At the PRovoke EMEA Summit in London last week, Allison Spray, chief data and intelligence officer for EMEA at Burson, sat down with PRovoke’s global head of features Maja Pawinska Sims to discuss the transformative impact of AI on communications. Spray highlighted the shift from AI experimentation to implementation, emphasizing the importance of data in client meetings and the rise of agentic AI which is AI that can make decisions and act independently.  

She also discussed the integration of AI into workflows, such as using generative AI for daily message monitoring as well as addressed common misconceptions about AI, stressing the need for clear communication and training.  

The resulting conversation has been edited for brevity and clarity.

Five key takeaways:


  • Data is now a fundamental part of AI and will change the nature of conversations with clients.
  • Businesses are moving beyond experimentation to integrate AI into workflows, necessitating a central repository and creative studio.
  • Organizations must communicate AI's intended uses and provide training, while being prepared for both successes and failures.
  • Firms should leverage AI to anticipate crises and test responses, managing stakeholder expectations effectively.
  • AI can impact every stage of the communications workflow, offering vast opportunities for improvement.


Maja Pawinska Sims (MPS): AI is probably the most transformational thing that's going on at the moment. So let's dive in. The AI landscape in communications is evolving rapidly. We've seen a lot of hype and a lot of experimentation over the past two years. Burson’s parent company, WPP, has made no secret of the investment bet it’s putting on AI. But from your perspective, what are the key signs that we're now moving into a phase of real implementation and tangible return on investment for AI in comms?

Allison Spray (AP): I think that the first thing I would look at is the reality of what happens when we go into a client meeting these days. I've joked in the past that I've spent most of my career in measurement and evaluation, in addition to AI. And measurement is always something that slides towards the end of that meeting more often than not and the reason for that is because data is the fundamentals for AI, and it's changing the conversations we have with clients, it's changing the kind of work that we're able to do, because we can see more holistically, we can understand more fully, and then we can start to predict more accurately.

The second thing I would point to, and some of this, I think, coincides with the rise of agentic AI as well as Gen AI, which obviously came before. It is just the speed at which we're starting to see businesses take these beyond that experimentation layer. It's not just about creating a sandbox anymore where teams can go and play safely, but actually thinking about their workflows and saying, where is it within my workflow that I might be able to use this technology that's going to allow me to use it at scale.

MPS: Many comms professionals are both excited and still a little apprehensive about AI. What are some of the biggest misconceptions you see regarding AI in our industry, and how can we address them?

AP: I think it probably sits on parallel. So two ends of a spectrum effectively. So on the one hand, you still have folks who come into AI and expect it to be a magic wand that is just going to completely change what you can do immediately, and they'll never have to do that kind of thing again. And that's not quite how this technology works, and it's not really where we are yet, even with agentic AI.

On the other side, you have folks who try something once and it doesn't quite work, and so they kind of sign out. So I think a lot of this comes down to when we're starting to roll AI out in our organizations, and that can be true if you're working in an agency or client side. What are we hoping that you're going to use these technologies for? How are we going to train and enable you so that you can use it in the way that it's designed? And then how are we going to continue to feedback some of the successes and failures into the organization.

MPS: Communications is often seen as an art, but over the past few years one of the things I think makes this industry especially fascinating and special is that it's increasingly becoming a science. How are you seeing organisations leverage data, intelligence and AI to make more informed decisions and drive better results?

AS:I  think the first thing I'll say is in addition to measurement, one kind of discussion I guess I've had a lot over the past year, is this idea that it's not enough anymore to think.  You have to know, particularly in the world right now, that is so polarized. You might put out a piece of collateral and just never expect it to have the reaction that it does. And some of that happens because we do live within our own worldview.

We can't always see how other people are going to receive a particular piece of collateral, and that's where, again, I think the AI story becomes really interesting, because we can start to use these tools, whether that's a sort of virtual focus group, or we can create a series of personas to respond and give us initial feedback on an idea or a concept.

MPS: Data-driven decision making and insight is obviously brilliant for impactful campaigns, but the sheer volume and velocity of information is overwhelming for communicators. How can brands use AI to cut through the noise and identify the signals that truly matter to their brand and reputation?

AS: So I think some of it comes back to that question, are you looking into the right channels. This is going to vary depending on your brand, where your audience lives, but pretty much across the board, there's a lot more audio and video content that we need to figure out how to access, because that is where a lot of these conversations start.

Increasingly, what we're observing, particularly from a sort of reputational crisis perspective, is those conversations start, definitely on social and often in audio and video. So something will be creeping along in a relatively low volume on a channel like TikTok, before it gets large enough that a national media title picks it up.

Then it's a bigger challenge for you to solve where, if you were able to see that before it got picked up by the national media, you might be able to interject. And this is where, again, that marriage of AI and more traditional approaches is really useful in that we need to have a crisis counselors. We need to look at our behavioral scientists and say to them, how do you communicate with people who are in a very different world view? How do we think about that and nudging them towards a different perspective.

MPS: Stakeholder expectations are constantly evolving. How can organizations use AI to stay ahead of the curve and ensure their communications resonate with their target audiences?

AS: There's a couple of different ways, and I'll go from the sort of broader to the more proprietary. There's lots that we can do with large language models. Generally in terms of being able to look at more data and make that more accessible, to democratize access to data. What we can do is take data, for example, out of a platform or a syndicated audience data platform and we can look at what are our audience is telling us. What are their preferences, what do they care about. And instead of using a spreadsheet to look at that, look at that through an AI agent.

So someone can have a more natural dialog like where would this person go for information? What would they like to do on the Saturday night, that kind of thing. And then you get to the more proprietary technology that we've been developing where we can use tools to actually test both that believability and shareability, but also the potential contentiousness of an issue. So if we're looking across two diametrically opposed groups, how likely is this to create an issue so you can be prepared for what that response is likely to be.

MPS: Crises can erupt quickly and unexpectedly. How can organizations move beyond reactive crisis management and proactively identify and mitigate potential reputational threats?

AS: So AI is a tool for all of us, but ultimately what it's going to do is help us do our work better. So you need to make sure that your team is optimized to work together [in crisis], to understand the workflow, to think about how that works. AI is just a tool, both to help us, as I say, see more, see some of this content before it becomes a problem, but also to help us prepare and reveal some of those challenges, because you don't necessarily know how a team will operate under pressure until they are. 

MPS: Looking forward, what are the key skills you think communicators need to develop to thrive in an AI-driven world?

AS: I think the first thing is just inquisitiveness. Are you interested in learning more? Do you want to understand how these technologies can come into your day to day, and are you trying them out? I know I've talked about this, moving into implementation, but this is technology that is moving faster than anything most of us have lived through before, so it is going to be both constant experimentation along with that implementation in parallel.

Next, prompt engineering is going to be massive in the future. Thinking about how we help our teams look at prompt engineering, understanding why the prompt is so important to the output. Again, that helps address some of those misconceptions as well around AI, because if you understand how you need to prompt a system, you're much likely to get a better result.

And lastly, we all need to become more data fluent. I think particularly in PR and comms, there's been a little bit of shyness around numbers, around data. But the future of AI, you can't have an AI strategy unless you have a data strategy. So we all need to up our fluency there, because it's going to help us massively as we start to integrate these tools.