I manage finances for vulnerable adults professionally. I can't reliably pay my own bills.

That contradiction isn't embarrassing. It's the entire thesis.

The dual position

I'm Matthew Grant. I work in financial services, managing money for vulnerable adults — people who, for various reasons, cannot manage their own finances. I understand regulatory frameworks. I understand budgeting, cash flow, debt management, savings discipline. Professionally, I'm competent.

Personally, I was diagnosed with ADHD at 43. The diagnosis explained decades of the same pattern: I understand money perfectly. I set reminders. I intend to pay bills on time. Then I don't — not because I forgot the principle, but because my executive function failed to execute it at the right moment.

Late fees. Overdraft charges. Emergency borrowing for expenses I knew were coming. Impulse purchases during hyperfocus episodes. The irony of helping others manage money while failing to manage my own was crushing.

I tried every compensatory strategy available. Budget apps. Reminders. Accountability partners. Spreadsheets. Standing orders. All of them required the one thing I don't reliably have — consistent executive function.

So I stopped trying to fix my brain and started building the infrastructure it actually needs.

The insight nobody else has

Most founders approach this problem from one direction. They either have lived experience of ADHD and build from personal frustration, or they have technical capability and build from market analysis. Neither position alone reveals the structural failure.

Working professionally with vulnerable adults under formal financial supervision, I see what happens when someone's lack of capacity is formally recognised — the system steps in and manages on their behalf. Working personally with my own ADHD, I experience what happens when executive function is informally insufficient — the system does nothing. You're expected to cope.

The gap between those two positions is where FLI lives. Ten million UK adults need execution support but don't meet the threshold for, or indeed want, formal intervention. They fall into the ADHD tax trap — too capable to qualify for help, too impaired to manage consistently without it.

Every safety framework assumes the user can act on its recommendations. Standard safety measures — requiring confirmation, presenting options, seeking approval — demand precisely the executive function we lack. I went back to first principles and recognised that safety mechanisms themselves can be the barrier. The solution is post-execution explanation with constitutional constraint, not pre-action approval.

Why this wasn't possible three years ago

The timing is not accidental. Four things matured within the past 24 months that make this viable now.

Constitutional AI

AI systems capable of operating within strict decision boundaries while maintaining contextual reasoning. Three years ago, you could build a chatbot or an automation. You couldn't build a system that reasons within constraints — executing autonomously while provably never generating novel financial strategy. Now you can.

Open Banking maturity

UK Open Banking has passed 15 million users. Payment Initiation Services are technically proven and regulatory frameworks are established. The infrastructure for an AI agent to move money within your own accounts, under your delegated authority, exists now. It didn't two years ago — not at this reliability.

FCA Innovation Pathways

The Financial Conduct Authority created a structured route for novel financial services to develop under regulatory guidance. This is the mechanism that allows a system like Syndic — autonomous financial execution for vulnerable populations — to be built within a regulatory framework rather than outside one.

NICE Digital Therapeutics

For the first time, there's a pathway to validate software as a therapeutic intervention for recognised conditions. Syndic can be assessed not just as a financial product but as a clinical tool — measuring outcomes against the Weiss Functional Impairment Rating Scale, pursuing the same evidence standard as pharmaceutical interventions.

Build this with me

One founder. One thesis. One product that needs to work before anything else gets built. If that resonates, join the waitlist.

Join the research waitlist

No cost. No commitment. Beta access subject to FCA authorisation.