward and return a better benchmark.
From Baxnet Technologies
A company chatbot can expose a real contradiction and invent a plausible confession in the same voice. The security boundary has to sit outside the model.