You're asking the wrong questions on AI, pt. 1
That might not be the fun thing to hear, but it is true. Let's start with what's actually happening, because the conversation around AI (in communications and beyond) has become almost wilfully confused. And I don't think that's accidental. Confusion is comfortable. It lets you keep doing what you were already doing while telling yourself you're adapting. We prefer just keeping on going on. It's fine. There's a term for that: Denialism.
But here is the reality. Our industry has looked at AI and asked: how do we use this to make more things, faster? That's it. That's the sum of the strategic ambition. More content, more campaigns, more output, more "presence." Scale the machine. Optimise the funnel. Ship more stuff. And I understand why. It's a reasonable answer to the wrong question.
Because the question was never how do we produce more. The question (the one that actually matters) is what are we even trying to do here? What does it mean to communicate well? What does it mean to be understood? What does it mean to build something with your audience rather than just at them?
Nobody wants to sit with those questions, because they're hard and they don't fit in a quarterly report. So instead we chase efficiency and call it progress.
But here's the kicker
We have been here before. This is important to understand, because there's a tendency to treat AI as though it's categorically new, as though nothing that came before it offers any useful guidance. That's wrong. Not just wrong, it's a convenient amnesia that stops us learning from what we already know.
Look at finance. Trading has been machine-dominated for decades. Not years, no, decades. The automation that people are now panicking about has been standard practice in financial markets since before most of us had smartphones. Execution is handled by algorithms at a speed and scale that no human can touch. And yet. The industry didn't shrink. It didn't collapse into a handful of servers and engineers. It grew. It reorganised. It expanded massively around the things that couldn't be automated: judgment, relationships, trust, the long-term management of perception and risk. Even more so: It evolved, to include more and more of the intangible world. Think the young domain of Digital Assets, for example.
The locus of value moved. It didn't disappear. What survived, thrived even, was the human layer. The relationship managers. The advisors. The people who understood that their clients weren't just looking for execution; they were looking for someone who understood their world, their risks, their ambitions. Someone they could trust when things got complicated. That has not changed. And will not change.
Communications is now facing the same inflection point. And it is responding, at scale, by doubling down on the part that is easiest to automate. Which is, to put it plainly, not great.
There's an assumption underneath all of this that I think deserves to be dragged into the light. The assumption that volume is a proxy for impact. That more content equals more influence. That if you just produce enough of it, something will land.Content isn't king. Context is.
But: The marginal value of another piece of content is approaching zero. Not because people don't read, or don't care — but because they are absolutely saturated. The supply of content has exploded and continues to explode. Distribution is fragmented, algorithmic, and largely outside your control. The traditional metrics - placements, impressions, engagement rates - have become detached from anything resembling real influence. You can be everywhere and matter to no one.
In that environment, efficiency doesn't create advantage. It compounds irrelevance. If you produce mediocre things faster, you have not improved your situation. You have automated your own obsolescence. The deeper issue is not technological. It never was. It's conceptual.
Over the last decade, communications has slowly operationalised itself into a production function. Optimised for output. Measured on activity. Justified through metrics that are, in many cases, closer to performance art than genuine insight. The work became about keeping the machine running — content calendars filled, channels maintained, announcements shipped, stakeholders serviced. Intelligence existed, but it was peripheral. A report here, some monitoring there, occasional insights that were acknowledged and then quietly set aside.
AI fits neatly into that model because it scales the existing logic. It makes the machine faster. But it doesn't question whether the machine is pointed in the right direction. And that question, the one about direction is the only one that actually matters.
So, then what?
Here's what changes when anyone can produce competent content: competence stops being differentiating. When language models can replicate tone, structure, and surface-level insight at scale, the baseline rises and the ceiling lowers simultaneously. What remains scarce - genuinely, stubbornly scarce - is not the ability to say things well. It's the ability to say something that matters, in a way that changes how it's received.
