The Two Clocks

The Stakes

A buyer asks an AI assistant which vendors to consider. The system returns a shortlist. That shortlist is two things at once: a commercial gate that decides whether the company makes the cut, and a reputational shorthand that shapes stakeholder opinion before the company has said a word.

Much of what produces that shortlist sits on surfaces Communications shapes. More than 95% of links cited by AI search engines are non-paid mentions and coverage, with 27% originating directly from earned media.^{1Gartner, Top Predictions to Inform 2026 Comms Strategies, February 2026.} The position gives Communications a new reputational mandate.

AI's reputational consequence is a board-level concern. 72% of S&P 500 companies disclosed at least one material AI risk in their 2025 annual filings, and reputational risk is the single most-cited category.^{2The Conference Board, AI Risk Disclosures in the S&P 500, October 2025.} Legal frameworks are forming across major economies — but deployment is outpacing regulation. In the gap between what AI systems are already doing and what the law can govern, reputational risk is compounding.

The governance response is forming in parallel. Frameworks are being built inside most large organisations. The Chief AI Officer (CAIO) role is emerging — with a mandate expanding rapidly from AI strategy and model governance into change management and organisational positioning. The review boards forming around them have a recognisable core: Legal, IT, Risk, Compliance, with HR increasingly added. Communications is rarely on the list.

That absence matters. Communications is the only function whose object is coherence itself — not the individual parts, but the total picture. Its unique horizontal position within the organisation makes that reading possible. It is also the function that understands legal compliance and reputational coherence are not the same discipline. Together, these make its contribution to AI governance non-substitutable.

At the same time, the same technology is automating the production work Communications has historically been measured on — a real risk for value propositions built on outputs. The readiness gap widens it: 88% of Corporate Affairs and Communications leaders say they are not fully prepared to lead an AI transformation in their function.^{3Boston Consulting Group, (Gen)AI Adoption in Corporate Affairs & Communications, March 2026.}

Communications functions that use AI to reclaim strategic bandwidth and redesign for AI-mediated reputation will solidify their position. The rest risk demotion to an executional function.

This framework is for leaders willing to act on what this shift requires — before someone else's logic defines it for them.

The Two Clocks

There are two clocks running simultaneously. Most Communications functions manage only the immediate one.

The immediate clock is outward-facing. It runs the operational demands that monitoring, content creation, and reputation management now carry in an AI-mediated world. Reputation is now mediated by synthetic intermediaries that compress an organisation into a paragraph and hand it to anyone who asks. Governance now extends to systems Communications has never had to engage with before. All of it is legitimate, all of it is urgent, and all of it will absorb every hour the function has — including the hours AI gives back by automating tasks.

The structural clock is inward-facing. It determines what Communications needs to become to meet the needs of the AI era — and it gets no such urgency. Nobody is asking the CCO to audit whether the function is being valued for what it produces or for what decisions it influences. Nobody is scheduling a board conversation about what Communications needs to become. That work has no external deadline. Nothing forces it onto the calendar. It will not happen unless someone makes it happen.

Both clocks, or neither works

A function that runs only the immediate clock ends up with new territory it cannot occupy. A function that runs only the structural clock ends up leaner and more efficient, optimised for a world that is changing around it.

The rational trap

Most Communications functions will prioritise the immediate clock over the structural one. It is a rational response to incentive structures. The immediate clock is where the visible risk lives, where the CEO is watching, and where the function's existing credibility is established. This response is also self-defeating. It is where mandate compression begins — quietly, through a series of decisions that each seem reasonable.

Even when the trap is named, the path might not change. The reason is not a lack of capability or bandwidth. It is that each clock requires the Communications leader to be professionally vulnerable in a different direction simultaneously.

The immediate clock requires the Communications leader to show up as a sophisticated technology and risk thinker, not just for people but for systems — a profile many have not yet fully inhabited publicly. The structural clock requires the Communications leader to dismantle parts of the function that currently define its identity — visibly, deliberately, under scrutiny.

Most will choose the vulnerability they can manage — outward-facing transformation — and quietly defer the one they cannot — internal redesign.

Protection over proportion

Communications is under pressure. Budgets are shrinking. Teams are anxious. The AI adoption window is narrowing. Running both clocks in equal measure is not an option for most functions, and pretending otherwise is unrealistic.

The principle is holding both clocks — they do not need to be split evenly. A function that protects ten per cent of its capacity for structural work is holding both — but only if that ten per cent is ring-fenced. Left to its own devices, it will be absorbed by the next crisis. The discipline is not the proportion. It is the protection.

