The Autonomous Marketer: The Shift from Operator to Orchestrator & How to Build Your AI Team

If you’re having trouble getting the results you want out of AI tools, the problem might be that you’re approaching it too much like traditional software.

That’s what Matt Hastings found when he started tinkering with LLMs a year ago. He started creating websites and apps using AI without an agentic approach, which he found limiting. He had to rely on his own review of a singular coding agent’s output—a challenge, since he doesn’t have any formal technical training.

 It also led to slower progress and more project restarts. With agents, have them check their own work and create more complex workflows. For example, a planning agent might also interact with a skeptical agent who pushes back on the plan, leading to iteration between the agents and a stronger outcome.

“Too often, folks will try an AI tool, not have it achieve the goal they wanted in one or two tries, and give up,” Hastings shares. “We're used to digital experiences being slick and working right the first time: I hit a button, I get an output. LLMs never work right the first time. We're not fully in control and the outcomes are probable, so we have to develop a new relationship with the tools.”

So much of AI advice focuses on the right outputs, but just as important to autonomous marketing workflows is honing in on the right inputs. That relationship is one of a manager working with a team instead of a user working with a tool.

“Once you start thinking that way, you realize your job is to create a situation where the AI is more likely than not to give you a non-average outcome—because out of the box, you’ll get average outcomes,” Hastings says. “How do I give it enough information, enough context, enough direction, and frequent enough feedback that it continues to track towards the outcome I want?”

In other words, you have to set it up for success just like you would a new employee on your team. This mindset has helped Hastings unlock new autonomous ways of working, and it’s now an approach he teaches other white-collar professionals as part of his business MVP Club. When done correctly, you’ll reap the rewards of any good manager: allowing you to step back from the day-to-day, spend more time on strategy, and seamlessly work on multiple projects at once. This will likely become especially true as AI becomes more introspective.

Here are some of the best practices Matt has found for training a team of AI agents and being a better AI manager.

You wouldn’t expect a new employee to deliver perfect results on their first day with no onboarding first—and you shouldn’t expect that of AI either.

That’s why Hastings says the first step in building a good AI agent is “documentation all the way down.” You can think of this as the training step, where you’re creating a text document to explain to an AI agent exactly what its job is and giving it all the resources and context it needs to give you the outcome you’re hoping for.

“One way to think about creating an agent is that you're writing almost a persona or a character,” Hastings explains. “You might give it description and backstory, procedures that it always should execute or things it should never do, how it should pick up a piece of the work and interact with other agents, etc.”

Hastings relies heavily on LLMs to help him write the documentation for his AI agent teammates.

“One activity I really like to do at the beginning of a project—before I set up the agentic structure or the documentation—is just have the LLM ask me questions about my goal,” he explains, sharing that this helps him unearth assumptions he might have forgotten to tell the AI.

Once he feels like he and the LLM are clear on his goal, he’ll ask it to do some research online and see how other people have structured their agents to achieve that goal. Then, he’ll take it further and ask the LLM what documents he should make, and even have the LLM structure and write the document so that when an AI agent reads it, they know what to do. “It’s a very two-directional working relationship, which is so different than software,” Hastings adds.

Then, with clear documentation in hand, he can start to upload all this context to build out your new AI teammate. He tends to do this with files and folders in VS code, but says it’s similar to information you’d upload in the project function of ChatGPT or Claude.

Even with great training, you can’t expect AI to work perfectly the first time—just like you wouldn't expect a new employee to never make a mistake. That’s why the next step here is letting your AI agent do the work, then refining with feedback.

“You have to delegate and you have to trust and you have to rely on feedback as your main tool to course correct,” says Hastings. “Your control is at the beginning and at the end, and in setting the pace of iteration.”

That’s why he says it’s important to move fast and provide lots of feedback when it comes to building AI agents. Get a version of your documentation in place, let the agent do its work, and then use the output to further refine the documentation.

You can almost think of these feedback sessions like performance reviews. Hastings will go back to the original LLM that helped him write the documentation and share the outputs he got and what was wrong with them. Sometimes, he’ll even have the LLM evaluate itself.

“I’ll confirm that it remembers my goals, and then I’ll say, ‘Does this meet the goal?’” Sometimes it’s right—sometimes it’s wrong. But either way, that becomes a conversation about what needs to change.

From there, he’ll ask the LLM how to update the documentation to fix those problems.

“And then we’ll run it again, and repeat and repeat until we’ve polished the stone and we’re getting consistent, good outputs,” Hasting explains. “Again, I think the best way to use these tools is to iterate as fast as possible because you're going to have to get through maybe five or six repetitions before you get to a really good system.”

Hastings also emphasises the importance of creating an orchestrator agent. “This agent has instructions to call different agents to action when tasks are appropriate. The orchestrator interprets a command or chat from me, and then decides which agents it should prompt into action,” he explains. “These subagents have their own context window and receive prompts from the orchestrator to follow.”

Yes, this takes time upfront—but that’s the work required to create an agentic team that’s going to make your life easier in the long run. Hastings adds that it’s helpful to have a little tenacity and trust in the ability of the tool to get where you want to go eventually, and with the right guidance.

The best thing about this new approach to working with AI? Whereas complex tech skills can feel unattainable, this type of management is something you can definitely learn.

The best place to start? Regular practice, which Hastings thinks about as doing reps.

“Use the tool every day, give yourself a goal, and work with the LLM to figure out how to get to that goal.” You might start with the smallest possible unit of work within a workflow—writing a piece of copy, doing research on the performance of different ad campaigns—and see if you can get an AI agent to do it.

Ready to shift from operator to orchestrator? Join The Autonomous Marketer and learn how to build AI teams that multiply your impact—with practical tactics delivered every other week.

If you’re feeling stuck, try leaning on the LLM for help.

“There should be a sense in which you never feel blocked again, because you have the ultimate unblocking tool,” he says. “You can always ask it, ‘How do I do this?”

Finally, he suggests finding a community of others you can learn from.

 “AI is going to disrupt all of our professional careers, and it's better to go through that disruption together rather than alone,” Hastings says.  “We're going to learn the most by sharing enthusiastic adoption of the tools.”