The Real Role of an Internal Communications Leader in an AI Transformation - and Why the Version Most Companies Are Using Is Ten Years Out of Date

Here is how most enterprise AI programs actually use their internal communications leader.

Someone in the C-suite decides the company is doing AI. Strategy is set. Tools are selected. Rollout timeline is determined. Budget is approved. And then, somewhere in the final weeks before launch, someone says "we should probably get comms involved" and the internal communications leader is handed a brief that goes something like this: employees need to understand what's happening, why it's good for them, and how to get started. Can you put together a campaign?

And the communications leader, who is good at their job and knows how to execute, builds the campaign. Town hall. Launch email. Intranet hub. Manager talking points. Training announcement. Maybe a video from the CEO. The calendar is populated, the content is polished, and on launch day the metrics look fine. Open rates are respectable. Attendance was solid. The dashboard shows awareness has landed.

Then eighteen months go by. And the program isn't working.

Not because the communication was bad. The communication was actually quite good.

It's not working because communications was never the problem and the company gave the communications leader a communications job when what the program actually needed was something entirely different.

This is the version of the IC role that is about ten years out of date. And it's the version most companies are still running.

where are you when the storm hits?

What the Outdated Version Looks Like

The outdated IC role in an AI transformation has three defining characteristics.

• The first is late entry.

  The communications leader joins the program after the strategy is set, after the tools are chosen, after the decisions that most directly affect employees have already been made. Their job is to announce a reality that was built without them, and to make that reality sound reasonable to people who had no input into it. This is the organizational equivalent of asking someone to do public relations for a house fire.

  
  

• The second is message ownership without decision ownership.

  The outdated IC role owns how the strategy is communicated but has no authority over the strategy itself. If leadership has decided that AI means certain roles will be restructured and nobody is willing to say that clearly, the communications leader is expected to write around it — to find language that is technically accurate but comfortably vague. They become responsible for the credibility gap between what leadership is willing to say and what employees actually need to know.

  
  

• The third is activity as output.

  The outdated IC role is measured on what it produces — the number of communications sent, the completion rate on training, the town hall attendance figure, the open rate on the launch email. In Post 4, we went deep on why those metrics are measuring the wrong thing. But the reason they became the metrics in the first place is that the outdated IC role was defined as a production function. Content in, impressions out. That's a useful role. It is not the role an AI transformation needs.

Gallagher's State of the Sector report shows that communicators are already taking on the role of change managers, yet only 2% have 'change' or 'transformation' in their job title. Almost half don't have a dedicated department budget.

Why This Version Fails Specifically for AI

The outdated IC model fails on most large change programs. But AI transformation breaks it in specific ways.

AI is not a bounded change.

Most major enterprise programs - an ERP implementation, a restructuring, an office consolidation - have a defined end state. There is a launch date, a go-live moment, and eventually a point where the program becomes the new normal.

AI doesn't work that way. The tools are changing. The capabilities are expanding. What was a pilot six months ago is being replaced by something more capable today. There is no finish line, which means a communications strategy built around launch cadences and milestone announcements doesn't map to the shape of the transformation it's supposed to support.

AI is emotionally loaded in a way most enterprise change is not. Employees are not anxious about a new expense reporting system in the way they are anxious about AI. The fear - of job loss, of being left behind, of being surveilled or replaced - is genuine, persistent, and not addressed by a reassuring email from the CHRO. Nearly half of employees worry about job losses due to AI, and only 41% say their employer has clearly communicated responsible AI use.

The IC leader who is handed a campaign brief cannot resolve that anxiety. The IC leader who is in the room when decisions about AI and workforce impact are being made and who has influenced those decisions can.

AI changes the credibility dynamics of all communication. Edelman's Trust Barometer found that 68% of employees already believe their leaders are withholding or misrepresenting information. Running leader communications through an AI tool and publishing the output does not help that number - it may make it worse. In an environment where employee trust is already fragile and the topic is one that directly threatens people's sense of job security, communications that feel polished, managed, or generated compound the problem rather than solving it.

The IC leader's job in this environment is not to make leadership sound good. It's to make leadership sound true and that requires being in the room where the truth is being decided, not in the communications room where it's being packaged.

What the Updated Role Actually Looks Like

I want to be specific here, because "strategic communications partner" is a phrase that has been written in IC job descriptions for a decade without actually changing anyone's calendar or budget. The updated role isn't a title change. It's a functional repositioning that shows up in concrete, observable ways.

• Entry point moves earlier. 

The updated IC leader is involved in the design of the AI strategy, not the announcement of it. That means sitting in the conversations where decisions are being made about which roles will be affected, what the transition timeline looks like, what support will be available, and what leadership is and isn't willing to commit to. The single most valuable thing an experienced IC leader brings to a strategy conversation is the ability to hear what's being decided and immediately understand how it will land with the workforce. That judgment is worth nothing if it arrives after the decisions have been made.

• Listening infrastructure replaces broadcast infrastructure. 

The outdated IC model is fundamentally about outbound, what leadership needs to say to employees. The updated model is equally about inbound - what employees are actually thinking, feeling, and doing, and what that tells leadership about how the program is landing.

It is a structured, continuous, segmented read on employee sentiment, specifically around AI, built to surface the signals that predict adoption failure before it becomes visible in the metrics. Fear-to-curiosity ratio by role. Trust in stated leadership intentions. Confidence in using AI for specific job tasks. These are the diagnostic reads that allow a program to course-correct in real time rather than discover the problem eighteen months later.

