AI Changed Everything About Reputation Except What Actually Matters

Speaker A: Foreign. So in 8 to 12 seconds, to be able to predict what any segment of the population in more than 63 countries around the world will likely think or react to has been a, ah, really a game changer for a lot of organizations. And I definitely see the portability towards making decisions about business.

Speaker B: Hello. We're here in Cannes at the Cannes Lions with all the big tech companies, all the advertising agencies, all the media companies. And I'm very delighted today to have someone whom I've wanted to interview for a long time, Chad Latz. He is the Chief Innovation Officer of Burson.

Speaker A: Thanks for having me.

Speaker B: Why don't you tell us first, what is Burson?

Speaker A: So Burson is a strategic communications company, and we're focused on helping our clients realize value out of reputation. And my responsibility at the firm is leading innovation there. So, you know, driving the different changes in adoption and AI.

Speaker B: So in layman terms, is Burson a, uh, PR agency?

Speaker A: Yes, I guess categorically we are a PR agency for sure.

Speaker B: Right, right. So a lot of people who are listening to us know, of course, what is a PR agency, but do not necessarily realize the level of change that is happening in PR or communication or crisis management. And, you know, in project management, we're very, very careful about stakeholders. Right. What they think about the project. In fact, a lot of the value that is perceived comes from stakeholders. So help us understand first, how is the current world changing? PR or stakeholder management?

Speaker A: Yeah, I mean, it's a great question. I mean, the reality is that we're living in probably the most volatile times in recent memory. Clients are constantly managing polycrisis, whether it's a cultural shift or something happening, uh, around the globe. And as a result, organizations need to figure out how they adapt to that change in a much more rapid fashion. And not only adapt, but be able to predict what's going to happen next. So as a result, it's changed the calculus we think about reputation, at least originally. It used to be a soft metric. And I think a lot of organizations did what they could with what was available to them in terms of those metrics. But once you're able to connect hard value to reputation, it really sort of changes the game for organizations and how they make decisions.

Speaker B: How do you measure reputation?

Speaker A: It was actually about a year ago here at can that we unveiled Reputation Capital, which is a proprietary platform and consulting framework and approach. And what we're doing is we're taking a look at unexpected shareholder returns directly attributable to reputation. I think one of the problems we saw was that a lot of organizations were using reputation studies or brand list studies. But by the time the report hit your desk, it was already out of date data. And we knew that there was a need to connect reputation more firmly in terms of unexpected shareholder returns for clients and for businesses.

Speaker B: Is it based on surveys that go out real time? How does it work?

Speaker A: So we've got an ensemble of models. We partner with the University of Oxford's Augmented Intelligence Lab to create the solution. It's about four different models and it's taking into account both fast moving and slow moving data. So what humans actually perceive and of course the volatility of the media ecosystem to be able to quantify the value of reputation against eight different levers, which are anything from products and services to innovation and creativity, to governance and to leadership. So organizations understand in this volatile environment their upside, opportunity and downside risk associated with movements. I think if you think about it Pierre, if uh, if a tweet can wipe out 13 billion in market cap overnight, there's uh, a greater sensitivity and need for these types of solutions for businesses.

Speaker B: Before we go back to maybe how this could apply potentially to managing projects and programs, because it's not just about corporations. Right, can you tell us a little bit about the evolution of the PR work? I read this interview from Martin Sorrell, right, sir Martin Sorrel, who said what really matters is what engines are saying, what the LLMs are saying about your company. And so you got to feed them with as much content as you can. It's a numbers game almost to influence the LLMs so that in return the LLMs influence people. Right, so what's your take on that?

Speaker A: So it's interesting. I think what's happened is that AI has become a mediator of understanding. And I think the thing that you're referencing is really the rise of GEO or generative engine optimization. And you know, as we're taking a look at that, all of a sudden, you know, it's not just a matter of where you show up in searches, it's how LLMs interpret the signals that drive brand perception. We're living in a zero click world these days. And as a result, you know, how machines read information and form opinion and then that opinion is then conveyed to uh, individuals is critically important. But we realized that there was also a fundamental flaw there too, because what it wasn't taking into account is the believability of the content among different stakeholder groups. So I think, you know, maybe to the point, I think the rotation that's happened is that there's a big preoccupation with being visible in generative engines and answer agents. Right.

