Something has shifted in how people find things out. More and more searches end without a click. The answer engine reads the web, summarises it, and hands back a paragraph. The Reuters Institute Digital News Report has tracked audiences moving toward AI-mediated discovery and away from the open link.

If that is where attention is going, impressions and press mentions no longer tell you what you need to know. They count activity on a stage the audience is leaving.

Why the old numbers mislead

Impressions once implied reach. In a summarised web, a page can inform an answer that millions read while its own traffic barely moves. Press mentions counted visibility. But if an answer engine synthesises coverage without naming you, the mention exists and the credit evaporates. The metric survives. The meaning is gone.

If the machine answers for you, the real question is whether it names you at all.

What to measure instead

A 2026 scorecard should track presence where decisions now form. Five signals matter more than the old dashboard.

  • Share of voice inside AI answers. When someone asks an engine about your field, does your organisation or your experts appear in the response.
  • Named experts in circulation. How many of your people are established enough to be cited by name, not folded into a generic summary.
  • Consistency. Steady, credible publishing that models learn to associate with your topic, rather than one-off spikes.
  • Engagement quality. Saves, thoughtful replies and shares from the right people, not raw view counts.
  • Downstream signals. Inbound conversations, inbound talent and inbound invitations that trace back to visible expertise.

Why named experts sit at the centre

Answer engines lean on sources they can attribute and trust. A named specialist with a consistent public record is exactly the kind of signal a model can latch onto and repeat. An anonymous corporate page is easy to blend into the background. The Stanford HAI AI Index charts how central these systems are becoming to how information is found. Being a source these systems recognise is fast becoming the point.

Build the scorecard now

You do not need perfect instrumentation to start. Ask the engines your buyers ask, and record who they name. Note whether your experts appear. Track it monthly. The trend line tells you more than any single reading, and it tells it early.

Comms teams that keep reporting impressions will look busy while their actual influence quietly slips out of view. The ones who measure presence in AI answers will see where trust is really accruing, and be able to steer toward it. New terrain rewards a new scorecard.

Ripple helps communications teams build the named expert presence that AI answer engines recognise and cite. If you are rethinking what to measure, let's talk about making your experts visible.