The Two Clocks in Five Dimensions

The Two Clocks framework has five dimensions. The first three are inward-facing — the production identity question, the redesign it triggers, and the talent profile that sustains the function in its new shape. They are the structural clock's work. The last two are outward-facing — signal architecture, the coherence of the organisation's signal across the surfaces AI systems draw on; and the trust consequence, Communications' contribution to governing the reputational risk those systems create. They are the immediate clock's strategic agenda.

The AI landscape will keep evolving, and faster than any framework can track in its particulars. What follows is not a map of the terrain but a set of orientations for navigating it. The five dimensions are the questions Communications has to keep answering; how the function answers them — and the terrain it answers them on — will keep changing.

The path through the dimensions will look different depending on where you start — the structure of the function, its current mandate, and its position in the organisation's decision hierarchy.

From Production Identity to Strategic Intent

Communications carries an inherited production identity — a value proposition built on the volume and quality of what the function visibly produces. Not every function sits inside it equally, but the gravitational pull is real.

The function was never only production. Interpretation, risk sense-making, the judgement to pace leadership through a crisis rather than amplify it — these were always the function's real contribution. Their premium was simply impossible to quantify. And what cannot be measured cannot be claimed.

The CCO who prevented the crisis walked into the CEO's office with nothing to show. The KPI was zero. Demonstrating the lack of something had no natural metric. That is the measurement gap in its most precise form.

The sequence that followed is not complicated. The measurement gap came first — without a credible mechanism to demonstrate commercial impact, output volume remained the only visible proxy for value. Output-based identity followed from that. Mandate compression — the gradual narrowing of the function's scope and the authority it carries — is the risk that follows from both: not abstract, not distant, becoming concrete with AI.

AI is not a productivity layer for Communications. It is a structural stress test. By automating the outputs the function has been measured on, it exposes whether Communications is genuinely upstream — embedded in how decisions are shaped, risks are governed, and signals are constructed — or whether it is a sophisticated production operation that has mistaken output volume for strategic influence.

The first move is uncomfortable and internal. It is an honest review of how much of the function's current value proposition rests on the production identity AI is now exposing — and how much on the upstream capabilities that remain irreplaceable. Most functions will discover the balance is not where they thought it was.

Set the strategic intent

Setting strategic intent means answering a specific question: what this function exists to do in an AI-mediated world, and what it needs to look like to do it.

In practice, that includes revisiting Communications priorities, scoping what the function should be involved in and to what degree, and what it will stop doing to make room for the new mandate.

The temptation will be to skip this and move to redesign — it is more concrete, more actionable, and easier to show progress on. That is the failure mode to avoid. Redesign without intent becomes whatever the immediate clock permits: faster outputs, leaner production, a more efficient version of the existing function. The intent has to be set first, or the redesign optimises the wrong thing.

Redesigning From Intent, Not Pressure

The strategic intent defines the function's purpose. Redesign is what has to change operationally for it to deliver on that. But it only works if it is governed by a principle the immediate clock will quietly absorb if it is not held deliberately: reclaiming strategic capacity and redeploying it intentionally. Efficiency is the byproduct, not the primary goal.

Only 34% of organisations report using AI to deeply transform their business. The remaining two-thirds are using it at the surface level or somewhere in between^{4Deloitte AI Institute, State of AI in the Enterprise, 2026 edition, January 2026.} — applying AI to existing processes without redesigning the underlying work. The advice available to Communications leaders overwhelmingly reinforces the same direction. Faster outputs, better monitoring, and new tools layered onto existing processes.

What redesign looks like in practice

Audit against the intent

The audit measures how much of the function's current activity and resource base supports the strategic intent.

What that means: mapping resource allocation upstream versus downstream, identifying what the function is producing that no longer serves the intent, and establishing what can be automated and at what ratio of human-AI involvement.

The upstream shift

Focus and investment should be on building or acquiring capabilities to use AI strategically, both in what Communications is now capable of doing and in what it can be measured on.

A few examples illustrate what that looks like. They are not a ranking and not the ceiling — they show the kind of shift the strategic intent makes possible. Tools and methods will evolve over time. The purpose they serve — the strategic use of AI — does not.

Attribution. When an issue surfaces in AI-generated responses, a Communications intervention can now be tracked against whether AI citations shift in response. For AI systems pulling live from the web, the feedback loop runs in days. For what models hold from training data, the lag is longer and attribution noisier — but the direction of travel is unambiguous: the loop is closing.

Risk simulation. AI-powered simulation allows message testing and crisis scenario modelling against dynamic, stakeholder-specific environments — live exercises that reveal where messaging holds and where it does not, before the real-world test arrives.

