Communications After AI: Why Operating Model Matters More Than Tools
Estimated reading time: 11 minutes
Boston Consulting Group’s January 2026 study, GenAI Adoption in Corporate Affairs and Communications, matters to SMEs as well, even though its cohort sits mainly in larger organisations. Conducted with 200+ senior communications leaders across Fortune 1000, Forbes Global 2000, and America’s largest private companies, and led by Russell Dubner with Professor Craig Carroll, it asks a serious question: how prepared is the communications function to lead AI change rather than merely react to it?
I am not interested in repeating its findings for their own sake. I am interested in what they reveal. My reading is that the study matters not because it proves communications leaders are late to AI, but because it exposes a deeper problem: many teams are still discussing AI at the level of tools, while the real pressure is falling on the communications operating model itself. That has obvious consequences far beyond large corporate functions. It should matter to Irish SMEs as well, especially those moving from small to medium scale.
The BCG signal
The headline number is stark: 88% of communications leaders say they are not fully prepared to lead an AI transformation in their function. Only 31% report meaningful progress scaling GenAI beyond pilots, while 68% identify as lagging. Those are not the numbers of a profession in denial. They are the numbers of a profession caught between awareness and redesign.
The most useful finding sits underneath the headline. BCG identifies operating-model design as the number one barrier to AI transformation, cited by 35% of respondents. Not prompts. Not access to tools. Not even budget in the first instance. The barrier is structural: how to define the roadmap, redesign workflows, assign ownership, preserve accountability, and embed AI systematically enough to produce business outcomes rather than scattered efficiencies.
That is why the report deserves attention. It shifts the discussion from enthusiasm to architecture.
More than marketing
The report makes another point that should not be lost, especially in SME contexts where communications is still too often collapsed into marketing.
BCG places Corporate Affairs and Communications second among enterprise functions on GenAI transformation upside, behind customer service and ahead of HR, marketing, finance, IT and legal. That alone should force a pause. If communications sits that high in transformation potential, then the function cannot be reduced to content throughput or campaign support.
The accompanying examples make that clear. The upside is not limited to drafting faster. BCG points to areas such as insight and foresight, reputation and advisory work, process change through agentic AI, and innovation beyond optimisation. In other words, the function under pressure here is not simply the content factory. It is the wider system by which a business interprets context, aligns internal understanding, advises leadership, protects trust, and coordinates what can and cannot be said.
“AI does not create the need for communications strategy. It raises the price of not having one.”
That distinction matters in Ireland. In many SMEs, the two functions still overlap heavily in practice, especially where one person or a very small team carries multiple responsibilities. But they are not the same function. Marketing is typically organised around market visibility, demand generation, customer acquisition, campaigns and commercial conversion. Business or corporate communications operates more broadly across governance, compliance, internal alignment, stakeholder relationships, process discipline, knowledge flows, approvals, and the business’s ability to speak consistently across different contexts. Under AI conditions, that distinction becomes more consequential, not less. The issue is not that marketing becomes unimportant, but that communications as a wider business system can no longer be treated as if campaign logic were enough to govern it.
Adoption or transformation?
The study also helps clarify a distinction that I find increasingly important: adoption is not the same as transformation.
A team can adopt AI and still leave the core of the function unchanged. It can use GenAI for first drafts, faster summaries, better research support, routine analysis, and more efficient production. Those gains are real. There is no need to dismiss them. BCG itself shows that many leaders are still oriented toward near-term improvements: 54% prioritise capacity gains, while only 22% prioritise future strategic capabilities.
That is understandable. Quick wins are easier to justify. They are measurable, immediate, and politically safer.
But they are not yet transformation.
Transformation begins when the function asks a more difficult set of questions. Which workflows actually need redesign? Where must human judgement remain decisive? Which approvals are still fit for purpose? What becomes centralised, and what becomes distributed? How is message consistency maintained when more systems participate in drafting, analysis, recommendation, and simulation? How is value measured beyond time saved?
This is where I think many AI discussions remain too shallow. They assume that accelerating existing activity is the same thing as changing the operating model. It is not.
If communications is a vehicle, AI is not simply a stronger horse. Nor is it enough to bolt an engine onto a carriage and call that transformation. Once propulsion changes, the whole design has to be reconsidered: control, balance, braking, coordination, safety. Core elements remain, but the logic binding them together changes.
The same is true here. AI can enter communications as an add-on. It becomes transformative only when workflows, governance, ownership and measurement are redesigned around it.
Operating model
This is the part of the report I find most valuable, because it forces the profession into a more serious vocabulary.
“Operating model” can sound like executive abstraction. In communications, it is not abstract at all. It means practical things: who drafts, who reviews, who escalates, who approves, which outputs are routine, which are sensitive, where the internal line is defined, how evidence is stored, how exceptions are handled, and who owns the final message when a business risk is attached to it.
That becomes even sharper when we remember another BCG number: 74% believe in the technology and its payoff, but only 7% say they are very confident in their ability to measure and communicate its value in a decision-grade way. That gap is telling. It suggests that the function is not struggling primarily with conviction. It is struggling with design, proof, and control.
For growing SMEs, the same logic appears in compressed form. The business becomes more visible. More people write externally. More proposals circulate. More content is generated. More tools are introduced. More decisions are distributed. Yet the communication system underneath remains informal, founder-dependent, or fragmented across functions.
Under those conditions, AI does not create order. It intensifies whatever order or disorder already exists.
That is why I would put the issue bluntly: AI does not create the need for communications strategy. It raises the price of not having one.
