For most of my career, the build vs. buy decision in software has leaned heavily toward buy, especially in the enterprise. It is understandable. Buy is faster. Buy is safer. Buy comes with support, training, and a roadmap.
But lately, I have been rethinking that bias. More accurately, AI coding assistants are forcing me to. Claude, ChatGPT, Gemini, and others have fundamentally shifted the landscape. We are entering a world where build is no longer a blocker: prohibitively slow, expensive, or complex.
In fact, in many cases, build might now be the smarter, leaner, and more adaptable option. Not just for internal tools, but for core systems that have historically been considered off-limits for in-house development. Think master data management, warehouse management, order management, or even parts of your ERP. Many of the so-called SaaS platforms and enterprise systems in these categories are clunky, outdated, and overloaded with features most teams do not need. Sometimes you do not need a full ERP — you just need a lightweight solution that works well for your business. The tools and talent to build these are more accessible than ever, and AI coding assistants are lowering the barrier even further.
Here is my take on how AI coding assistants will be reshaping the build vs. buy decision — and why I believe more enterprises should shift toward what I call Build-to-Flow.
1. Enterprises have over-bought and under-built
Most organizations are buried in a sea of platforms and SaaS tools. And yet, every new initiative still seems to require consultants, internal support teams, and complex integrations.
This is the dark side of buying for completeness. We buy big, configurable platforms hoping they will fit every edge case. But in practice, they end up overbuilt, underused, and expensive to operate.
I have seen quite a few projects fail not because building was too hard, but because what we bought never really fit, and drained the budget.
2. AI makes Leanform Builds possible
I have been personally using AI coding tools a lot lately, and I have seen firsthand how much they accelerate prototyping, frontend, backend, and even integrations.
You do not need a full team to build an MVP anymore. You need clarity on the problem, some domain knowledge, and a feedback loop. AI takes care of a lot of the heavy lifting.
This enables what I think of as Leanform Builds — fast, lightweight tools that wrap around real workflows and deliver value quickly. They are not trying to be platforms. They are purpose-built extensions of how your team already operates.
- Prototyping internal tools is now a one-person, one-week job
- Building a data workflow becomes more approachable, with AI stubbing out connectors and pipeline logic
- Wrapping a third-party API or enriching a SaaS tool is often easier to build around than force into an existing product
3. Build-to-Flow vs. Build-to-Spec
Traditionally, building meant collecting requirements and trying to recreate every feature in a vendor solution. That is build-to-spec — and it is usually a trap.
What I advocate is Build-to-Flow. Start with how the work actually happens. Build tools that match your team’s natural rhythms. Fit software to the business — not the other way around.
AI assistants and modern stacks make that possible in a way that simply was not realistic five years ago.
4. What about “build is not our core competency”?
This is one of the most common reasons companies default to buying: building software is not our core competency.
But the better question is: is owning and evolving our operating model a core competency?
In many areas like manufacturing, lead management, inventory planning, or internal data platforms, your processes are often unique to domain. They are often legacy with deep roots, and in other cases evolving quickly. Often they do not map cleanly to vendor software roadmaps.
Delegating those workflows to off the shelf tools often slows you down. You lose clarity and adaptability.
You do not need to be a software company to build what matters. With Leanform Builds, you can move surgically — focused on the workflows that differentiate your business.
5. Buying takes longer than you think
A typical sourcing or RFP process at an enterprise can take six months to a year. Requirements gathering, vendor evaluations, legal reviews, procurement — all before anyone touches the actual product.
Then comes implementation. More time. More consultants. More change management. And still no guarantee the final system will actually meet your needs.
Now imagine what you could achieve in that same time frame by building and iterating. You could have a lightweight, working version in production. You could be collecting feedback. You could be adjusting it as the business evolves.
This is what I think of as Iteration-Led Architecture — build just enough to drive value now, then evolve. In many cases, the “fast path” of buying is only fast on paper. Building might actually get you to value faster.
6. Why not build around what is already working?
Many large implementations fail not because the technology is flawed, but because of internal friction — organizational inertia, politics, and change management hurdles.
One of the advantages of building is that you do not need to rip and replace everything from day one. You can start small. You can build around what people are already using and extend it gradually.
Instead of forcing the business to change to fit a tool, you can Build-to-Flow — build tools that align with how the business already works, and improve from there.
That is how change actually sticks.
7. What I recommend
If you are a technology leader, here is what I would start doing:
- Re-audit past buy decisions. Would you still choose to buy that product if you had AI-assisted build velocity at your disposal?
- Pilot Leanform Builds more aggressively.
- Think modular. Build focused, single-purpose tools that can be replaced, iterated, or layered over time, not big monoliths.
- Do not wait for perfect. Use Iteration-Led Architecture to build something usable, get feedback, then keep shaping it.
- Balance build talent with build leverage. You do not need ten times more engineers. You need engineers who know how to use AI tools well.
8. The default should no longer be buy first
AI coding assistants have not eliminated the buy option, but they have ended its monopoly on pragmatism.
Building is back on the table.
It is cheaper. It is faster. And more often than not, it is better aligned to your business.
The future of enterprise technology will not be built entirely in-house, but it also will not be bought off-the-shelf.
It will be composed through Leanform Builds, shaped by Build-to-Flow thinking, and delivered through Iteration-Led Architecture — all powered by the leverage AI now provides.
We should start leaning into that now.