Historical signal as a forward strategy. A decade of coverage, earnings calls, and executive commentary can now be processed to identify which narrative frames generated sustained traction, which voices cut through, and which messages are worth building on — turning historical signal into forward strategy rather than retrospective reporting.

Causal measurement. Methods that establish links between Communications activity and commercial outcomes are in early development and not yet applied meaningfully to Communications work — but the category is worth watching.

Where the resources go

Internally, that means a deliberate shift in where the function's own spending goes. Investment in AI tooling and infrastructure. Upskilling across the existing team, not just concentrated on a new technical hire.

Externally, that means partners and tools primarily scoped to incremental execution or friction removal carry less strategic weight. AI is increasingly capable of doing that work. The partners worth weighing more heavily are those who bring capabilities the function does not have and cannot build quickly enough internally — systems thinking, governance fluency, AI literacy applied to reputational questions, to name a few.

37% of Corporate Affairs and Communications leaders are considering reducing agency spend by more than 5%, but only 12% are redirecting that spend towards strategic advisory work.^{5Boston Consulting Group, (Gen)AI Adoption in Corporate Affairs & Communications, March 2026.} The money is moving. Whether it moves to strategic capability is a decision whose consequences compound.

Governance starts at home

Redesign is not credible if the function's own AI use is ungoverned. How Communications uses AI — what the team produces with it, what review processes apply before AI-assisted material goes out — is a precondition for helping the wider organisation craft the same standards.

The New Capability Base

Communications carries capabilities AI does not replicate. Judgement. Craft. Relationship fluency. Second-order risk thinking. But they are only as valuable as the decisions they inform. Applied downstream, they shrink. Applied upstream, they are worth more and far harder to displace.

The other risk is over-reliance on existing strengths. They need to be complemented with new ones for the function to assume the required role.

The incoming capabilities will include:

• AI literacy — not technical depth, but sufficient understanding to engage governance conversations and challenge system design decisions.
• Reputational risk architecture — mapping where AI systems create trust exposure across decisions, surfaces and agentic deployments, before deployment, rather than after failure.
• Coherence judgement — looking across everything the organisation says about itself in the surfaces AI systems draw on, and reading whether those parts are telling one story or several.

These are the capabilities visible at this stage of the function's evolution. Others will emerge as AI evolves.

Resist the pull towards the visible hire: the person who knows the tech

The function does not need AI experts thrown at it. It needs people who understand AI well enough to govern it, and Communications well enough to know what governing it actually means.

When social media arrived, Communications functions acquired digital specialists. The capability came in. The discipline didn't transfer out. Channel expertise without Communications grounding does not fail quietly: Delta's 2024 Palestinian flag pin incident on X (formerly Twitter), where a reflexive reply from the corporate account drew international media coverage within hours^{6NBC News, 12 July 2024.}, illustrates what that failure can look like.

Any technical hire into a Communications function needs a structured Communications induction: not generic onboarding, but deliberate transfer of the discipline's foundational principles. The same logic applies to any AI agent the function deploys.

The Communications leader's job is to be honest about which capabilities must be preserved, which must be developed, and which must be recruited — in that order.

That order matters more now than it did before. AI automation is starting to bear most heavily on the entry and junior levels of the function — the work through which craft and judgement have historically been cultivated. A function that thins that pipeline without a deliberate strategy for how junior talent acquires those capabilities is not just reducing headcount. It is reducing its future capability base.

The roles that do not exist yet

Across enterprise functions, a recognisable set of new AI roles is now emerging at scale. AI Solutions Architects, AI Governance and Risk Directors, Agent Product Managers, to name a few.

None of these have Communications-specific equivalents yet. The three roles below are not job titles you will find in current market data — they are capabilities to be built, acquired, or distributed across existing roles. They indicate the direction the function's capability base needs to evolve. The specifics will keep developing, the direction holds.

• Agent Brief Manager — owns how the organisation's AI agents are briefed to represent it externally, across every agentic deployment.
• Head of Signal Coherence — governs the coherence of the organisation's total signal architecture, so that machine compression produces the intended representation.
• Reputational Risk Signaller — stress-tests the trust consequence of AI-system decisions before they are made, and surfaces it into the governance conversation.

The first three dimensions are the structural clock's work. The next two — signal architecture and the trust consequence — are the immediate clock's work, and the terrain it operates on.

Signal Architecture

AI is not just changing how information is distributed. It is changing how organisations are interpreted.

An AI agent can scan your last quarter of Communications and produce a one-paragraph 'reputation snapshot' in seconds. That snapshot may become the shorthand that investors, employees, or customers carry into their decisions.

