AI Should Be Changing What Web Projects Cost. Here's What to Look For.

I've worked in web services for more than fifteen years. Over that time I've seen plenty of "this changes everything" moments. Most of them changed a few things, a little, and produced a lot of conference talks.

AI is different. It has changed how web and software projects get built, and who can build them. The way those projects get priced is still catching up.

How we got here

For a long time, a serious custom web project needed a full cast of specialists. A strategist to work out the requirements. A UX designer, a visual designer, front-end and back-end developers, someone for integrations, a copywriter, a tester, and a project manager to keep them all pointed the same way.

Each of those people brought real skill. The work was genuinely hard, and specialization was how the industry delivered it well.

That made custom work expensive for two reasons. The first was the expertise itself. Experienced full-stack developers are skilled professionals, and they're priced accordingly. The second was coordination. Handoffs. Status meetings. Meetings to prepare for status meetings. A requirements document that three people read three different ways.

Together, those costs put a lot of useful tools out of reach for small businesses and organizations. Their problems weren't too hard to solve. They were too small to justify the bill.

What changed: fewer handoffs

AI tools now handle a large share of the production work that used to need a dedicated specialist. Not perfectly, and not on their own, but well enough that one experienced person can move remarkably fast.

One or two capable generalists can now turn a conversation into a working prototype in days, produce strong first drafts of designs, code, and content, and test and document as they go. When fewer people need to hold the project in their heads, coordination shrinks. (The status meeting survives, but it's shorter. Progress.)

What changed: the team you don't need on call

The bigger shift, for me, is this: AI has made it possible to do meaningful work without a development team on hand at all.

I'm not a full-stack developer, and I'd be the first to tell you so. I wouldn't take on heavy, complex development by myself, and I'd be wary of anyone who claimed AI makes that easy. But AI now lets me make code changes I couldn't have made on my own, and build simple, practical applications that just weren't an option for me before.

That matters because a huge number of real business problems are exactly that size. They aren't moonshots. They're "we really need a better way to do this one thing," and for years the only options were a spreadsheet or a development budget nobody had.

A real example

The Arabian Working Western Alliance (AWWA) runs several programs that involve people signing up and paying: Incentive Fund nominations, stallion enrollments, and Star Program enrollments.

Until recently, all of it was managed with spreadsheets. Payments came in by Venmo or by check. That wasn't for lack of effort. There simply wasn't an affordable online way to manage it. Every nomination meant manual entry, manual payment tracking, and manual answers to "Is my horse nominated?" Spreadsheets are wonderful tools right up until they become the system of record for an entire organization.

With AI-assisted development, I was able to build:

A database holding AWWA's existing nominees and enrollments, migrated from those spreadsheets
An easy-to-use interface for managing records, without needing to be a database person
An online portal where people can nominate horses, enroll stallions, and pay directly
Online access to program information, so members and trading partners can find what they need without waiting on a phone call or email

The result is less administrative work, fewer chances for errors, and noticeably better service to the people AWWA works with.

Under a traditional development process, staffed with high-cost full-stack expertise, a project like this would never have been affordable for an organization of this size. It would have stayed on the "someday" list indefinitely, next to the spreadsheet.

What didn't change

Here's the important caveat. AI makes producing the work cheaper. It doesn't make judgment cheaper.

Someone still has to understand the actual business problem, decide what's worth building and what isn't, notice when AI output is confidently wrong (it happens, and with impressive conviction), design something that will hold up over time, and keep the scope honest.

It also matters to know where the line is. Simple, well-scoped applications are now within reach for an experienced generalist. Large, complex systems with heavy security, performance, or integration demands still deserve experienced developers. Knowing which kind of project you have is part of the job.

So the real comparison isn't "people versus AI." It's experienced people working the new way versus anyone, of any size, still working the old way.

What this should mean for clients

If the work takes fewer hours, fewer handoffs, and in many cases fewer specialists, it's reasonable to expect a few things over time:

Costs should trend down for comparable work. Not to zero, and not overnight, but in a noticeable direction.

Timelines should get shorter. Seeing a working draft early should be normal, not a pleasant surprise.

Quality should go up. Faster tools leave more room for thorough testing, better documentation, and trying a few options before choosing one. Cheaper and better should travel together.

Smaller problems are now worth solving. If your organization runs on a spreadsheet, a stack of checks, and someone's very good memory, there may now be an affordable alternative.

The industry is still adjusting

To be fair to everyone in this business: change like this takes time to work through. Firms of every size are figuring out how to adapt their processes, staffing, and pricing models, and many are doing it thoughtfully. Billing structures built over decades don't reinvent themselves in a quarter.

That said, as a client you're entitled to ask how the shift is affecting your project. Some reasonable, friendly questions:

How is AI part of your process today? Anyone working with these tools should be able to answer this clearly.
How does that affect my timeline and budget? A good partner will talk about it openly.
Who will actually be working on this, and what will each person do?
Is this priced by hours or by outcome? Hourly billing and much faster tools make for an interesting combination. Worth a conversation.
How will the work be documented and maintained? Fast should still mean durable.

If prices and timelines haven't moved at all, and quality isn't noticeably better, it's fair to ask why. There may be a good reason. Ask anyway.

The bottom line

AI has changed both how web projects get built and who can build them. Work that once needed a team of specialists can often be done by one or two experienced generalists. Simple applications that were never affordable before, like moving an organization from spreadsheets and checks to an online system, are now realistic for small businesses and nonprofits.

That's good news for organizations of every size. If you're planning a project, make sure you're seeing that progress, and not only in the slide deck.

Running something important on a spreadsheet? I build right-sized custom tools, databases, and online portals for businesses and organizations, combining modern AI tools with more than fifteen years of hands-on experience. Get in touch and let's see what it would take today.