Service as a software
For a long time, the advice to founders was pretty simple. Build software, not services.
Software had high margins and scaled beautifully. Services had low margins, lumpy revenue, and every new client meant hiring more people.
I've been thinking a lot about whether that's flipping, partly because I'm living it with Equilo, where we're using AI and automation to rethink how bookkeeping gets done.
Everyone is mining for gold
Right now, almost every founder and investor is chasing AI software.
It makes sense. Building software has never been cheaper or faster, and there's a lot of money flowing into it.
But that's exactly the problem. When it's that easy to build and everyone is building the same thing, the space gets crowded very quickly.
It reminds me of the gold rush. Everyone rushed out to mine for gold, but you could probably have done very well by mining for silver or bronze instead, just because far fewer people were competing for it.
I think services could be that silver right now.
Why services are different now
Services businesses were never really seen as exciting. They were hard to scale because the work was done by people, and you could only grow as fast as you could hire.
AI changes that maths. A lot of service work is structured and repeatable, like bookkeeping, audits, claims processing or transactional legal work, and that's exactly the kind of work AI is getting very good at, very fast.
So a services business that's built with AI from day one can deliver the same outcome with a much smaller team, at a lower price, and with better margins than the old model ever allowed.
People want the job done, not the tool
The other thing I keep coming back to is what customers actually want.
Most business owners don't want another piece of software to learn. They want their books done, their claims processed or their contracts reviewed, and they want someone accountable for getting it right.
Selling the finished work rather than a tool also means you get better as the models improve. Every upgrade makes the work faster and cheaper to deliver.
Start with what's already outsourced
If I were looking for the next services opportunity, I'd start with work businesses already pay someone else to do.
There's already a budget for it, the customer is used to someone external doing it, and they're already paying for an outcome. You're not asking anyone to change their behaviour, just offering a better version of something they already buy.
Bookkeeping is a good example. Most small businesses already outsource it, and most of those providers haven't changed how they work in a very long time.
Pick the less crowded space
What I like most about this space is that it's not where the crowd is.
A lot of these industries are seen as boring, old-fashioned or too operationally messy for most tech founders to bother with. They'd rather build the next AI app than learn how a BAS gets lodged.
That's exactly why it's attractive. It's the same thing I've written about with careers, where you're better off finding a race with fewer runners than trying to be 0.1% better than everyone in the most crowded one.
What it means for the people doing the work
The part I'm still unsure about is what this does to jobs.
If services firms can do the same work with smaller teams, the traditional pyramid shrinks, especially at the bottom. That's the same bottom rung where juniors have always learned the craft.
I don't think these industries disappear, and I think a lot of these firms will thrive. But the people in them will spend a lot more time on judgement, relationships and advice, and a lot less on the repeatable tasks that used to fill a graduate's first few years.