Persuasion no longer starts from a blank page. It starts from a machine-made baseline. That baseline takes different forms — what the model absorbed during training, what it pulls from the web in the moment, or what an agent has been briefed on — but in every case, it exists before the communicator enters the conversation.

The baseline is not fixed. The same question asked twice can produce two different answers. The same question asked of different AI platforms can produce sharply different pictures. And sometimes the system fills gaps with claims that the record does not support. You are not managing one interpretation. You are managing a range of them.

This is a structural shift. It calls for a new role for Communications: signal architecture — governing the coherence of the organisation's reputational signal in a world where AI constructs a picture from that signal for anyone who asks.

The governance gap

The total signal picture — everything AI systems ingest when they construct an organisation's reputational shorthand — is not owned by any single function. This is the governance gap.

Investor relations owns the earnings narrative. Communications owns media and executive voice. Marketing governs brand. Legal reviews the disclosures. Each does its job, on its surface, to its standards. Nobody reads the whole.

That was previously manageable. Inconsistency was absorbed — through relationships, through context, through the latitude human audiences extend. AI is changing that calculus. There is now an intermediary capable of reading the sum of all parts — and handing it over to anyone who asks.

Incoherence was always there. It was never this visible, and its cost never this high.

From visible to eligible

Earned media, corporate narrative, executive voice, owned channels — the surfaces AI answer engines weigh heavily are ones Communications already shapes. The obvious move from that position is visibility — ensuring the organisation appears in AI search results. But visibility is a starting point. Organisations can be accurately represented and still not make the cut. The harder question is eligibility: whether the organisation's signals are credible enough to be selected.

Signals only work when they are costly to fake.^{7Costly signalling theory: Spence, Job Market Signaling, 1973; Zahavi, Mate Selection, 1975.} Earned media, regulatory filings, and analyst reports carry structural credibility because they are inherently harder to manufacture. Eligibility is governed at the input level, not the output level.

A buyer asking an AI assistant for a list of vendors to shortlist will get back a handful of names. The companies that appear are those whose signal architecture — consistent, authoritative, and coherent across sources — earns that position.

Generative engine optimisation (GEO) — under whatever name the discipline settles on — governs the technical surface and that matters. What it cannot do is govern the coherence logic underneath: whether the signals across every function's surfaces add up to a consistent, credible picture when compressed into a single shorthand.

That requires a different kind of judgement — horizontal, cross-functional, and second-order. It is upstream of every individual surface, each with its own stake in how AI lands. Communications is the only function whose object is coherence itself — the total picture. Every other function reads that picture through the lens of something else: compliance, investor impact, brand, adoption velocity. Communications reads it to determine whether it holds for the organisation as a whole. That is its specific contribution.

The operational form follows from it: setting the standard for how the organisation talks about itself across the surfaces it controls; maintaining a live map of all the surfaces AI systems draw on; auditing what the machine currently concludes and tracing it back to its likely source; convening cross-functional alignment to keep the total picture coherent.

The signal only Communications reads

Signal architecture is not only external. The narrative the organisation tells employees about AI — why adoption matters, what it asks of them — greatly impacts whether adoption succeeds and whether value actually lands.

That narrative is also an external signal. When what the organisation tells employees about AI diverges from what it tells the market, the gap surfaces — in employee commentary, in how AI is used, in what gets disclosed, and what does not.

HR and the CAIO shape pieces of the internal narrative. Neither is positioned to read it as an external signal. That reading — whether the internal and external stories hold together as one — sits with Communications.

Discovery before influence

Communications carries a discipline directly applicable here: discover before attempting to influence.

The same sequence applies to AI as a synthetic intermediary. What is the system currently concluding about the organisation? Which signals is it drawing on most heavily? Where is the story coherent, and where is it fragmented or absent?

The discipline transfers. The logic that governs it shifts.

With human stakeholders, influence is reciprocal — you brief, observe, and adjust. AI introduces structural asymmetry. It does not negotiate and, at present, cannot be corrected in conversation in any way that changes what it tells the next person who asks. The feedback loop exists, but it is mechanical rather than relational.

Operating in this mode requires two things the function does not currently have in most organisations: technical fluency to make the discovery questions answerable, and the agency to act on what is found by sitting at the governance table.

The Trust Consequence

Trust consequence is what AI systems do to an organisation's reputation, and how that lands with stakeholders. As of 2026, nearly three-quarters of companies are planning to deploy agentic AI within two years, but only 21% have a mature governance model for it.^{8Deloitte AI Institute, State of AI in the Enterprise, 2026 edition, January 2026.} That gap is not widely named as a reputation risk — but that is where the consequence lands.