Agentic illusion
The report’s forward-looking sections on agentic AI are important. 71% of respondents believe agentic AI will create impact within twelve months. BCG is also right to connect agentic systems with governance, decision rights and accountability. The more autonomous the system becomes, the less tenable it is to leave those questions vague.
This is where I want to extend the discussion beyond the report.
I think communications is now approaching a repeatable managerial mistake we have already seen elsewhere: the temptation to transfer contextual expertise out of human hands too quickly because the newest tools appear capable enough.
We saw versions of this in technical functions over the last year, especially around coding. Some firms moved from excitement to over-correction. They assumed they could cut external expertise, thin internal capability, and replace too much judgment with increasingly agentic workflows. In a number of cases, those assumptions proved expensive, and some of the burned bridges then had to be rebuilt.
Communications is not exempt from the same risk.
If anything, the risk may be greater. Communications depends heavily on context, sequence, organisational memory, audience sensitivity, political judgement, issue anticipation, and the ability to hold multiple stakeholder realities together at once. These are precisely the conditions in which fully unsupervised autonomy becomes brittle.
“I do expect communications teams to become mixed human-agent systems.”
I do expect communications teams to become mixed human-agent systems. That is probably the normal model for the coming years. But that is not the same as believing in one universal agent that can absorb the whole function.
BCG’s own material points in a different direction. Its illustrations move from narrower task agents to more complex orchestration and campaign agents, which already implies layered architecture rather than one monolithic intelligence. That seems right to me. A robust AI-enabled communications model is likely to require multiple layers: narrow agents for repeatable tasks, broader agents for synthesis or orchestration, and human professionals for judgement, escalation, interpretation, and accountability.
The more strategic the communications task, the less credible the fantasy of full replacement becomes.
Strategic horizon
Another revealing weakness in the report is horizon.
Most leaders are still pursuing AI through the logic of immediate optimisation: faster output, lower cost, more capacity, better short-term efficiency. There is nothing wrong with that. But it is not enough to define the future of the function.
The more interesting BCG slides sit elsewhere: insight and foresight, leadership advisory, audience simulation, synthetic testing, innovation, and new problem discovery beyond optimisation. That is where communications begins to move from production support toward strategic infrastructure.
This matters because AI is not only changing how work is done. It is changing what kind of work becomes more valuable.
Routine production will continue to be compressed. Basic drafting, formatting, summarisation and standard analysis will become cheaper and faster. That is obvious. The more consequential question is what rises in value as that layer is compressed.
My answer is: judgement, interpretation, governance, internal alignment, stakeholder reading, issue framing, escalation logic, and the capacity to redesign the system itself.
That is the real difference between adoption and transformation. Adoption improves the current task mix. Transformation changes the hierarchy of value inside the function.
The SME implication
The BCG cohort is not a sample of Irish SMEs, and it should not be misused as one. But that does not make the report irrelevant to smaller firms. In some ways, it makes it more useful.
If large, well-resourced communications functions are struggling with operating-model redesign, then smaller businesses should not assume they can improvise their way through AI change merely because their structures are lighter.
For SMEs on the border between small and medium scale, the problem is often sharper, not softer. They have enough complexity to create communication risk, but not enough structure to manage it consistently. Communications responsibilities are split between founders, marketers, operational staff and occasional outside support. Internal communications is underdeveloped. Approval logic is vague. Message ownership is implicit rather than explicit.
Under those conditions, AI can easily increase visible productivity while weakening coherence.
“Treat communications as part of business infrastructure before you try to scale its outputs.”
That is why I think the practical implication for Irish SMEs is not “use AI later” or “be cautious first”. It is more exacting than that. Treat communications as part of business infrastructure before you try to scale its outputs. Know where AI already touches stakeholder-facing work. Define where human review remains non-negotiable. Clarify what staff can say externally. Rebuild at least one important workflow before adding another layer of tooling.
If that sounds slower than the current market mood, so be it. In communications, false acceleration has its own costs.
What follows
The most interesting question raised by the BCG study is not whether communications leaders are behind on AI. It is whether the profession is ready to accept what AI is forcing into view.
If communications really is one of the highest-upside functions for transformation, then the conversation cannot stay trapped at the level of tools, prompts and quick wins. It has to move toward operating model, governance, redesign, and the changing boundary between human judgement and machine capability.
That boundary is where the next serious debate begins.
If teams cut agencies before they redesign the function, what exactly are they saving - cost, or merely the visible form of expertise? If internal communications becomes a mixed human-agent system, who owns truth, escalation and exceptions? If leaders remain focused on productivity tables while ignoring message governance, what kind of trust architecture are they actually building? And if communications is still treated as a subset of marketing in many SMEs, what happens when AI starts stressing the parts of the business that marketing never governed in the first place?
“Strategic advisory support for the communications function during periods of change could provide answers.”
Those are not rhetorical flourishes. They are practical questions about how businesses will decide, explain, coordinate and remain credible under AI conditions. For many SMEs, one of the most logical responses is strategic advisory support for the communications function during periods of change. An external communications strategist can help tailor the system to the business’s goals, constraints and internal talent, leaving behind not only a more workable vehicle, but also the trained key people needed to drive it.
That is why I do not read this report as a story about late adoption. I read it as a warning that the communications function is being pushed toward a more demanding form of maturity.
Not bigger tool stacks. Better system design.
That conversation is only beginning.
About the author
Jurij Blazejewski is an independent communications strategist working with Irish SMEs on structured communications systems, governance and AI-related organisational change.