How I lost £1.2M trading — and what retail brokers taught me on the way down.
I lost £1.2M of my own money trading gold and bitcoin over three years. This is not a redemption arc. It is a technical dissection of the machinery that took it, and why I now build AI systems inside the firms on the other side of that trade.

The number is £1,247,000. It went in blocks. Some to gold on leverage, more to bitcoin on leverage, a final tranche to a strategy I still think would have worked with a fifth of the size and a tenth of the emotion. I know the exact figure because I have looked at it every day for two years while I built the business you are reading now.
This note is not a redemption arc. It is a technical dissection of the counterparty. Because in the process of losing that money to five different retail brokers across three jurisdictions, I learned more about how modern broker machinery actually works than I did in the decade of tech work that preceded it. That knowledge is the reason KJ Capital's broker practice exists, and this note is the honest version of the story.
The four things brokers did that I would now build for them
Every broker I lost money at was doing four things well. Not perfectly — some were sloppy — but well enough that a client with my psychology profile was always going to lose. Retention outreach that hit at exactly the wrong moment. Deposit funnels that recovered the exact abandoned session I had walked away from. Dealer-desk routing that got sharper the deeper I got into the book. And a compliance envelope thin enough to be plausibly deniable but thick enough to survive a supervisor's ask.
None of these are villainous. They are what a well-run broker does. The problem is not the machinery; it is the machinery meeting a client who should not have been trading at that size. And the reason I now build these systems for brokers is that the ones that skip them are worse — they lose the informed clients and keep the dependent ones by accident. That is a worse outcome for everyone.
The machinery, honestly named
- Dealer-desk routing that read my flow within the first fifty trades and priced me accordingly for the next five thousand.
- Retention outreach on my two worst weeks that recovered a deposit I had already emotionally written off.
- Deposit funnels that failed over from a card decline to an e-wallet in under twelve seconds — I remember the timing because it was the moment I should have walked away.
- IB and affiliate channels that recycled me across three brands in two years without a single one of them realising it (or admitting they realised it).
- Compliance envelopes that ticked every box while never once slowing the machine at a moment it would have cost me less.
Why I now build AI for the firms on the other side
The obvious reaction to the story I just told is to hate the counterparty. I do not. The reaction I had, after eighteen months of processing, is more useful: the same machinery that took me out at scale is the machinery that lets a well-run broker survive the compliance environment, price risk accurately and treat vulnerable clients better than they are treated today. AI is what makes it work. AI with a compliance envelope designed first is what makes it survive.
Every broker I now work with knows the founder story before we sign anything. It is not a sales asset; it is a disclosure. It means I understand the psychology on the client side better than most consultants and the machinery on the broker side better than most vendors. It also means I have a specific view of what good looks like — a broker that would not have taken me out at that size, because their retention model would have flagged me on week six and their compliance layer would have made the intervention non-negotiable.
The machinery took me out because it was good, not because it was evil. Building better machinery is the redemption arc.
What we do differently at KJ Capital
Every KJ Capital broker engagement includes a Consumer-Duty and vulnerable-customer layer designed in — not retrofitted. It is not marketing. It is the direct result of losing seven figures to firms that had those layers implemented as posture rather than architecture.
That is the founder story. The rest of the practice — dealing desk AI, retention brains, KYC orchestration, surveillance triage — sits on that foundation. If you want to understand what we build and why we build it that way, this note is the answer.
FAQ
Is the £1.2M figure verified?
Yes — trading statements from five brokers across three jurisdictions. Available to any prospective client on request under NDA.
Do you have a conflict of interest working with brokers?
The reverse. Every broker I work with wants a partner who understands both sides of the screen. The founder story is a qualification, not a disqualification.
Does this mean you dislike brokers?
No. The industry has firms of every quality. I work with the ones that want the compliance envelope designed first.
Why publish this at all?
Because prospective broker clients deserve full disclosure before signing anything, and because the machinery is worth naming honestly.
Where does the £15k Diagnostic fit?
It is the entry point. Two weeks, a costed plan, and — for broker engagements — an explicit Consumer-Duty overlay from day one.
Want this rigour applied inside your firm?
Start with the free 5-minute AI Readiness Score, or go straight to the £15k Financial AI Diagnostic — a two-week engagement that produces a costed build plan mapped to your regulator, your stack and your P&L.