Agency, Freelancer, or AI-Augmented Solo Dev?

The app development market changed more in three years than in the previous decade. An honest comparison of three ways to actually get your app built.

Three Models, No Obvious Answers

A CEO or founder looking for someone to build a mobile app has three options today: a software agency, a classic freelancer, and an independent engineer working with AI agents (the newcomer of recent years). I'm the third one myself, so I'm declaring my bias upfront. But I'll make the comparison honestly, including the downsides of my own model.

The Software Agency

When it makes sense: large scale, compliance requirements, a project measured in person-years, guaranteed continuity independent of any single person.

Pros: processes, replaceable people, SLAs, experience with large organizations.

Cons: you pay for the whole apparatus: project manager, sales, office, and margin. Your project is built by people you never met during the sales process. Communication goes through layers, and every scope change is a formal procedure. Typical entry budget: high five to six figures.

The Classic Freelancer

When it makes sense: a small, well-defined scope; fixes and iteration on an existing app; a tight budget.

Pros: you talk directly to the person writing the code. Flexibility and price.

Cons: one person working traditionally has a hard throughput ceiling. Backend, design, store publishing, testing: something usually falls outside their specialty. And the risk: they vanish, get sick, or take a better contract.

The Solo Dev With AI Agents

This model didn't exist three years ago; in 2026 it genuinely competes with small teams. An independent engineer delegates to AI agents what used to take several people: code search and refactoring, tests, audits, release preparation. I wrote about that workshop separately in AI-First Development in Practice.

Pros: the scope of a small team without agency overhead. One person owns the whole thing, from architecture through release, so nothing falls between "departments". Decisions take hours, not sprints.

Cons (because there are some): continuity still depends on one person (ask about documentation and code handover). For projects needing many specialists working in parallel, say a dedicated ML team plus a mobile team, the solo model runs out. And finally: "dev with AI" is sometimes a cover for people who couldn't build an app without AI at all. How to tell the difference? Below.

Red Flags, Whatever the Model

• A quote with no questions about your business. If someone prices your project off a feature list without asking who will use it and why, they're pricing screens, not a product.
• No apps in the stores. A portfolio of mockups is not the same as products that passed App Review and have real users. Ask for App Store and Google Play links.
• "Everything is possible." An experienced developer says "no" and proposes a cheaper alternative. A yes-man says yes to everything and bills you for every yes.
• Silence about maintenance. An app lives on after launch: OS updates, backend costs, error monitoring. Whoever doesn't bring this up before the contract will leave you alone with it after launch.

Questions Worth Asking Before You Sign

1. Which apps have you shipped to the stores, and what's happening with them today?
2. What happens to the project if we part ways halfway through? (The answer should include: your repository, your accounts, documentation.)
3. What will this app cost per month to run after launch?
4. Which of my ideas would you cut from the first version, and why?

That last question is the best filter I know. A good developer always has a list of things to cut.

Looking for someone to build your app and want an honest conversation instead of a sales pitch? Get in touch. Even if it ends with me recommending a different model than mine.