Freelancer, Agency, or AI: Who Should Actually Build Your Website?

A version of this question lands in our inbox almost every week.

A founder needs a website. Either there's nothing yet, or whatever they built isn't really doing the job it's supposed to do. And now they're staring at three options.

Hire a freelancer. Hire an agency. Or just throw it at AI and see what comes back.

Each one has people who swear by it. Each one has rebuilds in progress somewhere. And almost everything you'll read on this was written by either a freelancer or an agency. So weigh accordingly.

We're an agency. Same warning applies here.

But what I can tell you is what we've actually seen.

One client came to us after a freelancer built them something that "kinda sucked" (their words) and needed a full rebuild. AI-forward team, smart people. They didn't need more AI output. They needed direction.

Another founder tried to shortcut the whole thing with AI tools. Ended up with a site that was technically functional and strategically empty.

And a handful walked in with literally nothing. No brand, no positioning, no design. They wanted us to take it from idea to launch, because a Figma file and a goodbye wave was not what they were paying for.

Three different shapes of the same question.

Here's what we've learned.

The freelancer option

A good freelancer is excellent for the right job.

Specialists. Fast. Cheaper than an agency. If you've found one whose work you already trust, the experience is hard to beat. For a clearly scoped deliverable (a landing page, a design refresh, a single feature), a skilled freelancer is usually the smartest call.

The problems show up at the edges.

Freelancers are individuals. So when life happens (illness, a bigger client, a stretch where the inbox got out of hand), your project slows down.

The person designing your site is often not the person building it. And the handoff between design and dev is where a lot of stuff quietly falls through the cracks.

If your needs shift mid-project, you're renegotiating with one person who may or may not have room for what you're now asking.

None of this is a knock on freelancers. It's just what working with one person looks like. The question is whether your project fits that shape.

When a freelancer makes sense: scope is clear, you've worked with them before or have a strong referral, and you don't need ongoing support after launch.

The agency option

An agency brings a team, a process, and accountability.

When it's working, you get design and development thinking together from day one. Someone's always reachable. There's a track record you can check before you commit.

The risks are different from a freelancer's, but they're real.

Agencies vary wildly in quality. Some are sharp strategists and middling builders. Some make beautiful work that doesn't perform. Some put a senior team on the pitch call and a junior team on the actual project. (Pay attention to that swap. It happens more than people admit.)

The tell is usually in how they ask questions before proposing anything. An agency that quotes you without understanding your business, your users, or what success looks like is selling you design, not results.

Cost is the obvious factor. A good agency is a real investment.

It makes sense when the stakes justify it. When the site is core to how the business runs. When you're raising and need to look the part. When a previous build went sideways and you're fixing it properly this time.

When an agency makes sense: the site is genuinely business-critical, you need strategy and execution together, you've been burned and want accountability, and you're ready to invest properly.

The AI option

(Most agency content on this is either dismissive, protecting turf, or breathlessly enthusiastic, trying to look current. Neither is useful.)

We use AI tools constantly. Cursor is basically a team member at this point. We've used it heavily building Bento Japanese, our own product, and it's in the workflow on most client work. These tools are real. We're not going back.

Here's the part that usually doesn't make it into people's takes: AI makes experienced folks meaningfully faster. It's brutal to people who are still building judgment.

The mechanism is simple.

AI will confidently generate something that looks completely right and is subtly wrong. A nav structure that buries your value prop. A page hierarchy that confuses on mobile. Code that works until it doesn't.

If you've shipped enough sites to know when something's off, you catch it in seconds and move on.

If you haven't, you follow the AI down a path that costs you a week and ends with something that looks professional and doesn't actually work.

That's the gap AI doesn't close. It gives you the surface. It doesn't give you the thinking underneath.

(Yet, anyway. Ask me again next year.)

When AI tools work: you're early, you need something live to test, and you have someone in the loop with the experience to evaluate what gets generated.

When they aren't enough: the site is a real business asset and you need someone who knows what right looks like. And can tell when the AI doesn't.

The question nobody asks

Most founders default to two questions: what's cheapest, what's fastest.

Neither is the right question.

The actual one is: what does this site need to do, and what does it cost when it doesn't.

A bad website isn't free. You pay in slow loads, in visitors who bounce before they understand what you do, in leads that don't convert, in the rebuild you commission eighteen months later.

(How do I know? We've done the rebuilds.)

Figure out what the website is actually for. Then work backwards to who should build it.

A quick decision guide

Choose a freelancer if: scope is clear and contained, budget is tight, you have a strong referral, and you don't expect to need ongoing support.

Choose an agency if: the site is core to the business, you need strategy and execution together, you've been burned before, and you're ready to invest properly.

Use AI tools if: you're early stage, you need something live fast, you have someone internal who can fill the gaps, and you know you'll build it properly later.

The honest answer most of the time is a combination. AI to move fast and prototype. A freelancer or agency to execute properly when the stakes are clear.

The mistake people make is using early-stage tools on late-stage problems.

One more thing: do you even need a website?

This is the part agencies usually skip. So I'll say it.

Not every business needs a website. And of the businesses that do, not all of them need it built by a professional.

If you run a restaurant, a Toast page with your menu and hours probably does the actual job. If you're early stage, a strong Instagram and a clear DM funnel can carry you for longer than you think. If most of your customers aren't even online, does it really make sense?

A website is a business asset. It should not be treated as a default.

Build one when there's a good reason. Skip it when there isn't. And before you spend real money on it, make sure it's actually the thing the business needs.

(Yes, we're an agency telling you, maybe you don't need one. That's the point.)

Ready to figure out which one you need?

Book a free 30-minute consultation and we'll give you a straight answer. Including whether you need us, a freelancer, or honestly just a good AI prompt. Or, you know, no website at all.

Book a call →