That's not a production problem. That's a strategic one. And strategy, in communications, is inseparable from understanding your environment. Most organisations still treat their communication landscape like a set of channels to be managed. Owned. Earned. Paid. Plans are built around these categories. Success is measured within them - and ideally operated in siloed teams that don't communicate well internally. Everyone is happy to do their own thing on their own island (pretending there is no sea of context connecting them). Now, this is even more disconnected from how information actually moves.
What actually exists is a continuous, uncontrollable surface of visibility where everything interacts: media, creators, employees, customers, algorithms, communities, conversations you never see. A statement doesn't "live" in a press release. It propagates. It mutates. It gets reframed, indexed, surfaced in contexts the originator never imagined. The idea that you can place a message and then manage it from there. That idea is not just outdated, it's a liability.
The buck does not stop at earned. The buck does not stop at owned. The buck does not even stop at paid. There is so much more to it.Control was always an illusion. It is now an expensive one
What replaces the illusion of control is something more honest and more demanding: orientation. Knowing who shapes perception in your category. Understanding where interpretation actually happens. Seeing how narratives form — not just tracking where your messages land. This is where relationships stop being a soft, secondary concern and become a hard strategic asset. Not relationships as access. Relationships as shared context.
A journalist who understands how you think will interpret you differently than one who only receives your press releases. A client who has experienced genuine value from you - beyond the deliverable, beyond the product - will defend you differently when things go wrong. An employee who actually understands the story of the company will represent it differently in the spaces you'll never see. These are not tactics. They're structures of meaning. And no, you cannot automate them.
Let me be direct about where this leads, because I think the risk is real and I don't think enough people are saying it clearly. If AI is used primarily to increase throughput (which is the dominant use case right now) two things happen in parallel:
First, organisations start to train themselves, and their audiences, to accept lower standards of substance. When speed is rewarded internally, depth becomes optional. Not explicitly, it just quietly stops being prioritised. Over time, this hollows out the capability that would actually differentiate you. You get faster at producing things that matter less. Until you don't matter at all.Second, the external environment fills with homogeneous language. The patterns converge. The tone flattens. Distinction disappears. And in a landscape where everything sounds the same, being good at this, like genuinely good, becomes indistinguishable from being invisible.The paradox is clean and brutal: the more efficient communication becomes, the more expensive attention becomes. Not in media terms. In cognitive terms. You are asking people to give you their time. Time is the one resource that doesn't scale, doesn't replenish, and can't be bought back. If what you're offering in return is interchangeable, they won't.
Communication is not the distribution of information. It is an exchange of value
We don't ask ourselves this question often enough:
"What do people get in return for the time they give your content?"
If the answer is awareness, you're already behind. If it's information, you're competing with infinite supply. If it's "content", god help you. True value in this context is experiential. It's clarity where there's confusion. Perspective where there's noise. Meaning where there's fragmentation. Sometimes, if you're doing it well, even something as rare as genuine enjoyment. Not polished. Not on-brand. Actually enjoyable.
This is why the companies that outperform over time are not the ones that communicate the most. They're the ones that accumulate narrative advantage: slowly, deliberately, through consistency of perspective and coherence of story. Apple is the example everyone reaches for and for good reason. Fifty years of deliberate and consistent storytelling (‼) have created a position that is, in practical terms, unassailable. Their AI assistant might make you wince. Doesn't matter. The myth holds. You still want the product for reasons that have little to do with its specifications, and everything to do with what owning it gives you.
That is not a campaign. That's the compounded return on decades of caring about how you are understood. Most organisations will not get there and that's fine. But the ones that move in this direction - that start asking better questions, that invest in understanding over output - will not just communicate better. They will be understood differently. And that is the only advantage that truly compounds.
Used properly, AI can be part of this. Not as a content engine on steroids, no. First, it is an environmental sensor. Not to replace thinking, but to expand the field in which thinking operates. To map conversations and networks faster than a human team could alone. To surface shifts in how your category is being talked about before they become crises. To test interpretations, explore angles, stress-test assumptions. To increase quality, not just speed.
So my honest advice: Stop treating communication as output. Start treating it as an accumulation of meaning. Everything else follows from that.