Speaker B: Which is a number first step. Right. You want first to be there and then worry about what it actually says about you.

Speaker A: Yeah, that's exactly right. So, you know, of course you have to think about, you know, how do you create a, uh, structure of content? We've got a bit of a 3C's framework that allows us clients to think about that. But equally important, we published a report called a credibility paradox, which took a look at the believability or the credibility as a result of what was coming out of individual large language models. Eight different large language models across 85 companies. And we were able to see, start to see like those kinds of signals that were coming across. That helps. And, uh, you know, to go back to your prior point about how the discipline is changing, it's really helping us to think through how do we change content, what content is surfaced, how do we think about structuring reputation and reputation inputs into what LLMs are reading?

Speaker B: And so currently, based on what you're observing with all your clients, the 85 maybe that you. You've been referring to, do companies do a good job at influencing their reputation on LLMs?

Speaker A: I think what we're seeing here is that there is a lot of movement in the space. Um, there's a whole lot of preoccupation. It's becoming a bit of a boardroom conversation because it's a strategic signal that's feeding what reputation ultimately looks like for clients. So what we're seeing is a lot of interest in that area. What we're encouraging companies to do is to not just think about that as a sort of single audit moment. Like, you know, for example, what are the eight major large language models saying about my company or brand? We're thinking about this as an ongoing practice and a strategic signal that we take into account that then shifts communications, programming and approach.

Speaker B: So you're the chief Innovation officer of Burson? Yeah, Burson is the largest or maybe one of the most reputable companies in your space worldwide. So I'm going to. Let's try to do a little bit of lateral thinking. Many of the people listening to us today, manager, what we call Giga projects, projects with billions of dollars of investments that have massive impacts on communities, sometimes a lot of complexity in the political environment, you need to go find more funding, you run into unexpected issues and you want to make sure your stakeholders still buy into the project. Has Burson ever thought about Applying all these great things you're doing for brand reputation to a program reputation or a project reputation, I would go so far

Speaker A: as to say, you know, a company's reputation really impacts their license to operate, regardless of what sector they're in or what industry that they're focusing on. And I think one of the things that's really critically important is that you get those foundational signals to really understand what creates the opportunity and the risk. You understand and can quantify that risk. Then you can predict, of course, what different movements, as it relates to different vulnerabilities, project in the macro environment might generate for the company or the brand. And as you're able to predict that, you're able to then, um, strategically pivot your approach to make that happen. I feel like you're talking a little bit about how do you use data as part of a decision architecture.

Speaker B: I'm thinking more about your project, potentially is losing the trust of your main stakeholders. So can you anticipate that other techniques, other tools that you could use from what repetition management is about to handle that better for your big project, big program, you're building a bridge, communities start to get worried about it. You're building data centers, communities start to be very worried about that, the effect on their local power grid, on water, on, um, environment and everything. Right. So how do you. Is there a way to learn from what is being done in corporate brand reputation management and in fact, transfer this to almost a standard practice when you manage, let's say, a project of a big size that has potentially significant impact. And even if you think about more mundane things like internal projects, that sometimes you want to change entirely your entire ERP in the company. We know that transformation is hard and sometimes the effect of that project could be very negative on people's perception about the company. Right. So are, uh, there ways to use and leverage all your techniques and tools to program and project management?

Speaker A: Uh, I really love that question, Pierre. So I think, you know, one of the things that we've been really focused on is our work in cognitive AI. So with all of the preoccupation about generative AI, what we're really seeing is some power in cognitive AI, or the ability to predict how different stakeholders are likely to think, react or respond.

Speaker B: Okay, that's very interesting.

Speaker A: Yeah.

Speaker B: How do you do it?

