For years, the startup playbook was fairly predictable. Raise some money, build a product, hire a few engineers, then hire some more. As sales grew, so did the team. Before long, the company had people for marketing, customer support, finance, recruitment and everything else that founders had once been doing themselves.
AI is beginning to change that formula.
Some startups are discovering they can grow significantly without growing their headcount at anything like the same rate. Rather than hiring another engineer, marketer or customer support person, they are increasingly turning to AI tools and agents to take on work that would previously have required another human.
Butternut AI, a no-code website builder, is one example. The company had grown to nine employees before cutting the team to four, with AI agents taking on some of the engineering work. Despite the smaller workforce, the company has around 10,000 paying customers.
Lindy, which makes AI “teammates”, has taken the idea further. The company has reduced its marketing staff and replaced some work previously carried out by creative agencies with generative AI tools. It now serves more than 500,000 users with just two people in marketing.
The numbers are striking because this isn’t simply about making employees a little more productive. If an engineer can use AI to do the work that once required several engineers, the company doesn’t necessarily need to hire those additional people as it grows, which changes the traditional startup equation.
A company generating £10m in revenue might once have needed dozens, perhaps hundreds, of employees to support that level of activity. An AI-native company can potentially reach the same revenue with a fraction of the workforce. The costs haven’t disappeared, of course. AI models, computing power and human oversight all have their own bills. But the economics can look very different when your biggest expense is no longer another salary.
There are signs that this is already happening. Startups founded in 2022 had an average of around 20.7 employees by their third year, compared with 32.5 employees for startups founded in 2018.
The temptation is to see this simply as another story about AI replacing jobs, but the more interesting shift may be happening earlier than that. AI means some companies may never create the jobs they would previously have created.
That doesn’t mean humans have become redundant. Someone still needs to decide what to build, understand customers, manage relationships and take responsibility when something goes wrong. As companies get bigger, those requirements don’t necessarily disappear either.
What is changing is the assumption that growth automatically means hiring.
For a generation of founders, the ambition was to build a big company. The new ambition might be to build a very big business with a very small company. If that becomes normal, the question won’t just be what jobs AI replaces. It will be how many jobs companies decide they need to create in the first place.
