I have intentionally tried to steer clear of AI related topics since it is not the lane I want to concentrate on in these articles. However, I have to acknowledge that there is a huge intersection between the use of AI tools to create a meaningful business and the ability to start and manage small teams. So, I'm biting the bullet as they say and giving my opinion on what makes a good AI centric team and specifically a company composed of just that small team.
I want to get very granular in this article. I want to concentrate on what I consider the best composition for an AI centric team that intends to create primarily cloud-based software. There is so much potential at the moment for small teams to quickly deliver enterprise and commercial software at a fraction of the time and cost. However, this window won't last forever. Large corporations may be slow to act, but with their vast resources and nearly bottomless funds, they will eventually out compete these more agile teams. Even AI could easily reach a point where software is so easy to create it becomes an individual, completely customizable endeavor where there is little money to be made. Thankfully, we're not there yet. This is the time for small teams to act.
The beauty of AI tools is that you can get a lot more done with a lot fewer people, but it's critical that they are the right people, filling the right roles. We are inundated with examples of individuals creating these impressive applications by using a single, simple prompt. There's a whole lot of misdirection going on with these vibe coding experiments. These "tech demos" are not releasable pieces of software, and they are most certainly not monetizable. What's needed is a distinctive workflow that allows for repeatable, high-quality results. While I have a very specific workflow that I've developed over the last three years, I'm going to concentrate on the roles you'll need to succeed rather than the process itself. This typically means a team of between five and seven people. That's it. That's all you need. With this number you'll be able to create world-class software in a plethora of niche areas.
Before going into the specific roles, let's start with one universal qualification. Without exception, every single person on your team must already be proficient using AI tools. They should already be using them in their day-to-day life. You don't have time to teach them these basic skills, and if they haven't already embraced AI on a personal level, they're not likely to have the passion for it that's necessary to help your team succeed. With that out of the way, let's start talking roles.
You will need one or more domain experts. These are people who know the industry/niche you want to operate in. If you are one of these "experts", definitely hire a second person. You're going to be busy enough leading the team and won't have time to always serve as the subject matter expert. These people are absolutely critical in determining the requirements that will satisfy your customers' needs. Requirement documents are a must. Without solid requirements documents, AI coding agents will go off the rails much quicker than you expect. This step cannot be skipped. Luckily, your subject matter experts are already using AI tools (see above), so they will be able to use these tools to efficiently and thoroughly create the documentation you'll need.
Without question, you will need a senior software developer. AI can do the work of dozens of people, but it needs the guidance of someone with deep technical knowledge. This person will define the technology stack your application will use rather than letting AI decide on some random approach that won't allow for your long-term needs. You'll need someone who understands the code being written, even if they don't have to write most of it themselves. They'll want to do code reviews for every bug fix and feature that goes in.
Next, you'll need a DevOps person, one who can build and maintain the environment that will host your application. Things go wrong in the environment, and you'll need someone who fixes problems quickly and ensures you have the right level of resiliency and redundancy required for your product. Ideally, this person should also have some good security chops as well. Some operational activity can be outsourced, especially if your application is so critical that it requires nearly 24/7 uptime, but this will depend on your specific situation.
You're also going to need someone to deal with the finances. This is a full-time position that you can't just pawn off on one of your other roles or, heaven forbid, yourself. People who like to work with numbers, really like to work with numbers. For the rest of us, it's a special kind of torture.
I would also recommend a dedicated support/sales person, someone who will handle the influx of questions and concerns about your product as well as be able to assist with minor customer issues. This role is a bit broad, I know, and could be a bit of a catch-all for a number of customer-facing activities. It will depend heavily on the nature of the software you're creating and your intended audience.
Finally, you'll need someone to marshal the troops and tie it all together. That's you. You are the leader of this team after all.
You'll notice I haven't mentioned quality assurance, and that's no accident. On a team this small, you probably can't justify a dedicated QA hire, so testing becomes everyone's job. But be careful. When something is everyone's responsibility, it has a nasty habit of becoming no one's. Just keep this in mind and adjust as needed. On the other hand, if you can afford a dedicated QA person and the software warrants it, by all means, hire that person.
The roles are one thing, but of course you have to hire the right people. For that, I'll direct you to my many other articles on good hiring practices.
So, what are you waiting for? Find your brilliant idea and start building your team.