Speaker A: So, you know, uh, we developed a solution called Decipher. We partnered with a cognitive AI company called Limbic. And what it looks at is, are two critical signals. It looks at the believability of content, messages, topics Themes, opinion leaders. And it looks at virality, the degree to which piece of content is likely to generate engagement. And we use hundreds of millions of artifacts to train AI models that will ultimately predict what will happen in the future, how an audience will likely react or respond to any stimulus in the future. And I think one of the things that's been really powerful about this solution we talked about sort of those, let's say, those signals around reputation, you also need sort of a fast twitch response mechanism or technology. So in 8 to 12 seconds to be able to predict what any segment of the population in more than 63 countries around the world will likely think or react to has been, uh, really a game changer for a lot of organizations. And I definitely see the portability towards making decisions about business. I think fundamentally, PR and strategic communications isn't just about sort of outreach and media engagement. It's a critical component in driving business value and understanding what's going to create the opportunity and the risk for brands.

Speaker B: How do you see all of these being impacted in the next 18 months by the increasing power that AI brings? Let's not think about 10 years down the road, but 18 months, right? How. What's your roadmap?

Speaker A: I think, you know, what we're seeing here is organizations really trying to maximize the value that they can extract from AI. If you think about how fast moving the conversation is, you know, I guess maybe if even over the last three years here, it can. The conversation began with, well, look at what AI can do. The, uh, conversation last year was, what should we allow AI to do? And then the conversation this year is like, what can humans do as a

Speaker B: result of doing so many things that we're struggling to catch up and now realize that we need to understand, in fact, in the end, what we are going to do?

Speaker A: Yeah, that's right. And the tension. I think people have this idea of, let's say AI being a collaborator, sat in and listened to a panel yesterday, and the fundamental notion was AI is not a collaborator. Collaboration is something that's reserved for humans. AI is a tool or an instrument that allows humans to do better work, to add value for brands. And I think that's a really powerful way of thinking. It's of it. You know, at Burson, we think of ourselves as an agency that's created for the intelligence era, and we're seeing AI give a lot of advantages for that.

Speaker B: So maybe you don't like Copilot being called Copilot.

Speaker A: Yeah, that's. That's interesting. Yeah, I think, I think all of us, you know, we're working with a lot of organizations to navigate the AI transformation. And you know, once executives bring the technology on board, they say, check. We've done that. You know, we've enabled the organization with the technology. What we're finding fundamentally is that workflows need to change. We've, you know, evaluated more than 230 different workflows, bundled them into different task classifications to quantify the value and the efficiency that could be driven by AI. But more importantly, how do organizations need to change to capitalize on that? And so to go back to your earlier question or point about where is it going, you know, as agents and more a to a or agent to agent interaction happens, we have to take a much more strategic view around where humans come into the equation. That idea of humans in the loop, I think we're getting a bit sour to that notion. It's more about humans at the helm and how we're directing the AI.

Speaker B: Yeah, it feels to me that we are at the point where we need to start deciding what we delegate to AI totally without any control, what we may want to delegate with some control with the human in the loop and what we don't want to delegate at all to AI. And that discipline of thinking and accepting that some stuff will be delegated without any control. I think it's not there yet. I'm not seeing that much.

Speaker A: Yeah, I think very, very low value tasks are easy to automate. But when business leaders are making critical decisions based on data and have to make critical movements, the idea of outsourcing all of that critical thinking to an AI, or let's say crit decisions in an end to end way doesn't feel right to a lot of organizational leaders. You know, they want to understand where the expertise is brought in and just. And from a strategy perspective, they want people who are bringing that combination of human plus technology to deliver, you know, the best answer to the best approach.

Speaker B: You know, one thing that I've noticed lately is now I have to edit a lot because I'm editing AI, which is very frustrating. How do you protect yourselves against that kind of lower race to the bottom in your own agency?

Speaker A: I think a thing that a lot of organizations are grappling with, they're seduced by the speed. And I think the way to think about the workflow is that there are certain aspects of working that are accelerated by AI, uh, first draft, sort of early ideation. But then you might find yourself actually spending more time on the strategic refinement of the outputs to get to something that's really high quality. And I think at the end of the day, whether there's technology involved or not, the critical question is are you producing something that's a high quality output that's going to make an impact for your business or your brand? And that was a call to all leaders long before the technology became so prevalent. So I think one of the things that we're looking at, it's not just human in the loop, it's taking a critical look at what uh, is being produced by the large language models. A lot of them of course are you think about averages of the Internet, right? You know, a lot of accumulated knowledge and that might be very valuable for condensing phases of activity like research. But you know, some of the deep critical and strategic thinking that needs to happen still needs to be driven by humans and you know, guaranteeing that quality of what's coming out of the models before you sort of put it out into the world. We talk about AI slop. It's a real problem now.

Speaker B: Uh, and what is your message for your own teams? Do you let them, encourage them to use AI, Let them use AI, tell them not to use AI, what's your own message? How do you see your even profession evolving? And I'm saying this because of course there are many perils with project management and project managers. Uh, do we encourage people to do everything with AI to go faster and then just check? Or do we tell them no, you can't do everything with AI, don't do it.

Speaker A: We sent an organizational mandate uh, a couple years ago and we have an initiative called Future Work which is really about enabling our workforce, leveraging AI. But we also delivered that as a consulting practice practice to our clients. And you know, a call to action was really to get everybody engaged in the technology and the platforms and we hit 100% adoption by our professionals last year. So I'll start by saying we are strongly encouraging our organization and our team members, uh, to leverage AI. And so much so that we've of course empowered them to build their own agents which has been a really how

Speaker B: do you do it? Like you, they have cloud and they build their own agents. How does it work?

Speaker A: So we have an end to end AI powered platform by uh, wpp, which is our parent company called WPP Open. And we've literally put the ability for individuals to build agents at their own discretion to serve different work tasks.

Speaker B: And it's based on which like LLM,

Speaker A: they have availability of all of the large language models.

Speaker B: Okay, so they can build agents based

Speaker A: on whatever they want, whatever they want the underlying model to be, whatever they feel delivers the best output. And for us, Pierre, that was a game changer. Once you gave everybody the capability to think about how they could best apply AI and create themselves, it takes a lot of the anxiety out of the process. You know, I think we work with a lot of organizations and the, uh, anxiety around AI is like, well, how's the technology going to change my job? Is it going to impact whether or not I have a job? When you empower an organization to leverage best in class tools and technologies to build solutions that help them deliver their own work, in addition to giving them things like super agents that allow them to do common tasks, it's a fundamental game changer. That was what got us to the level of adoption that we have.

Speaker B: It's interesting you're saying that. The other day I was talking to a young partner in one of the large management consulting firms, and they had received two weeks earlier a message from their CEO saying, the same thing. You can now use our platform, develop your agents at your discretion. And then two days later they received another message saying, well, in fact, I didn't mean what I said. There are some restrictions. You can do this, you can't do that. And the rumor is within the company that it's because on day one they had over a million dollars of cost coming from the tokens. So how does it work at WPP or Burson? Ah, Are you not afraid of a free management of agents and the consequences on the cost of operating your business?

Speaker A: It's a reasonable question. You know, what does it cost to actually run these models? And there's been a lot of high profile examples of very expensive token usage. We burned our whole year's allocation of tokens in a month's time. You, uh, know, I think that that's something.

Speaker B: My CFO is very worried about that, by the way.

Speaker A: Yeah, for sure, for sure. Uh, understandably so. I think, you know, part of what organizations are doing is taking that into the business planning process. We're of course, evaluating that effectively and we're also taking a look specifically at the data on how people are utilizing different agents that either they're creating or that we're creating and we're bundling them into clusters. We're seeing whether or not there are common tasks that many people are doing across the organization and as a result build new agents that are deployed globally. So it doesn't require that individuals spin these instances up on their own. And, you know, we're starting to get baselines around what, uh, the spend is. You know, we work against an allocation for technology and for tokens that we monitor on.

Speaker B: Is it exploding? Right now you're seeing your cost exploding?