The broken chain

When reputational issues arise in human-mediated systems, there is usually a traceable chain: a decision was made, by someone, in a specific context, and Communications can work back through it. That traceability is imperfect, but it exists.

AI-mediated systems break that chain. When reputational consequence originates inside the system itself — a hallucinated output, an agent acting outside its parameters, synthetic content in circulation — the chain is difficult to reconstruct, harder to correct, and in many cases impossible to attribute cleanly without the ability to trace what the system did and why — a capacity most organisations do not have.

Air Canada's 2024 tribunal ruling, which held the airline liable for refund guidance fabricated by its chatbot, is the canonical example.^{9Maria Yagoda, BBC, 23 February 2024.} The airline argued the chatbot was a separate entity responsible for its own outputs. The tribunal rejected that entirely: deploy any system, whether rules-based or generative, own what it produces.

The second reading

Being an afterthought has always been the more costly configuration. In an AI-mediated world, where failures scale instantly, the cost is categorically higher.

AI governance is a matrix responsibility. The CAIO, Legal, IT, Risk, Compliance, and, increasingly, HR all have defined roles. What none of them is positioned to provide is the total trust consequence view.

The CAIO mandate is the closest of these to that view. The contribution Communications brings is not a substitute for that mandate but a second reading of it: a view of how the organisation's reputational picture is holding together, read independently of the function driving AI adoption.

The regulatory context is forming simultaneously and unevenly, but the direction across major economies is consistent: organisations will be held accountable for what their AI systems do and produce. Reputational accountability extends further. How an organisation uses and discloses AI — and the economic, social, and environmental repercussions of that use — is now a reputational variable in its own right.

Disclosure is not neutral. 73% of experts expect AI to have a positive impact on how people do their jobs, compared with 23% of the public^{10Stanford HAI, AI Index Report 2026, April 2026.} — a gap that does not close by issuing a responsible AI statement.

Agents as company spokespeople

In agentic workflows, AI agents become operational representatives of the company. They will require training — or in Communications terminology, spokesperson briefing. This is especially critical when different functions deploy their own agents, each briefed in its own terms, to its own stakeholder logic. The result is agentic cacophony: one organisation, multiple agents, multiple voices.

Communications has a specific contribution to make at each stage of how agents are briefed and governed.

• It develops and provides the corporate training material pack — the positioning, messaging frameworks, executive voice, and earned media record that forms the common narrative foundation all corporate agents draw from, regardless of which function deploys them.
• It checks whether the total signal architecture — after machine compression — produces the intended representation at the other end. Whether the picture the other agent receives about the organisation is the right one.
• And it brings the trust consequence view to the risk and escalation logic — which topics an agent should not speak to, when it should escalate to a human, and where the reputational exposure lies if those parameters are set incorrectly. Legal compliance is not the same as reputational coherence.

How and on what terrain Communications delivers this contribution will evolve as the agentic era matures. The principles will not — narrative coherence and the trust consequence view belong in the design of agent deployment, not in addition after the fact.

What Happens If Communications Does Not Move

Every month, the structural clock — the work of redesigning the function — runs without deliberate attention, and the ground shifts further.

What deferral produces is not irrelevance. It is structural demotion — the function being narrowed in scope and authority through a series of reasonable, defensible choices, each of which makes sense in the moment. More plausible, more gradual, and therefore more dangerous than an abstract threat.

It looks like this. An AI governance committee is formed, and the CCO finds out after the first meeting has already taken place, without Communications. Marketing builds a generative engine optimisation plan and shares it with Communications as an FYI. Leaders draft their own quotes (increasingly assisted by AI) and run them past Communications for editing rather than asking Communications to draft them — the function that once governed executive voice is now copy-editing it. A board paper on AI risk moves through Legal, IT, and the CAIO and reaches the board with no Communications input, because nobody thought to ask.

None of these is a crisis. They are the small, quiet decisions that reshape a function's mandate without anyone setting out to do so. Other functions — better positioned, more prepared, already present in the conversations that matter — naturally occupy the vacant territory. The CCO role becomes executional, one missed invitation at a time.

The compression accelerates when management frames Communications through the production lens. Shifting that framing — ask by ask, by answering the strategic question underneath the production request — is part of the internal redesign.

Communications is the discipline whose professional identity is built on anticipating consequence and thinking ahead of the problem. It rarely turns that instinct on itself.

The most consequential application of Communications' second-order risk thinking is to what is coming for Communications itself.