You've invested in AI. Now, how do you get your frontline staff to use it?

November 21, 2026
Lead Contributor Gary Wyatt

Back in February 2023, I wrote a blog called 17 Ways to Engage Customer-Facing Staff in Knowledge Management. It was all about getting frontline teams involved in KM initiatives and, more importantly, keeping them involved.

Fast forward to today, and organisations are investing heavily in AI assistants, copilots and all sorts of clever technology to help their customer service teams. But many face a familiar problem.

You've spent a fortune on the technology, launched it with great excitement, and a few months later discover that half your agents are still using their old methods.

Involve Front-line Human Agents Before You Launch

One of my biggest recommendations back in 2023 was to involve frontline staff from day one, and I'd say the same about AI.

Get agents involved in discovery sessions, let them test the technology and, most importantly, listen when they tell you something doesn't work.

Don't involve only your most experienced agents, either. New starters often have a completely different perspective.

I even suggested letting agents name the new KM system. Why not do the same with your AI assistant? Run a competition, let them vote and give them something they feel they've helped create.

A bit more interesting than announcing that everyone must start using "Enterprise AI Assistant Version 3.2" on Monday morning!

Don't Ignore the Rogue Human Agents!

I've always had a soft spot for the so-called rogue agents. They've got their own cheat sheets, shortcuts and little collections of useful information because the official knowledge base doesn't quite do what they need.

Rather than telling them off, get them involved!

They're often the people who can explain exactly why an existing system isn't working and what your shiny new AI assistant needs to do differently.

Of course, not every workaround is good, but there's usually something worth learning.

Get Out There and See What's Happening

Another suggestion from my original blog was for KM professionals to spend time sitting with human agents, watching how they work and occasionally even taking calls themselves.

I still think that's one of the best things you can do.

If you're introducing AI into a contact centre, go and watch agents using it during real customer conversations.

Are the answers useful? Does it save time? Are agents having to double-check everything? Or is it actually making their jobs harder?

You might discover rather more from an afternoon on the contact centre floor than from a dashboard full of adoption statistics.

Make It Fun

In my original article, I suggested launch parties, competitions, Knowledge Champions, branded mugs, fidget spinners and, of course, cakes.

I'm not suggesting a slice of cake will magically make your AI implementation successful. Although it probably wouldn't hurt.

But there's something to be said for making people feel part of something rather than simply announcing another change.

Recognise agents who provide useful feedback, celebrate improvements and show people how their suggestions have made a difference.

And please, when an agent reports that the AI has given a completely ridiculous answer, don't let that feedback disappear into a black hole.

Some Things Haven't Changed

AI might be able to find information, summarise conversations and suggest resolutions in seconds, but it can't make your employees trust it simply because you've paid for it.

That comes from involving people, listening to their experiences and demonstrating that the technology genuinely helps them do their jobs.

Looking back at my original 17 suggestions, I don't think I'd remove many of them.

The technology has changed enormously in three years, but the people side of Knowledge Management really hasn't.

That's something worth remembering before the next big AI rollout.

________________________________________

Gary Wyatt is an award-winning customer service, knowledge management transformation consultant with 25+ years’ experience improving frontline performance in complex, high-volume service environments.

He helps organisations turn fragmented knowledge, content, and processes into trusted, scalable ecosystems that improve frontline performance today and lay the foundations for AI, search, and automation tomorrow.

He's delivered measurable impact across major organisations, including 10–30% AHT reductions, 448 FTE-equivalent UK savings, fewer escalations, improved FTR, higher NPS and multi-million-pound/dollar technology and vendor-cost savings.

His approach is hands-on, collaborative and focused on delivering measurable results. Whether you're building a KM function from the ground up, transforming an existing capability or preparing your knowledge foundations for AI, Gary brings the clarity, experience and insight needed to drive real impact.

Connect with Gary at LinkedIn...

Back to main blog