Speaker A: You know, I think what's happening is that frontier model producers are also looking for ways to make the models and make a bit less expensive, I think, because of all of the concerns that you have. So, you know, there's this constant adjustment. You know, every week, uh, there's a new upgrade to a new frontier model. And so part of what we do on the innovation side of things is track and evaluate the progress associated with them, um, the performance versus cost, you know, and maybe you don't actually have to run the latest model to deliver a solution. You can run a, uh, model that's a little bit cheaper, a couple generations older, and you sort of assess the models for the use cases. But, you know, we've, we've sort of moved past the point of driving enthusiasm for the technology. I think our organizations and our clients are seeing the real value and what comes as a byproduct of using it. But you have to direct it against the right use cases to realize that value.

Speaker B: Thank you. I'd like to go back a little bit to your platforms.

Speaker A: Sure.

Speaker B: Reputation Capital Decipher. Sounds to me that you're accelerating the pace of feedback from stakeholders from the environment so that you can adjust faster. At, uh, pmi, for many years, we've been working around the concept of enterprise agility. We recently launched a manifesto for enterprise agility, where we defined enterprise agility as the ability to react, to change, or proactively transform, but without losing coherence. So how do you manage this tension between constantly worrying about what people think and remaining steady with your strategy? Because you can't react to every single opinion or every single shift in the mood. So how do you avoid creating more anxiety, more volatility? By giving people almost too much information

Speaker A: Too often you make an excellent point. Not all things are equal. And I think one of the things that AI has allowed us to do and our innovation portfolio that we've shaped is able to quantify that opportunity, the upside opportunity and downside risk, and predict what will happen when a particular situation comes to pass. And the thing that we really never had before was the ability to get that intelligence in 8 to 12 seconds. Whether or not organizations can move at that pace becomes another question or another consideration. And that's where you start to think about, well, are there next steps that fall out of the Things that we're detecting, for example, through, um, you know, we have a proprietary social intelligence platform platform called Sonar which detects, you know, signals from all aspects of the Internet. We then shift that into prediction. Well, what do those signals mean in terms of their potential for impact on a client's business or business outcomes? And what's the follow on impact to reputation? So reputation capital, we think about these things in a bit of orchestration and how we pass the data and the information not only between the solutions themselves, but back to executive stakeholders to drive key decisions. And you open the conversation by talking about the fundamental shifts in pr. What we're seeing is that this level of intelligence to drive strategic communications is becoming a board level priority. If all of a sudden you can walk into a room and say, our reputation is worth $13 billion and here's what our risk is against this particular lever, it's really a game changer as opposed to, you know, getting a survey back that where the data was fielded over the course of several months with no sort of real clear signal around what the actual impact of business is. And so, you know, inside of organizations you also have to be prepared to react quickly.

Speaker B: Yeah, I think there's always been this notion, uh, of crisis management. Right. It's where you've got to be very fast. What I'm hearing is that even when you're not in a crisis situation, having real time data about your reputation can bring a lot of value.

Speaker A: Absolutely. Uh, I think one of the things that we know that organizations focus on a lot is how do we become part of the culture? How do we become part of the cultural conversation from an opportunistic standpoint to generate, uh, value and positive movement for the brand. So a lot of the intellig intelligence and technologies that we've talked about today are things that we're using to fuel different aspects of our other practices. We've recently released a framework point of view called Culture up, which looks at those opportune moments. It sort of shifts the paradigm from issues in reputation management and response towards how do you take advantage of cultural moments to build brand value. So I think you articulated it really well. It's in a way these, these technologies, these capabilities allow us to take a look at two different sides of the same coin.

Speaker B: Chad, I have one last question for you. Uh, how long have you been coming here to Cannes?

Speaker A: Oh, it's been, uh, you know, probably a little, a little more than a decade for sure.

Speaker B: So if you look at the mood today compared to the mood last year, Two years ago, three, four, five. Like, what is the trend? Like, what's, how is it different?

Speaker A: I think one of the things would, uh, probably be of no surprise that I'm focused a lot on technology and technology executions.

Speaker B: But this is Ken, right? So everybody's here, all the big AI companies, tech companies. I used to come here, we had advertising agencies, media agencies, and then they got completely, completely taken over by all the big tech companies. And now we're seeing all these AI companies everywhere.

