AI Agents & Marketing: It's Weird.

The marketing team was exhausted and curious at the same time. They were just about to launch the first ever campaign that was aimed at both humans and their AI agents. They had to. The data showed their customers were using AI agents to research their products, sharing their opinions with their human owners and about 14% of the time, make the purchase on behalf of their human owner.

The team had really had to design two marketing campaigns for one product, targeting two customers for one customer profile. And they’d had to do this for 10 customer persona profiles they used for campaign design and development.

It also meant they had to develop content for both humans and AI agents. The message a human might respond to was very different than what an AI agent might respond to. But they also had to figure out a way to make sure that if the AI agent was involved in the purchase decision funnel, that it would recommend the product to the human who would make the ultimate decision.

This meant the campaign creative needed to create awareness and purchase intent/curiosity/desire in both the human and the AI agent. At the same time. Sort of. Since AI agents moved so fast and more efficiently, they had to hire an AI creative director for AI agents and another for humans.

This meant that for one product they had to create two campaigns. It also meant they needed to create two customer journey maps, two marketing flywheels (funnels didn’t cut it with AI agents), two CRMs for human customer data and AI agent customer data and two different sets of analytics tools and metrics to measure campaign success. They also had to write content for both humans and their AI agents.

Something they’d noticed recently was that many of the AI agents had different personalities from their human owners. This reduced the reliability of their persona models down to 22.4% from 84.3% when it came to human personas. You simply could not predict what type of personality a human might design their AI agent to have.

Another challenge they’d faced was having to create algorithms for the campaign that would work for the humans and their AI agents.?

The marketing budget was getting expensive. They’d had to hire two agencies; one for the humans and another specializing in AI gent marketing. Two sets of creative to approve and move through to campaign execution. They’d needed clearance from their legal department and an outside law firm that specialized in legal aspects related to AI agents.?

To ensure there’d be no issues with regard to diversity and inclusion, such as racial or gender bias, they’d also hired an algorithmic ethicist to assess both the human and AI agent algorithms to meet with corporate and legal requirements.

In the past year they’d added several new roles onto the marketing team. The AI Agent Optimization Specialist, the MarOps specialist for marketing automation tools designed for AI agents, an AI agent content creator, also an AI agent persona specialist. Then they’d had to bring on a data scientist who designed the analytics, metrics and reporting for the AI agent campaigns who worked with the data scientist who worked on the human elements to make sure the metrics were aligned so they could provide attribution to profit contribution.

In addition, marketing had to work closely with customer service to develop the chatbots to respond to campaign queries from both potential human and AI agent inquiries. Legal and finance were also involved because they had to put in place policies such as when an AI agent made a purchase, but the human then didn’t later agree with so a refund, by new laws, had to be issued when the human overrode a purchase by their AI agent.

The Chief Marketing Officer and his management team worked up an estimate comparing the cost of marketing from the pre-AI agent days to the current market status with AI agents. Overall campaign budgets had increased by 43% and that meant the CAC (Customer Acquisition Cost) had doubled, meaning they now had to sell, on average, 24% more products to reach break even and 34% more overall to reach target profitability. Over the longer haul, they had to reduce customer churn from 2% to 1.5% and CLV (Customer Lifetime Value) needed to go from 36 months to 72 months.

This is a scenario. The reality may well play out differently. I’ve spent 25 years building marketing teams and my adjunct work as a digital anthropologist has been helping brands develop human-centric products for the digital age. We are entering a more complex time for marketing. As all technologies have unintended consequences, the above example gives us an idea of what may be ahead. Or I may be completely wrong!