Internal communicators are used to stakeholders questioning a draft. AI changes the shape of that challenge.
A senior leader can now take a carefully developed announcement, ask an AI tool to improve it, and come back with a rewritten version. Or a communications plan. Or both. The result might be useful. It might also be confidently wrong.
That leaves communicators with a new question: what is their role when senior leaders across the business can produce content, critique messages and develop communications ideas for themselves?
This was the starting point for Poppulo's latest webinar, Everyone's an IC Pro Now: The New Stakeholder Tension When Every Exec Has AI. Moderated by Poppulo's Andrew Hubbard, the discussion brought together Shel Holtz, senior director of communications at Webcor, and Jenni Field, founder of Redefining Communications.
The panel was not arguing that executives should stay away from AI. The question was how communicators can distinguish useful input from plausible-sounding noise, preserve productive stakeholder relationships and retain professional accountability for the final communication.
Is this happening everywhere?
Holtz began by saying he had not seen the issue emerge at Webcor—for good reason. That is largely because the company has a clearly defined review process. Stakeholders are asked to check communications only for accuracy. They are not being invited to rewrite them according to personal preference.
Field's experience, particularly with her UK clients, is different. She is seeing leaders use AI to draft communications and build plans, although less often to critique finished comms.
That distinction is important. This is not necessarily a universal problem yet. But it is a very real possibility, especially where the role of internal comms has never been properly explained.
Field's concern was that some teams have allowed themselves to be seen mainly as content producers. If that is what stakeholders think comms does, then AI appears to offer a quick alternative to sitting through a meeting or waiting for someone to explain the wider context.
What that points to, more than the tool itself, is the strength of the relationship behind it.
The value of comms cannot just be the words
Field kept returning to the importance of credibility and trust. Communicators need to show stakeholders why they ask the questions they ask, what expertise they bring and how they improve the outcome of a communication, not just the wording.
That work has to happen before there is a problem.
If leaders have never understood why comms needs to explore the impact of a message or challenge an assumption before it goes out, they are unlikely to see the point when AI produces a fluent draft in seconds. The tool becomes the easier option.
Holtz took the discussion back to his time at Mattel in the 1980s, when desktop publishing gave every department the ability to produce its own newsletter. Before long, people were receiving fifteen, twenty or thirty of them, many containing information that had nothing to do with their jobs. The technology had made publishing easier, without making communication any better.
This, he argued, is why internal comms needs to act as "message mission control". Someone has to help the organization decide which tools are appropriate for which messages, and stop every new channel becoming additive.
AI belongs in that same conversation. Internal comms is not simply there to write things. It should help the organization work out what needs to be communicated and how best to deliver it.
Some things can be good enough
Neither speaker suggested that every piece of communication needs to be crafted entirely by hand.
Field made the point that communicators can sometimes become too protective of their own process. There is a danger in assuming that anything produced with AI is automatically bad, or that a different method must mean an inferior result. Not every hill is worth dying on.
Holtz agreed. He is comfortable using AI to help produce a short item for a wellness newsletter, for example, as long as he checks the facts and edits the tone. That is very different from handing over a message about organizational change.
For more sensitive communication, he writes the message himself and then uses a tool he calls his communications "red team". It tests the message against different types of employees, including people who may be difficult to convince or hard to reach.
That is a useful distinction. AI can help a communicator think through a message without becoming the author of it.
Field also warned about what she called "AI slop": writing that looks polished but leaves the reader with the real work of deciphering what the sender actually means, having saved the sender's time at everyone else's expense.
Push back when it matters
The panel was equally wary of the reflex to condemn anything touched by AI.
Field's advice was to judge the output and the outcome, rather than the method. If a stakeholder has used AI and produced something perfectly serviceable, there may be no reason to object. The better question is why they went to AI in the first place.
The comms team might have been unavailable, or the stakeholder might not fully understand what the team does. The relationship might be weak, or past requests might have run into too many barriers.
Find that out and the problem becomes much easier to solve.
Holtz put the threshold slightly differently. If AI-generated communication has caused a real problem, bring the evidence: a bad decision, incorrect information passed along, confusion that can be traced straight back to the message.
"Keep the receipts," as Hubbard put it.
That is a much stronger basis for introducing governance and training than simply saying the content feels artificial.
Where the human line still matters
The discussion became more definite when the panel turned to AI-generated video.
Field drew a firm line around simulated executives. If video is being used to build trust, replacing a real leader with an artificial version risks weakening the very relationship the video is meant to support.
Her view was different for graphics-led or infographic-style videos, where no individual is being represented artificially.
Holtz offered one exception to his own caution about cloned leaders: translating a CEO's real remarks into other languages for international employees, provided the use of AI is disclosed.
He also described an animated management series produced with AI tools. The technology creates the characters, voices and visuals, but Holtz writes the scripts and directs the work. Without AI, he said, the project would probably never have been possible on the available budget.
What came through across the discussion was that AI can take on more of the research, editing, testing and production, but it can't take on the relationship between a leader and the people being asked to listen.
That stays with the person doing the communicating.