Speaker A: Yeah, well, it's because the intersectionality of what the technology can enable through the lens of creativity. In the earlier years, we were seeing examples and instances of where AI was applied to, uh, a creative execution. And those were sort of interesting and novel uses. As we discussed, we started to see the conversation shift a little bit away from like, oh, well, look at what this AI can do and more towards what should we allow the AI to do? And I think there has been, at least for the last several years, this tension between technology and talent, technology and human capability, which I think is obviously going to continue to be a trend as we move forward. I'm really excited about taking a look at some of the work and seeing how technology is being applied. We had this. I, uh, had a funny discussion with a colleague and we said, so, you know, are we finally, at this moment, where AI has to be less central to the conversation? We no longer ask the question, well, was this thing produced by Photoshop, for example? The technology sort of fades into the background and the human capability and creativity is brought to the fore and accelerated. And so, you know, we think about, if we sort of think about that through the lens of reputation, we feel strongly that reputation is earned and creativity is a big part of that. And so, you know, we're, you know, for me, what I'm looking at is that constant mix and the tension between technology, creativity and human capability, which is obviously constantly changing.

Speaker B: Right. And if you had one piece of advice for our young audience, so people who just maybe are still in college or, uh, are looking for their first job or just started with a company, what kind of advice would you give them? In this very unknown, uh, world that we're in? Right. The evolution is so unclear. So what piece of advice can you give them? Whether they're currently in Mumbai, in Hanoi, in Seoul, in Rio, in New York, in la, wherever you are in the world right now, or in Europe, here, in Paris, in London, what is the piece of advice you would give them?

Speaker A: So I think that there's a lot of anxiety, particularly among emerging Professionals as a byproduct of what AI is doing to industry and what we're seeing inside of academic institutions is a similar kind of tension. Do we step away from sort of traditional teaching practices and move more towards enabling our students with AI and technology? One of the things that I would strongly encourage, both in the academic environment, but also as an emerging professional, is to dive headlong into utilizing and utilizing the agentic AI.

Speaker B: Right, not just conversational AI.

Speaker A: That's exactly right. And in addition to that, seek out the human interaction, mentorship and engagement that will allow you to develop in your professional career, learn from people in addition to learning from the technology. Sort of an interesting thing that we observed, Pierre, Actually, in the earliest days of our organizational adoption of AI, it was actually our youngest cohorts that were the most hesitant about utilizing the technology.

Speaker B: We're seeing this in project management.

Speaker A: Sometimes it's a bit of a head scratcher. You know, we're thinking, oh, you know, like this is a generation that is inherently digitally native. We would have expected that, uh, they'd be using the technology all the time. What we realized pretty early on is that they were experiencing that tension inside the university setting. So, you know, in first moments, we wanted to just encourage them to utilize the technology as part of what they're doing, but to bake that in as a total academy approach at the organization. So I guess to your question about advice, seek out and utilize the power of technology, but don't abandon the human relationships that will also be required to learn.

Speaker B: But the point you're making is making me think about something. Some companies do not allow people to use AI that much internally. They are very slow because they want to control, they have confidentiality concerns and all of that. Do you think today it is difficult to attract talent if you don't let them leverage AI almost as they wish?

Speaker A: Yeah, absolutely. Without question as they wish. We'll let that maybe remains to be seen, but as a critical component of what you're offering as an employer, it's absolutely essential. And I think organizations that are not embracing those approaches, fortunately, what we're seeing is even in the most highly regulated industries, with the right guardrails, the right guidance, the right organizational transformation approach, AI acts both as a talent attractor and also an accelerant for the business. So I think, you know, even the highest regulated industries, we're seeing obviously significant uptake in the use of the technology. So, gosh, if you're not utilizing these technologies and you haven't enabled your organization to do so, you're going to have a whole lot to do to catch up with the rest of industry and

Speaker B: even the reputation problem, potentially 100%. All right. Thank you, Chad.

Speaker A: It's a pleasure. Thank you.

Speaker B: It was a great conversation. Thanks for your contribution, and I hope you enjoy your time here in camp.

Speaker A: Thanks so much. You as well.