So here it is; August already.
About six or seven months ago, just back from the Christmas break, I told the team I didn’t think it was proving very fruitful to be putting all our energy into potential investors and partner banks at that stage as it was going to be time-consuming, and our priorities were unlikely to be theirs. So I said: “let’s discuss another approach, where we go and do this ourselves”. Everybody talked it over and broadly agreed it would be hard but viable.
Six months on, most of the things I thought would be hard, we are starting to solve. But other things that I hadn’t considered have proven to be harder and taken longer than I ever expected.
Two big things that I worried about
Going into this, there were two matters that I knew were foundational and which I worried could have the potential to derail us.
The first was licensing. The worry was that everything Lucie did would be captured as personal financial advice or credit assistance, and would therefore have to operate under an AFSL or an ACL. Some of what Lucie will eventually do will need those licences in time, and that has always been planned for. My worry was the bigger one: that every single interaction would be caught, with all of the compliance machinery that follows. It was daunting, and it worried me.
But between our own research and some external legal support, we now have a pathway through it, framed by a boundary between ‘unregulated’ and ‘regulated’ activity; which is a weight off my shoulders.
Of course what it leaves is the question of where regulated activity starts and unregulated activity stops. Except that isn’t a line; it’s a boundary, and it’s flexible in size, opacity and permeability, and a good deal of our work from here is going to be about navigating that. But that’s a defined problem. We can work on a defined problem. Before, it looked like an overwhelming one.
The second was Lucie getting the numbers wrong. A wrong balance, or a spending figure that isn’t right. The risk of that has the potential to destroy everything we do, and it’s the other thing that kept me awake.
But I think we now have a pathway through that; a verified-data pipeline architecture in which the figures on your screen are meant to come through structured, deterministic paths, with Lucie’s AI agents focussed on framing the explanatory commentary. We’re not there yet. But if it works properly, it’s fundamental.
Those were the two big imponderables at the start. Having pathways through those forests of danger and ambiguity is frankly pretty huge.
Where did six months go?
Our initial build plan had a first phase, which was done in a few weeks. Then a second one: get accounts and data ingestion working, and get to the point where you could look at money accurately in the app in a straightforward way. That was meant to be another short(ish) phase; the one before the interesting work of building money management tools.
Well that hasn’t worked out, has it? Here we are. We’re very close now, but we’re still going. A one-month thing became a six-month thing, and written down like that it looks horrible.
James, my CTO, says this is the hard part; that the core platform is the difficult thing to get right, and that once it’s in, the rest gets more straightforward. I hope he’s right. I look at what’s still to do and I can see difficulties there too. Though I suspect some of it is simply that one full-time person and a part-time team is a slower machine than I’d allowed for.
Our open banking contract has gone the same way. We started discussions ages ago and we’re at the final contract stage, working through the last of the legal detail. But it’s taken a long time. It isn’t for want of motivation on anybody’s part; everything just takes longer than you expect. A few days waiting on a reply. A few more getting round to the next thing. Then stakeholder consultation, then legal advice, then round again.
Six months, and it sometimes feels like a year.
One bit that isn’t working
Nobody really knows we exist yet.
I’ve kept up what I said I’d do: an article and a TikTok and Instagram video every week since March, and a monthly stakeholder email. Some of that has been genuinely enjoyable, and there’s been a bit of interest off the back of it. But five months of it hasn’t produced anything like the followers or the waitlist signups we will eventually need. Beyond that weekly output I’ve done very little on marketing and brand building, because we’re still so immersed in the build. That has to get some proper focus, and it will.
What I have got out of it is a much better sense of how little I know about social media marketing. What’s the saying: the more you know, the more you realise you don’t know? On that basis I must be very wise.
But there are plenty of wins
We’ve just rebuilt the Lucie Admin Portal: the internal tool we use to see what the system is doing and to configure how it behaves. The old one had been ‘cobbled’ together a while back to get us going but was past its use-by date and was becoming increasingly hard to use. The new one is the opposite of that, and the reaction from the team has been better than anything I expected (which is gratifying since I designed all the new UX).
Two things have come out of it, I think. It demonstrates that this is a serious venture building things properly, which is a message worth sending internally. And it forced me to interrogate every detail of how our system fits together: the workflow intricacies, the minutiae of the rules that govern them, the internals of the ingestion pipeline, the specifics of how tools get called to fetch information or structure how it is presented. I showed it to someone for the first time yesterday, and I could see immediately that they understood how the thing works in a way they hadn’t before.
We’re nowhere near polished on the actual workflows. But I can now put a screen in front of somebody and say this is how it works, and here is where you change it.
Everybody has a life
This is the part I genuinely didn’t anticipate.
At the moment I’m the only person on the Lucie team full-time. Everyone else has other work, or family, or their own circumstances, and over these months every single one of them has had something come up. Life gets in the way, priorities change, crises crop up, and what people are able to contribute waxes and wanes. And so does the pace of our work.
There was a month, maybe more, when, to me, it felt a bit like nothing moved much at all. Some people had things going on. Other people were away. To a degree it felt a bit like it was just me, keeping the thing going on my own at that stage.
It’s different now; we’re making great strides and there’s a real sense of progress. But pace comes and goes doesn’t it? I guess one advantage of being as old as I am is that I’ve been here before. I know these stretches happen and I know the answer is to keep going. That doesn’t stop me lying awake at night working through every indicator of failure I can conjure up, but at least I don’t dwell on it. And tomorrow is another day, circumstances change, and you never know. Understanding that helps.
Would I do it again?
Yeah - I wish I’d done it ten years ago.
It’s unquestionably the hardest thing I’ve taken on. When I started my own consulting business twenty-five years ago, that was riskier: I knew nothing about running a business, I had no idea how to win consulting work, and I had serious financial commitments. In hindsight I can’t believe I threw myself at it with such elan.
This is a different sort of hard. It is infinitely more complex, and half the things we are grappling with don’t have answers yet. The joys of riding the AI wave. But it’s infinitely more enjoyable for exactly that reason. Working out how to solve complicated things, and learning enough about them to know what I’m doing, has been both extraordinarily valuable, and humbling.
And here is the part I can’t account for. A year in, not much sleep and working silly hours six or seven days a week, and I’m healthier (touch wood) than I’ve been at any point in the ten years before this. Steadier. More positive about the world. Fitter. Hopefully nicer to be around.
I have no explanation for it. But hopefully I won’t write this and then the wheels will come off the cart tomorrow :).