Showing Up in AI Answers Isn’t the Same as Changing Behaviour

Stuart Bruce is co-founder and co-CEO of Purposeful Relations, which advises communications teams on AI, governance, and measurement. He has been in PR for more than thirty years and is a Fellow of the CIPR, PRCA, and AMEC. The firm also provides AI visibility auditing and advice on AI as a stakeholder.

Plenty of tools will sell you a number for how often you show up in AI answers. His argument is about what that number represents.

The prompts businesses need to see are the ones no tool can reach

Traditional search showed people your material. “AI is not doing that. AI is giving you something that it has synthesised, that it has created using all of those different sources that are available to it.”

“Every time somebody puts in a question or a prompt, it’s synthesising new information.”

“That’s why there’s a lot of talk about snake oil in this space, because there are 101 different tools out there that claim to measure your presence or your visibility in AI answers. And the reality is they can’t possibly work.”

“For traditional search, you’ve got access to data. For traditional media monitoring, you’ve got access to data. For traditional social listening, you’ve got access to data. And guess what, with AI answers, you don’t have access to the prompts that people are actually putting in and you don’t have access to the answers that AI is given. So the best these tools can do is directional telemetry.”

Some vendors do offer panel data drawn from real users. “Have you ever shared your prompts with one of these companies? No, and I don’t know anybody that does. Some basically buy the data from companies that get you to install browser extensions, which scrape your prompts.”

That is more plausible for a consumer brand, where people are more likely to have such an extension on a personal machine. “But there aren’t many corporates or institutions where you can install any old browser extension on your work computer.”

So the prompts a business would most want to see, from buyers, regulators, and other professional stakeholders, are the ones a panel built that way is least likely to capture.

The profession has run on a made-up number before

Bruce isn’t arguing that nothing can be measured. His concern is that a number built this way gets reported as though it were observed.

“We’ve got the hideous thing called advertising value equivalence, where basically people would measure a piece of media coverage and say, ooh, that would cost X pounds if you paid for it as an advert. And then they times it by a random number because editorial is worth more than advertising. And it is total bullshit, pardon my language. It’s just fraudulent. It’s made up. But despite that, because it’s a simple number, some people still use it. And my fear is that visibility is the new AVE.”

“Let’s say you come up in 80 per cent of answers. So what? What companies need to be focusing on is not just looking at whether we appear in answers, but how is it affecting stakeholder behaviour? Are they going to buy something? Are they going to apply for a job? Is it going to help a politician make a decision about some new legislation? And it’s linking those to the outcome rather than just looking at output, which is visibility.”

His own approach starts with the stakeholders the sample is meant to represent. “We’ll identify different types of stakeholder personas, such as a journalist, customer, employee, or politician, and then create a series of prompts that we think that persona is going to use.”

Human insight and AI make the initial call, but focus groups are the key check. “It’s a really good way of validating those other two methods. Because it would be quite expensive if you just did focus groups, but actually if you do those other two bits of research and then just test it, you see if we actually speak to people, are we getting similar answers back.”

For a small comms team the value isn’t time saved

Asked for his one piece of advice to a senior communicator in a mid-market organisation, he didn’t talk about tools at all.

He points to GCS Assist, the AI tool used across the UK government’s Government Communication Service. He describes its biggest gains coming in the smaller teams, where people had fewer colleagues to turn to.

“When they’re looking at using AI, it’s to improve what they do. It’s not about saving time. It’s really hard for them to develop best practice because they’ve got nobody to bounce it off. AI is the perfect foil for red teaming, bouncing your ideas off and basically upping your game. So it’s to embrace it, but not to do your job for you, but to improve how you do your job.”

That distinction comes up repeatedly in aibl’s mid-market work. Time saved is easy to count, but speeding up an individual task doesn’t tell you whether the outcome has improved.