(MENAFN- PRovoke) The future of crisis communications will involve massive amounts of data provided and analyzed by artificial intelligence, which will also have the ability to assess a wide range of scenarios and predict the impact of specific messages on tightly defined audiences-giving the human communicators involves a higher degree of confidence in their chosen strategies.

That was the core message of a session at our PRovoke Global Summit in Chicago earlier this week, moderated by Jim O'Leary, CEO of Weber Shandwick North America, examining“The New Architecture of Crisis Management” and including a demo of the agency's new crisis management agent.

The session began with a prediction about the impact of AI on the profession that echoed some earlier panels:“We're in a moment in our profession where there is going to be more change in the next five years than there has been in the previous five decades,” says O'Leary.“And of course, one of the main reasons for that is the AI revolution.”

O'Leary pointed to research conducted by Boston Consulting Group and presented earlier during the event:“Based on the analysis they did, they've determined that 80% of the work that we do in communications fits into what they call the AI collaboration sweet spot, where AI is a force multiplier for the work humans can do on their own.”

Those thoughts were echoed by Shannon Thornton Susko, former chief communications officer of Centene Corporation:“We've all seen the kind of scary headlines about reduction in force from AI and opportunities to optimize the function through cost cutting mechanisms. But the reality is, we actually have a unique and timely opportunity to optimize the function in ways that leverage and bring the best of human led and AI assisted functionality and being able to focus on things that can drive greater value across the organization.”

The upside, she says, is the potential“to leverage AI to increase your speed, your agility, your precision. A human led AI assisted capacity is going to bring you the ability to optimize your function, both during normal course and transformative periods.”

She acknowledged that while some communicators are experimenting with AI, many remain hesitant:“You have organizations who are like, 'in '26 we're getting Copilot,' and you have organizations that are running agentic AI solutions and using LLMs and bots to be able to advance the work they're doing, not only within a comms function, but much more broadly, from an operational perspective across their enterprise.”

So while there are a wide range of new tools being developed,“I think the ability to optimize and leverage the use of AI in corporate reputation and risk management are critical. The ability to drive real-time response that is more precise based on data and insights and gives me an edge of information, whether I am dealing with a risk that is local or global.

“AI particularly for reputation management and crisis preparedness is going to help put us in a better position. Five years ago, we would be doing research for weeks on something that we can do now in minutes, and so the ability to make decisions on a dime in real-time that are data backed, insight, informed, and again human-led, AI-assisted.”

O'Leary discussed Weber Shandwick's technology partnership with Google and the development of the agency's Halo“AI spine.”

“Within Halo, we are building agentic systems that are very specific for certain things, including crisis and risk management. The simple way to think about an agentic system is that it is built with a ton of very specific task agents that are operating autonomously, autonomously, and they're orchestrated together.”

Dustin Johnson, global chief commercial and innovation officer, emphasizes that the tool“is powered by custom data partners and decades of Weber knowledge built into the system. The last thing we want is the sea of sameness that comes with traditional, off the shelf, LLMs. We build our data into this and our knowledge. And we partner with our clients in very secure environments, Google secure environments, to load their data in and their knowledge in. So when we're working with them, everything is truly customized.”

Setting up a demo of the crisis system using a hypothetical client, Johnson explains:“We already have our monitoring agents going out and looking across the web for potential risks. Once it's met that threshold, our agents will create a crisis detail page and immediately fill out all the information that it has, an assessment by customer, employee and investor. It's going to tell us, how fast is this thing moving? How much virality does it have, how much influence does it have? And then ultimately, what's the kind of boycott or emotional intensity risk of the thing?”

Johnson says there are 10 different agents putting this information together to ensure it as detailed as possible.“And then lastly, the agents also take a cut at multiple scenarios. So what if this takes off in virality? What if it maybe kind of fizzles out, if we don't want to take too much action, because that can actually create more damage. The system has already taken an assessment at which of these it thinks is highly likely and what the probability is,” adding a new level of analysis to traditional scenario planning.

Patrick Chaupham, head of applied innovation and I/O consulting, chimes in:“The human in the loop is more now in the business of context engineering. We've talked a lot about prompt engineering. We're actually moving into an age of context engineering, which is what is going on in the world, what's going on with stakeholders, and then what are you gonna do about it?

“So now we can start to map these scenarios into how we would manage a crisis, you know, using our own expertise.”

In the specific scenario that forms the basis of the demo a multinational company has a production issue, and has also been making layoffs against a backdrop of activist activity. And one of the laid-off employees has started a Go Fund Me page. “They were laid off, and the employer is choosing not to contribute to it, which is causing other chattering across the system,” says Johnson.“And you can see the boycott intensity actually increased as emotional intensity increased because of this.

“So the system immediately goes out, takes the context of the crisis and looks for comparables, looks at other crises that it that are publicly available or within our knowledge base that we load into the system and ask, what can we learn from this? What can we apply just another layer of knowledge, another layer of intelligence that we give to the practitioner to help assess the situation.”

The system is then capable of generating initial assets. Says Chaupham:“Based on the analysis scenario, the input that we've gathered from media landscape and our client knowledge, it generates a holding statement. It generates a press release, some Q&A, some social posts, a summary of the timeline, as well as a letter to employees.

“This is just the agent's assets, so the comm strategist has a starting place just to react to. This is not something we just automatically publish.” The system will also offer advice on how different stakeholders, from employees to investors to activists, are likely to respond to different statements.

Adds Johnson,“This just gives us another layer of validation confidence. It gives us high confidence in advance for whatever response we make.”

And O'Leary emphasizes that the system can provide a high level of granularity.“We have built this with custom synthetic audiences that are much more precise than the three that we're showing here. So we could literally have a synthetic audience loaded in here that is the exact community that is affected. You could have one for the local political envriment, for the residents and first responders. And then you can be testing your response against all of those.

“You can test everything that you can imagine against those audiences before they go out into the world.”

Adds Susko:“If your CEO or the board calls and wants to understand what are the implications of engaging or not engaging, you now have a real-time data backed insight, informed way to answer. That's not to say that you're always going to know 100% but you're going to know a lot faster and with greater conviction than you would otherwise.”

Ultimately, the tool exists to inform the human decision-maker, rather than replace that person. Says O'Leary:“There should never be a scenario where the human is not the final reviewer. Because we've all seen the tech hallucinate. So we would never allow for a scenario where you would just go out into the world with something that you know, the system is auto generated for you, without that proper review.”

Though even there, there is a role for the technology:“We have agents, their entire jobs are essentially to challenge the results of the other agents that have created the content. You can even have agents that are designed to correct maybe just your own personal unconscious bias.”

At the end of the day, the balance between machine and human input into decisions might change, but both will have a critical role.

MENAFN06112025000219011063ID1110307707