The End of Build Once, Sell Forever
Distinct, or becoming extinct?
For decades, the path to impact was clear.
If you wanted to change the world at scale, you built a product. Services were boutique, constrained by time, ultimately unscalable. SaaS was the dream because software scales geometrically - write once, sell infinite times.
The trade-off was obvious:
Products: Scale infinitely, but serve the average use case
Services: Perfectly customised, but constrained by human hours
Ambitious founders chose products. Consultancy was the ‘safe bet’.
The speedboat problem
A speedboat going in the wrong direction doesn’t get you there any faster. And suddenly, everyone has a speedboat.
When anyone can build software, the question shifts from “can we build this?” to “should we build this?”
The constraint isn’t execution speed anymore. It’s judgment about what to execute.
This affects products in three ways:
New customer acquisition dries up
Why would a startup in 2025 pay for Salesforce when they can build a custom CRM tailored exactly to their workflow?Existing customers defect
Klarna didn’t just reduce their Salesforce spend - they invested in internal tooling that compounds in value. The economic calculation flipped.The standardisation tax becomes visible
Products serve the average use case. When customisation becomes cheap, “good enough for most people” is suddenly a massive liability. You’re paying for features you don’t need and missing features you do.
The paradox: Products were supposed to benefit from AI (faster development, better features). Instead, AI made the alternative to products - custom solutions - viable.
Decoupling value from hours
But services have their own transformation happening.
The old constraint on services was simple: you sold time. A consultant could only serve N clients. A fractional CTO could only advise so many companies. Scale meant hiring more people (and dealing with quality control, management overhead, margin compression).
AI changes this by creating leverage without standardisation:
Audit 10 codebases instead of 2 (AI-assisted analysis)
Deliver a 50-page strategic report in days instead of weeks
Explore 100 design variations instead of 5
Maintain relationships with 20 clients instead of 5
But here’s the crucial difference from products: the customisation is the point. You’re not building one thing for everyone; you’re using AI to deliver bespoke judgment to each client faster.
The value isn’t “we built software you can use.” It’s “we applied expertise to your specific context.” AI makes that expertise more scalable without making it generic.
So what's left to defend when anyone can build anything?
More here: