Here’s a situation that’s becoming more common.
A founder we coach is an expert in a specialized service area. He built high-quality, AI-powered software that does the work faster and better, but it still needs a human expert to review results, correct mistakes, and exercise judgment.
He noticed something frustrating. His old service business, where clients hired him for his expertise, was much easier to sell than his new software product with a human on the back end. So we explored the inverse. What if he put the human expert front and center, and let the software power the delivery behind the scenes?
That raises some old fears for him. Wouldn’t investors hate that? For years, the standard advice has been that venture investors want software companies, not service companies. However, that preference is starting to change.
Why investors have been wary of services
When investors say services don’t scale, they’re mostly using shorthand. Deloitte and Accenture employ hundreds of thousands of people around the world. Services can clearly get big. With hindsight, we all wish we could have invested in Accenture when it had five people.
The more precise problem is that traditional services scale linearly. If you sell hours, the only way to grow is to hire more people to sell more hours. Revenue grows roughly in step with headcount, and margins stay below traditional software margins.
There’s a second concern too. Service revenue can be a sugar high. It comes in fast, it feels great, and it can distract founders from building the product. Service work and software development require different trade-offs, and investors worry founders will keep choosing the service work because it pays today. That’s the trap we described in Don’t Build Business A to Fund Business B.
Every B2B software company is already partly a service company
The dirty little secret of B2B software is that services have always been part of it. Implementation, onboarding, support, customization, and customer success are services. In many software companies, consulting services are hidden behind titles like “forward deployed engineers.”
Even “Software as a Service” is a bit of a misnomer. Most SaaS is really software by subscription. You’re renting software, which isn’t the same as receiving a service from an expert.
So the question was never really whether a software company would do service work. It was how much, and whether the service work pulled focus away from the product.
What AI changes
AI is collapsing much of the gap between services and software.
When software and AI do most of the delivery, the cost of providing a high-quality service drops dramatically. The expert spends their time on judgment and relationships, while the system handles the repetitive work. That’s one reason you’re seeing more models built around forward-deployed engineers and expert-led delivery supported by software.
The result is a business that can grow faster than traditional services without needing a proportional increase in people for every new customer.
The market-size math
There’s also a huge market-size argument.
Far more customers know how to buy a professional service than know how to buy, implement, and get value from software. That second group is smaller than most founders assume.
Here’s a simple example. Say a pure software business has 90% gross margins and reaches $1M in revenue. That’s $900K in gross profit.
Now say an AI-powered service business has 60% gross margins, but can reach a market three times larger because customers already understand how to buy it. At $3M in revenue, that’s $1.8M in gross profit or 2x the software play.
You still have to grow into that larger market, and AI can help with that too. But many founders and investors would happily accept lower margins for access to a much larger pool of buyers.
A different growth curve
Traditional services grow in a straight line. Software at its best grows like a sharp hockey stick.
AI-powered services may land somewhere in between: a wider, more gradual hockey stick that starts from a higher base. You can win customers earlier because the offering is familiar and easy to buy, and growth can accelerate as more of the delivery is handled by the system.
How this fits with “don’t build A to fund B”
This might seem to contradict what we’ve written before about avoiding the agency-to-software trap. The difference is whether you’re building one business or two.
The trap is running a service business to fund a separate software product you hope to build someday.
An AI-powered service company is one business. The service is how customers buy it. The software and AI are how you deliver it. Every engagement should make the system better, feeding the kind of improvement loop we described in Running an AI Company Is a Different Kind of Business.
A useful test: does each new customer make your delivery system smarter, faster, or cheaper? If yes, you’re building one compounding business. If the service work and the product are improving separately, you’re probably back in A-to-B territory.
What this means for founders
If you’re building an AI-powered service company:
Make it easy to buy. Package your offer the way customers already know how to purchase it.
Keep the expert at the front. Trust is often what closes the deal, especially in B2B.
Let AI carry more of the delivery over time. Track how much of the work the system handles and how that affects margins.
Tell investors the story clearly. Explain how the service wins customers, how AI improves delivery, and why the business can grow faster than a traditional services firm.
Service used to be something founders hid in the back of the pitch deck (or left out completely). With AI, it may become the most effective way to get software into the market.





