AI Insights

For decades, the advice to CTOs to ‘buy, don’t build’ has been sound advice. Take something off the shelf, install it and move on, because speed has mattered more than almost anything else, and SaaS was the answer. There was no need to worry about building capability yourself, shifting culture inside your organisation or running long transformation programmes just to get value. But AI has shifted that argument, and things are changing quickly, driven by two forces at the same time – one around demand, and the other around delivery.

On the demand side, everything with software in it now needs an AI story, from products and services through to board-level conversations. Stakeholders and customers expect it, and if AI isn’t clearly on your roadmap, you’ll be answering questions about that very soon.

On the delivery side, there is truth in the idea that AI can now generate software at a pace we haven’t seen before. Marketing hype would have us believe it does everything itself and makes humans redundant, which is dramatic rather than accurate, but it does fundamentally change the economics of delivery. If you’re a software business and you don’t pivot how you deliver, you won’t have a business in five years. AI isn’t replacing engineers, but it is replacing the inefficient delivery systems around them. The old model of selling seventy developers into a customer environment is diminishing, and where it still exists those seventy people increasingly operate in small teams delivering far more innovation, at a speed that simply wasn’t possible before.

The new model

The biggest advantage SaaS ever had was speed, because off-the-shelf was always faster than bespoke. But when AI-assisted engineering removes that gap, the trade-offs start to look very different. If you can build quickly, it becomes harder to justify accepting something that only partially fits your business.

This becomes even more obvious when you look at AI itself. By design, it’s meant to be unique, shaped by your data, workflows, risks and your operating model. Trying to create repeatable AI products that work the same way everywhere is incredibly hard, which is why we’re seeing such a sharp move towards bespoke AI in more mature industries like financial services, healthcare, retail and manufacturing.

The numbers reflect that shift. Bespoke AI product development is growing at over 40% a year and is forecast to become a bigger market than traditional application development in less than 18 months. Buyer behaviour mirrors this, with 67% of existing enterprise customers planning to build AI into their products this year, and 65% of new business pipelines already including data and AI work, often before the go-to-market story has fully caught up. Customers are moving whether organisations feel ready or not.

How buying decisions get made

In established markets, buying is about efficiency. Budgets are planned, competition is crowded and organisations optimise for cost control. In emerging markets, buying is driven by competitive pressure and fear of falling behind, where speed and outcomes beat maturity every time. People don’t buy because something feels safe, they buy because not moving feels riskier.

For CTOs, the trade-off increasingly comes down to whether to buy capability faster, or risk dependence on someone else’s roadmap for something that is critical to survival. With the ability to build faster, learn quicker and adapt continuously, while keeping control of what genuinely differentiates the business, ‘building’ is firmly back in the frame.

The chaos AI has introduced into the enterprise world feels risky, but as the old saying goes, in the midst of chaos, there is opportunity. New markets don’t have established leaders yet, and there is a short window where speed matters more than scale, and where becoming the obvious choice is still possible. “Buy, don’t build” worked when software engineering itself was the hard part. Now, the real value lies in understanding what to create, how fast it needs to evolve, and how closely it needs to fit your business to matter.

Joe Wolski, CTO, Godel

Posted 25 Mar 2026

share