December 9, 2025
The World Media Group recently hosted its second Agency Future Leaders event, where Antony Mayfield, CEO of Brilliant Noise, shared practical advice for marketers navigating the AI landscape. Here are Antony’s top tips for young leaders – or anyone looking to harness AI’s potential.
Start with AI literacy, not tools
Mayfield was keen to stress that AI literacy isn’t about learning to code or mastering the latest platform. It’s about developing something far more fundamental: your own thinking.
“AI literacy is an evolving set of skills that’s about understanding what AI systems can and can’t do, where the limits are, which one’s right for [the task]. But it also involves critical thinking,” Mayfield explains. “This is a very different kind of tool, of computing, to other kinds that have come before. It really works extremely well with our own thinking systems.”
The challenge is that we rarely examine our own thinking processes. Mayfield compares it to delegation – when you delegate a task, you must unpack your automatic processes and explain what you do instinctively. Working with AI requires the same skill.
The three levels of AI mastery
Mayfield breaks AI competency into three distinct levels:
Level 1: Prompting
Learn how to use structure prompts that bring out the best in the system you’re working with. Mayfield dismisses the term “prompt engineer” as tech people trying to overcomplicate things when what’s actually required is simply the ability to explain yourself clearly.
Level 2: Tool-making
Create tools that work for your mind. “The best way to think of AI is not as a tool, but as a tool-making tool; it works with your mind.” Mayfield says. He recommends using features like Claude Projects to build customised assistants for specific tasks. Mayfield shared his own example: “I’ve got a tool that’s like a copy editor. It’s based on a really acerbic, harsh copy editor at Random House…because I don’t want an AI to be nice to me.” He also has a fact-checking tool in Chat GPT that finds statistics and claims, providing references for verification.
Level 3: Agents
Once you’ve built individual tools, that’s when you’re ready for agents. These are AI systems that can use other tools, remember context, and work towards objectives rather than just completing single tasks.
Until you’ve made some things for yourself that are making your work better, Mayfield advises sticking to one platform.
Make it daily practice
Mayfield’s magic formula for improving his team’s AI literacy is lifted from Google’s approach. The tech giant mandated that all employees, technical and non-technical, spend an hour daily using Gemini AI. Weekly, teams of eight held ‘Show and Tell’ sessions where everyone shared something they’d discovered.
“Our most popular meeting of the week is our Show and Tell,” says Mayfield. “Everybody does five minutes, and it’s accelerated everybody’s learning. Seeing it is what changes minds and accelerates learning, not having it explained to you as a theory.”
The key is deliberate practice, just like learning a sport, instrument or language. And the work you’re doing doesn’t have to be professional. Experimenting with personal interests helps you understand the boundaries.
Challenge and test relentlessly
Mayfield also recommends arguing with AI systems. Push them around and don’t accept the first answers. “It is not like any technology that’s come before – like Google or a database where it gives you an answer, it’s giving you what it thinks might pass for an answer,” Mayfield explains. “It is a conversation.”
Test the system with topics you know well so you can sense-check the responses. This helps you to understand where AI excels and where it falls short.
Focus on real problems, not hype
Mayfield’s work with BMW demonstrates AI’s practical impact. His team now handles hyper-localised content across multiple European markets – something previously too expensive or complex. They’ve created approximately 900 content assets across different motorbike models, personas and markets, with localised messaging for each.
AI doesn’t replace human judgement; it helps to manage the complexity that makes personalisation viable. “The complexity of making all of that work is the bit that gives everyone a headache or means you have to invest a lot of money with technology,” Mayfield notes. “But you can also create your own way of doing it.”
Address client concerns head-on
When clients resist AI adoption, Mayfield identifies two types of objections: emotional responses based on unfamiliarity and fear, and rational concerns often based on outdated information.
For GDPR worries, he points clients to compliance documentation. For concerns about data sharing, he clarifies that enterprise versions of tools like ChatGPT operate as locked-down, local instances.
Mayfield’s fundamental pitch for learning and investing in AI is that “AI literacy is the one thing that’s not going to go out of fashion, because all the platforms are changing so quickly. But humans understanding it and being able to gauge whether it’s BS or whether it’s useful – that’s a really good investment.”
Antony Mayfield was interviewed by Emma Ball, Senior Commercial Director, Europe & APAC, Reuters Professional.
For more insights from the Agency Future Leaders event, read our companion piece on Overcoming Imposter Syndrome and Building Unshakeable Client-Agency Trust.