Will AI replace marketing agencies?

It's Thursday afternoon, and someone on your team is building a shortlist. Forty creators, one spreadsheet, a column for followers, a column for what the last brand paid, a column for who replied. It'll take two days.

Everyone in the room now knows a machine could do most of it before lunch.

Good. That's two days back for the work your clients actually remember you for.

What AI takes off your plate

Let's be specific. These are the parts of agency life that machines now handle well:

• The first draft of a campaign plan.
• The first creator shortlist, and the second one after the client says "a bit younger".
• Chasing approvals, sending reminders and fixing the logo on the deck again.
• Building the monthly report out of screenshots and exports.
• The weekly status email nobody on either side reads.

None of those is why a client hired you. They hired you for your ideas, your taste and your calm when something goes wrong. That's exactly the part AI leaves with you.

One person with AI can match a team

Here's the study every agency owner should know.

In 2025, researchers from Harvard Business School and the Wharton School ran an experiment with Procter and Gamble on real product problems. 776 professionals were split four ways: working alone, alone with AI, in a pair, and in a pair with AI. The paper, called The Cybernetic Teammate, came out through the National Bureau of Economic Research in April 2025.

"Our findings reveal that AI significantly enhances performance: individuals with AI matched the performance of teams without AI, demonstrating that AI can effectively replicate certain benefits of human collaboration."

The Cybernetic Teammate, National Bureau of Economic Research, 2025

Two more findings are even better. Normally, technical people suggest technical answers and commercial people suggest commercial ones. With AI, that split disappeared, and everyone produced balanced ideas. And people said they actually felt better while working this way.

That's not a picture of agencies disappearing. It's a picture of a small team doing the work of a big one, with fewer handoffs and less friction.

A better process beats more talent

In June 2005, a chess website held a tournament with one unusual rule: anything goes. Bring a grandmaster, bring a computer, bring both. Teams of top players entered with several machines each.

It was won by two men from New Hampshire with no chess titles at all. Steven Cramton, 28, coached soccer at a prep school. Zackary Stephen, 24, was a database administrator. They played on three ordinary desktop computers. ChessBase reported on 19 June 2005 that they won the final 2.5 to 1.5, against much stronger players who had machines of their own.

How? They knew their programs inside out. They knew which one to trust in which kind of position. They picked a few candidate moves from their own experience first, split the analysis between them, then compared notes and argued it out.

Garry Kasparov, the former world champion, summed it up in The New York Review of Books on 11 February 2010:

"Weak human + machine + better process was superior to a strong computer alone and, more remarkably, superior to a strong human + machine + inferior process."

Garry Kasparov, The New York Review of Books, 2010

Read that as an agency owner. Process beat talent, and process beat hardware. A clear way of deciding, run by people who know their tools, is exactly what a great agency sells.

The best ideas still come from people

Leonard Boussioux, Jacqueline Lane, Miaomiao Zhang, Vladimir Jacimovic and Karim Lakhani ran a challenge for new business ideas and published the results in Organization Science in 2024. 125 people entered ideas. The researchers also had AI produce ideas, some on its own and some guided by a person. Then 300 judges made 3,900 separate ratings across 234 ideas.

"Our results indicate that while human crowd solutions exhibited higher novelty, both on average and for highly novel outcomes, human-AI solutions demonstrated superior strategic viability, financial and environmental value, and overall quality."

Boussioux and colleagues, Organization Science, 2024

In plain words: people came up with the most original ideas, and people guiding AI came up with the most workable ones. So the winning agency does both. People first for the big idea, AI second to test it, cost it and make it real.

And here's the commercial upside. When every agency in a pitch starts from the same tools, a truly original idea stands out even more. Originality becomes the thing everyone wants to buy.

What to change this quarter

1. Price around your thinking, not your hours. Move your fee onto strategy, ideas and results, while it's still your decision to make.
2. Give the tools to your juniors first. Erik Brynjolfsson, Danielle Li and Lindsey Raymond followed more than 5,000 support workers through an AI rollout and published the results in the Quarterly Journal of Economics in 2025. The newest people improved 34%. Your juniors will grow faster than ever.
3. Keep a human idea step. Write down your own thinking before anyone opens a prompt. That's where the originality comes from.
4. Name who checks the machine's work. Treat AI output like a draft from a fast new hire, and give someone the final proof, the same way you already proof every deck.
5. Learn your tools like the chess players did. Not which tool is best in general, but which one to trust for what. That knowledge is written down nowhere, and it's worth a lot.

Where Lumienzo fits

We built Lumienzo to take the paperwork off agencies' plates. Lumi learns each client and researches what's trending. It writes every campaign plan and the first ideas, and finds creators using verified audience data. It tracks every sample and every post, and builds the report your client wants to see. Every client's campaigns run side by side, and your team just approves.

Your ideas, your taste and your client relationships stay yours. That's why agencies are one of the five groups we built Lumienzo for. See how it works for agencies.