AI is rapidly transforming the way agencies work, from strategy and insight to creativity and customer experience. To explore what this shift really means in practice, FutureWeek spoke with Ed Freed, Global Chief Transformation Officer at RAPP. In this conversation, Ed shares where agency employees are getting AI use wrong, if GenAI actually creates compelling narratives, and how best to use AI agents.
Which AI use case are you most proud of so far?
One standout use case for our company has been our content intelligence and dynamic creative optimisation, powered by AI. Content Intelligence analyses every piece of content a client produces, along with its in-market performance results, and can understand at a pixel-by-pixel level what is driving performance with an audience.
We then use GenAI to combine this analysis with individual customer interactions to optimise the next bit of content they see, automatically adjusting campaigns in-market to maximise impact. This approach has not only dramatically improved campaign performance – like boosting engagement rates, ROI, and media effectiveness for our clients – but has also freed up creative teams to focus on strategy and conceptualisation rather than execution.
What does it mean to work collaboratively with AI?
Many people approach AI as if it were a tool like Word or Excel, as something where you enter an instruction once and expect the perfect result immediately. When that doesn’t happen, it’s easy to assume the AI isn’t capable. In reality, the real power of AI comes when you treat it less like static software and more like a teammate or assistant.
Working collaboratively with AI means engaging in a back-and-forth process: presenting a problem, letting the AI ask clarifying questions, refining your inputs, and iterating toward the solution together. For example, if I’m given a complex challenge by a client or internally, I might ask the AI to help me break down the problem, highlight gaps, or suggest approaches. This kind of interaction is conversational and iterative, not a one-off transaction.
How will this change the way people in agencies work?
AI blurs traditional boundaries between roles. Strategy, data, creative, and technology functions increasingly overlap when anyone can generate documents, imagery, or even technical explanations with prompts.
While empowering, this can also dilute quality and risk bypassing the expertise of specialists. That’s why agencies must adopt structured ways of working with AI. AI can often get you 70 percent of the way there, but the final 30 percent – human judgement, subjectivity, and craft – is what ensures quality and differentiation for clients.
How is AI reshaping agency work?
The first is strategic insight, where AI can pull together multiple data sources, including structured performance data and unstructured market and competitor intelligence, into client-specific datasets. What used to take days can now be achieved in minutes, which means repeatable knowledge work is automated and our teams can focus on higher-value, strategic interpretation.
The second area is generative creativity. While standard tools often produce generic, stock-like outputs, our focus is on training AI with client-specific content, such as brand imagery, voice, and guidelines, so the outputs are not only high quality but also on-brand and campaign-ready. This approach also opens the door to generative personalisation, where creative can be generated in real time, tailored to individual customer behaviours and needs.
Finally, we are exploring new frontiers in creativity. AI excels at remixing existing material and predicting what content will resonate with audiences, but it cannot yet achieve true invention – the subjective leap that comes from human intuition and hunches. That is where human creativity remains essential, and where the most exciting possibilities lie: combining AI’s ability to process, remix, and extrapolate with human ingenuity to generate ideas that go beyond the obvious. In short, AI is automating the repeatable, accelerating insight and content creation, and enabling agencies to move up the value chain – focusing less on routine execution and more on innovation and impact.
How are you using AI agents?
AI agent technology is maturing incredibly quickly, and we’ve already experimented with its potential. About a year ago, we built a proof of concept where a chain of agents could take a client brief, conduct research, translate that into a creative brief, develop campaign-specific executions, and even generate assets like voiceovers and video segments – all with no human intervention. While the experiment proved that this kind of end-to-end automation is possible, it also demonstrated the risks: when we shared it internally and with a client, it was seen as too extreme and was ultimately rejected. The real question with agents isn’t whether they can run autonomously – they can – but where you introduce the human into the loop.
Our view is that agents are most powerful when used collaboratively, as an extension of the team rather than a replacement. For example, in our structured workflows, we might use an agent to take a brief, conduct overnight research, and deliver a package of insights for strategists to build on the next morning.
Can GenAI actually generate creative that’s compelling?
It really depends on how narrow the use case is. For example, in paid media we can already see how GenAI could slot into Dynamic Creative Optimisation (DCO). Instead of orchestrating a fixed set of content variants and letting the system choose the winner, GenAI can create an inexhaustible stream of new variants and optimise them continuously against audience performance and cost-per-acquisition. That kind of narrowly defined use case works well and could even be run all year round with little intervention.
But when it comes to delivering a cohesive customer experience across channels, things get more complex. Most businesses are still organised in silos – social, CRM, paid, owned – and as Conway’s Law suggests, those internal structures are reflected externally in fragmented customer experiences. GenAI can help at the channel level, producing personalised content such as email variations based on recent user behaviour. However, ensuring that all those touch-points add up to a compelling, consistent brand experience requires human judgment and structured orchestration. Agents can support by automating narrow tasks, but for broader brand storytelling – like a major product launch spanning multiple channels, you still need people in the loop to ensure the creative holds together.
What’s been your number one challenge when it comes to AI implementation?
The biggest challenge has been the speed of change. Traditional software projects often run on long timelines, but with AI, that approach simply doesn’t work. We’ve seen clients spend a year or more developing an AI system, only to find that by the time they’re ready to release it, the technology has already moved on and their solution is outdated.
That’s why AI projects need to be handled by small, agile teams working in short cycles – no more than a few months – with a focus on proofs of concept, rapid iteration, and the ability to pivot quickly. Agencies tend to adapt more easily to this pace, but many clients struggle because their organisational structures and governance models aren’t set up to support fail-fast experimentation. It’s particularly difficult in industries like banking, where leaders are asked to commit to major ROI projections while also facing the reality that some cutting-edge AI initiatives have a high chance of failure.
Why is it so important to be agile, especially right now?
Over the same period we’re putting agents into our workflows, people are going to start carrying agents on their phones. And once that happens, the whole paradigm shifts. If my agent already knows me – what I care about, what I don’t – why would I sit and read every brand email myself? I’d just have the agent summarise what’s important.
That’s going to fundamentally reshape how brands connect with people, and in my view, it’ll spark almost a renaissance in customer relationship management. So while we’re busy thinking about how agents can make our own work cheaper or more efficient, the real game is preparing for the fact that customer interactions themselves will change radically. That’s why agility is so important – it’s about being ready for both sides of that shift.
