There’s a revolution happening in how software gets built, and most of the market hasn’t priced it in yet. For years the math was fixed. A modest web application with auth, a database and a few dozen screens meant a sizable team, Gantt charts and deadlines set in quarters. Everyone accepted that because it was true and building software took time.
This is no longer the reality. What used to take six months a skilled team ships in weeks. The quality didn’t get worse, the work got lighter. Teams that have made the transition to this new reality can do dramatically more to further your IT goals in significantly less time, and therefore much less cost.
The opportunity is being first to notice
When the cost of producing something drops by an order of magnitude, that value has to land somewhere. The smartest buyers are the ones capturing it early, before the rest of the market catches up.
The right partner passes that gain to you: faster timelines, leaner budgets, more shipped per dollar. You’re not getting a discount. You’re getting the actual new cost of the work. You should demand transparency about AI use from your vendors now. Not next quarter, next milestone, or when it’s time to renew the contracts, but right now.
The uncomfortable part of the conversation is that vendors are heavily incentivized by outsized margin gains to continue billing under a 2020 regime while adopting AI opaquely.
The exciting part of the conversation is that arriving in the new world means ambitious things that were “next year, if the budget holds” are now shipping this quarter. Smaller teams with skilled engineers ship just like big ones used to.
The distance between an idea and a working product is shorter than it has ever been.
“But isn’t AI-built software unreliable?”
In the wrong hands it can be, and that’s exactly why choosing the right partner matters more now than ever. Almost everyone using ChatGPT and other AI chat services has experienced AI shortcomings first hand, in the form of confident-but-wrong answers. Those same shortcomings are actually less prevalent writing code than prose. However, architectural considerations in software creation are seldom prompted correctly from the outset by non-professionals, which does lead to the examples skeptics point at: systems that buckle under load and real security gaps in hastily shipped software.
“AI can’t build production software” misses the point. AI doesn’t build anything — a person does. The gap between a skilled holder of the tool and a careless one didn’t shrink, it grew. In one sense, this means you’re no longer buying software, you’re hiring judgment you can trust to drive powerful tools to good results.
The bar actually went up
The doom perspective gets this wrong. The new tooling didn’t make engineering trivial — it moved the value to where it always belonged. Not in producing lines of code, but in knowing what components make up the system. Understanding the use cases and the work is more important than ever, and now the focus can shift squarely there. That work got harder to fake. True discipline in the craft of IT is now the whole game.
A skilled partner with these tools isn’t a cheaper version of the old team — they’re a better one.
Have the conversation
Talk to your vendors. The ones worth keeping will welcome the questions. Ask how their approach has changed in the last year. Ask what a leaner, faster version of the plan would look like, and watch whether they light up or get defensive.
The answers will tell you who has actually evolved and who is still working — and pricing — the old way out of habit. The new math is here. Find a partner who’s already operating on the level.