Building this infrastructure and interpreting what it surfaces is an IC function - but it requires a fundamentally different definition of what IC is there to do.

• Change architecture, not content architecture. 

The outdated IC role designs a content calendar. The updated IC role designs the change architecture - the sequence, pacing, and structure of the interventions that move people through the stages of change. What does the Awareness stage actually require, and how do we know when it's complete? What is the specific barrier at the Desire stage for frontline employees vs. managers vs. senior individual contributors, and what does the right intervention look like for each? What does Reinforcement mean when the change has no defined end state? These are change management questions. The IC leader who can answer them - and who understands both the organizational change theory and the communication mechanics - is the person this program needs at the table. Not after decisions are made, but while they're being made.

• Manager enablement is a primary deliverable, not a secondary one. 

Internal comms has always run on a secret subsidy: managers willing to take messaging and make it make sense — adding context, translating it for their function, and using their credibility to sell decisions they didn't make. That subsidy is running out. Managers report increased workloads, often absorbing eliminated roles. The implication for AI transformation specifically is significant. If the manager layer is overwhelmed, undertrained, and not bought in to the AI program themselves, no amount of company-wide communications will reach employees in the way that matters. The updated IC role treats manager enablement as a primary strategic deliverable - not a cascade of talking points, but a genuine capability-building program that gives managers what they need to have real, honest conversations with their teams about AI and their specific roles.

• Measurement accountability shifts. 

The updated IC leader is not measured on what they produced. They are measured on behavioral indicators that tell leadership whether the program is producing the outcomes it was designed to produce. We covered the full measurement framework in an earlier post. The point here is that accepting accountability for those outcomes — not just for the content that was designed to drive them is what repositions the IC function from a service to a strategic investment.

The Argument I Hear Most Often And Why It's Backwards

When I describe this repositioned role to communications leaders, the most common response I hear is some version of: "I'd love to operate this way, but leadership doesn't give me a seat at that table. I don't have the authority. I get called in too late."

I understand why this feels true. And in many organizations it is true, structurally, right now.

But here's what I've come to believe: the authority is not given. It's demonstrated.

The IC roles most at risk are built on execution alone. The ones that will thrive are rooted in counsel, judgment, and the ability to hold up a mirror to leadership - even when the reflection is unflattering.

The IC leaders I've seen successfully expand their role into transformation strategy didn't wait to be invited to the planning meeting. They showed up to the business problem. They built the listening infrastructure and brought leadership data they didn't have. They named the adoption risk before it became a crisis. They translated what employees were actually experiencing into language that leadership could act on. They connected the communication strategy to the business outcome in terms leadership could hold accountable.

That is not a communications function. That is an organizational effectiveness function.

"Communications" signals a simple service function, you have a message, we deliver it. The real opportunity comes with a complete identity shift. Forget "internal communications," try "organizational effectiveness." Drop "change communications" for "transformation enablement."

I'll go further than that: in an AI transformation, the IC leader who positions themselves as the person responsible for whether the behavioral change actually happens - not whether the message got out - is the most valuable person in the room. And they're almost never in the room right now.

What This Means for the CHRO Reading This

If you're a CHRO or CPO with an AI program that isn't producing the behavioral change you expected, there are three questions worth asking about how you're using your IC function.

1. When did they join the conversation? If the answer is after the strategy was set and the tools were selected, you brought them in too late. The most valuable thing an experienced communications leader offers is not the ability to package decisions, it's the ability to see how decisions will land before they're final and to influence them accordingly.

   
   

2. What are you holding them accountable for? If the answer is campaign output - communications sent, training completions, town hall attendance - you have defined the job as production. A production function produces. It does not transform. If you want transformation, the accountability structure has to reflect it.

   
   

3. What is their relationship to the listening infrastructure? If your IC leader is producing outbound content but has no structured, continuous read on what employees are actually experiencing with the AI program - and no authority to bring that data into leadership conversations and push back on decisions based on what it surfaces - you have a postal service. Reliable and necessary. But nobody invites the mail carrier to the strategy meeting.

The Version That Actually Works

The IC role that an AI transformation needs looks like this:

• present when strategy is being designed, not when it's being announced

• responsible for listening as much as speaking

• accountable for behavioral outcomes, not content output

• the person in the room who names the adoption risk before it becomes a crisis - and who has the relationships and the data to back up the call

That role requires a different skill set than the one most IC leaders were hired to bring. It requires genuine change management fluency - not a workshop certification, but the ability to apply behavioral change theory to a specific organizational context and design interventions accordingly.

It requires analytical capability: the ability to build and interpret sentiment data, to see the patterns that predict failure before they show up in the metrics. It requires a kind of courage that is harder than writing: the willingness to tell leadership what they don't want to hear, backed by evidence, and to hold the line on it.

That version of the IC role doesn't wait to be called in. It has already been in the room, shaping the decisions that will determine whether the program works.

Most organizations aren't running that version. They're running the version that writes the launch email and builds the intranet hub and reports back on open rates.

That version won't save an AI program that's failing. And as the programs keep failing — as the boards keep asking where the ROI is and the dashboards keep showing green while the behavior stays flat — the organizations that figure this out first will have a significant advantage over the ones that don't.

The IC leader who already understands this, and who can walk into a CHRO's office and name it in those terms, is not a communications resource. They are a transformation asset.

That's a very different conversation to be in.

Next - Check out Post 6 of 6 → What to Do in the Next 90 Days: A Practical Reset for AI Programs That Are Measuring the Wrong Things and Managing the Wrong Way