PR is measuring the wrong things. AI just made the old metrics indefensible

Your CEO wants more coverage. Your team wants to know whether AI will take their jobs. Your audience wants to know whether anything your company says is true.

And the monthly report says impressions are up.

This is the uncomfortable position PR finds itself in. We have extraordinary new tools for producing messages, just as the authority of the people delivering them is being questioned.

The trust problem is real, although it needs describing accurately. Business remains the most trusted of the four institutions in Edelman's 2026 survey. Yet its 2025 report found 68% worried that business leaders deliberately mislead people, up 12 percentage points since 2021. Trust in business does not mean automatic belief in its spokesperson.

Younger audiences sharpen the challenge. A 2026 survey of 2,000 Americans aged 18–28 found 72% trusted customer reviews when judging a brand's credibility. Only 46% trusted PR campaigns, stunts and branded activities. This is one US survey, but the warning is clear: the official message competes with people who have used the product, worked inside the company or checked the claim.

A bigger megaphone will not fix a credibility problem.

Meanwhile, inside communications teams, there is what I call "Anx-AI-ety". Will my skills still matter? Is another agency doing this faster? Am I falling behind? The symptom is FOLO: fear of losing out.

The temptation is to demonstrate progress through production. More posts. Faster releases. Cheaper reports. Another AI subscription. Everyone looks busier and more modern. Whether anyone believes us remains unclear.

That is why the measurement debate has become urgent. If AI helps us produce ten times more, we need to explain what became ten times more valuable.

Consider a hypothetical sustainability launch. The coverage is excellent. The CEO's messages appear everywhere. Then an employee posts evidence that the claim does not match reality. A researcher picks it up. Customers share it. Journalists start asking questions.

The dashboard is still green. The reputation is heading the other way.

A useful monitoring system should spot the challenge before it dominates the headlines. But detection is only the start. Who is making the allegation? Is the evidence sound? Who trusts that person? Which journalist, employee group or customer community could carry it further?

This is where stakeholder intelligence becomes central to PR. Follower counts tell us potential reach. We also need to know whose word carries weight on this issue, with this audience, at this moment.

The answer may be an engineer, a customer or an independent researcher. Paying a bigger influencer will not necessarily help. Credibility has to fit the claim.

Next, check the company's memory. What did we promise last year? What did employees tell us? Have we answered this concern before? Connect those records to the live debate, so the team can brief leaders with evidence rather than scramble through old emails.

Now the briefing earns its keep. It explains what changed, who matters and what to do next. Perhaps the answer is a correction, an independent review or a change in practice. Sometimes the best communications advice is to stop repeating the message.

Then engage, record the commitments and check what happens. Did the company fix the problem? Did the people affected accept the evidence? Did employees understand the response? Did customers become more willing to recommend us?

Those questions give alerts, maps, briefings and dashboards a purpose. They connect what PR produces to what people believe and do.

AI can support that entire sequence. Agents can track developments, retrieve past commitments, update influence maps and prepare options. People must check the evidence, make the final judgement and own the response. Shared knowledge should grow with every engagement, rather than disappear when someone leaves.

Even AI visibility needs this discipline. Being mentioned in an AI answer does not prove that anyone trusts you. Counting citations can become the new counting impressions. Check what the answer says, why it says it and whether it helps people make an informed decision.

For the next campaign, ask a harder question before approving the creative: why should this audience believe us? Name the people they trust. Identify the evidence they need. Check the promises we have yet to keep. Then decide what success would look like in their response. That conversation deserves as much attention as the launch plan and media target list.

We still need audience research and honest limits. A change in trust may reflect better products or conduct. PR should show where it contributed, rather than claim every success.

The summit's theme, "From Message to Mandate," is an invitation to lead. We should use AI to understand people earlier, find credible, practical answers and help our organisations act on them.

Anx-AI-ety will not be cured by producing more noise. It will ease when we can show why our work matters.
---

Terence Lyons is Founder & CEO of TSC.ai, a stakeholder intelligence company working with multinational corporations across extractives, energy, financial services and consumer sectors. He is an Annual Guest Lecturer at Oxford Saïd Business School and speaks at MIT EmTech Asia, TEDx Singapore, and the Public Affairs Council.