Much like the embers of a campfire, a danger sits underneath the surface of an organisation’s daily life. Embers that are never fully extinguished, giving off enough heat to dry out everything around them. Without much imagination, the wind changes, a single spark lands in the wrong place, and the risk everyone thought was handled announces itself in a sudden line of flames.
Most organisations’ communication decision matrix looks exactly like that campfire: it is stable on the surface while a decision debt builds underneath. The architecture of tools, channels, and habits that “just arrived” over the years carries risks that remain invisible as long as things mostly work until a moment of stress or the arrival of AI forces every hidden compromise into the open at once.
This is how decision debt works. A series of small choices not made. A conversation that never happened about what, exactly, the system is supposed to do. A governance question deferred until next quarter. A meeting that kept getting pushed because the urgent always crowded out the important. Each one of these non-decisions is borrowed time. The interest compounds in the architecture of the organisation, until a crisis brings it all to reality at once.
Most organisations communications are not struggling because their strategy is wrong, or not only. They are failing because communication decision rights were never designed. Who owns the company’s story? Who can commit on behalf of the organization? Which channels count as the official record, and which are just noise? These questions are at the intersection of corporate communications, IT, legal, HR, and business operations. Because they belong to everyone, they are designed by no one.
The result is a familiar landscape. A CRM that was a good decision at the time. Email that has always been there. Teams or Slack because it was easy to say yes. A couple of WhatsApp groups because a key client wanted it that way. A document system, an intranet, a shared inbox that two people manage but belongs to neither of them. A corporate communications function somewhere in the middle, expected to sustain reputation, manage risk, and support growth across a stack that nobody ever treated as an architecture.
Every piece answered a specific problem at a specific moment. Taken together, they form a communication system that was never designed, in which the story of the organisation is fragmented across a dozen surfaces, held partially in each one, fully in none. That fragmentation is the first layer of decision debt.
Then, as is most often the case, AI arrives in the same pattern as everything else. An AI assistant in Outlook. An automatic meeting summariser. A chatbot on the website. A summarisation tool over internal documents. Each one feels like progress. Each one answers a real pain point for a specific team. And each of them increases the interest rate on existing debt.
The AI tools learn from the surfaces they can see. Some email threads. Some CRM fields are updated sporadically. Some meeting summaries capture the slide deck, but not the moment where a client pushed back and a commitment was made. The AI builds its model of the organisation from this. It learns who the stakeholders are, what priorities appear to be, and which issues look urgent. It learns the corporate story as told by the most legible channels. Then it acts on that partial story with confidence.
The senior account director commits during a Teams call. It never becomes part of the CRM. It does not make it into the official contact report. It lives in memory and in a private notebook. When the AI assistant is asked to draft a client update, it draws only on the systems that look authoritative from the outside. It produces a fluent, professional, completely coherent message that is wrong.
From the client side, the contradiction is simple. One message says one thing. Another message says something else. By the time anyone within the company notices, the conversation has moved from the quality of the relationship to the company’s reliability. The issue now becomes about whether the organization can be trusted when it speaks.
The postmortem revealed that the organization has no authoritative account of its own commitments. No single narrative spine connects what was agreed to, what was approved, and what was communicated. Corporate communications has messaging. Legal has clauses. Sales has decks. Operations has tickets. AI did not create the gap. It just made the consequences show up faster and more publicly.
This is what decision debt looks like in a communications context. Every meeting ends with an informal “we will figure it out later.” Every campaign is launched without clarity about who has the final say. Every approval chain that relies on personal relationships rather than defined authority. Each one is borrowed time. The cost compounds within the underlying architecture until a moment of stress makes it all real at once.
Most of us have been there in one way or another, but AI changes the economics of that debt. It removes the buffer that once existed between internal ambiguity and external perception. When communication output can be generated at scale and speed, the tolerance for architectural ambiguity collapses. You cannot bolt generative capacity onto incoherent governance and expect reputation resilience to come out the other side.
There is a way to frame this that comes from narrative theory. An intradiegetic story is told from inside the world in which it takes place. The narrator is a character, subject to the same constraints and stakes as everyone else. That is a useful way to think about what a modern corporate communications department is being asked to become.
If you build a communications function and the AI within it that operates intradiegetically, you are building a system that understands itself to be part of the organisation’s story. Inside the actual narrative of commitments, relationships, and context. Not sitting beside it, scraping the surfaces of CRM and email, and treating that as reality. The difference matters.
An AI that only reads archives and dashboards is always working on a representation of the company’s story. It sees what has been captured, formalised and filed. It does not see what was implied in the room, negotiated at the edge, and revised at the last minute. It fills gaps with plausible inference, which is another way of saying that it invents where the record is thin.
An AI that operates inside a deliberately designed communication architecture is working with the story itself. That architecture is what corporate communications should be claiming as its core mandate. It is the decision to define where the official story lives. Which channels carry authority.
How commitments are captured. How informal intelligence becomes part of the record. How ambiguity is surfaced rather than smoothed over.
In practice, this means three decisions that most organisations still treat as operational detail instead of structural design.
The first is deciding what counts as the official communicative record. When a senior leader says something in a town hall, where does that live twenty-four hours later? When a key client conversation changes scope, how does that move from a chat transcript to the systems that everyone, including AI, recognizes as authoritative? Without an answer, every tool is working from its own partial canon.
The second is deciding where human authorship is non-negotiable. In a world where AI can draft almost anything, the question is not whether it can. It is where it must not. There are moments where corporate communications must be unmistakably human. The apology that carries legal and moral weight. The commitment will be treated as contractual. The statement that will define how the organisation is understood in a crisis. If those boundaries are not designed, AI will eventually speak with institutional authority at exactly the wrong moment.
The third is deciding how the system handles what it does not know. A communication architecture that silently encourages AI to gloss over uncertainty with smooth prose optimizes for fluency over truth. Instead, the architecture needs explicit rules for how unknowns are flagged, how draft status is signaled, and how human review is triggered when a message touches regulated topics, sensitive stakeholders, or ambiguous data.
None of these decisions are technically complex. They are organisationally uncomfortable. They require an honest conversation about how the company actually works, rather than how the org chart suggests it should work. They force executives to confront the fact that communication is not just messaging. It is the operating system that connects decisions, actions, and accountability.
For corporate communications leaders, this is the opening. The function that has often been seen as a downstream service provider is suddenly facing the question that shapes everything else: where does the story of this organization live, who authors it, and how do human and artificial intelligence operate within it rather than beside it? Answering that question is not about drafting an AI policy and filing it on the intranet. It is about asserting architectural authority over the company’s narrative infrastructure.
Some organisations will only do this after the crisis. They will rationalise channels in twelve weeks that could have been designed in ten years. They will retrofit governance around tools that are already live. They will rebuild trust with stakeholders who have already seen the seams. The debt will be paid on terms set by events, not by leadership.
The organisations that treat AI not as a tool to add but as a reason to decide will end up with something else. A communications architecture that can support intradiegetic AI. A department that is not just producing content, but curating and governing the story in which everyone else is already acting. An operating environment where decision rights are clear enough that when the next crisis comes, speed and coherence are not at war with each other.
The debt is already there. The choice is whether to keep adding interest through non-decisions or to treat corporate communications as the place where the story’s architecture is finally decided.
Leave a Reply
You must be logged in to post a comment.