What happens when agentic AI becomes your contact centre analytics team

As anyone reading this knows, most contact centre leaders aren’t short of data, they’re short of time to make the most of it. Picture it: dashboards pile up, analyst queues grow longer, and somewhere in the gap between the data and the decision, opportunities get missed. Sound familiar?

That was the starting point for EdgeTier’s latest webinar: 1,000 Analysts, Always On: How Agentic AI is Transforming Contact Centre Analytics. Joined by Caroline Cooney-Hurrell, Global Customer Support and Contact Centre Operations Leader at Explora Journeys, and Milo Oostergo, Principal Solutions Architect at AWS, EdgeTier Co-founder and CCO Bart Lehane walked through what this shift in AI capability actually means for contact centre teams, and what it looks like in practice.

A startup running at luxury scale

Before the conversation moved to technology, Caroline gave some context on Explora Journeys that set up everything that followed.

Explora is a luxury cruise brand launched in 2023, part of the MSC Group. It currently operates two ships, with a third arriving this year and a total of six planned. Caroline joined in 2020 to design the customer experience function from scratch and has grown the team to over 130 agents managing upwards of 3,000 to 3,500 interactions daily, operating across South Africa, Manila, the US, and Europe.

The challenge that shapes everything at Explora is one that many contact centre leaders will recognise, even if the stakes are different: how do you keep a genuinely personal guest relationship intact as you scale? When a booking can represent tens of thousands of euros and a customer’s entire annual holiday, a poor interaction doesn’t just affect satisfaction scores. It affects retention.

“We have people who are spending a lot of money, and are highly invested in spending their free time with us,” Caroline said. “For us, it’s about knowing very quickly when something goes wrong. Very quickly, when something is right. And then being able to decipher in the hundreds and thousands of minutes of interactions what really is important to people.”

What’s actually changed in the underlying technology

Milo from AWS provided useful grounding on why agentic AI is possible now in ways it simply wasn’t a few years ago.

The shift, he explained, comes down to three things. Foundation models accessible via services like Amazon Bedrock have made it possible to summarise, reason, and generate at scale. Conversational AI has moved from rigid intent-based bots to models that handle natural, interruption-prone dialogue. And critically, connecting AI agents to existing business systems has become dramatically simpler.

“A few years ago, to get an agent to interact with your systems, it required integrations for every backend,” Milo said. “Today, with standardised interfaces like the Model Context Protocol and services like Amazon Bedrock Agent Core, you can give agents access to your tools through a consistent interface, and the agent can reason what to do and calls the right APIs. You don’t have to build hard-coded workflows anymore.“

That last point matters for contact centre teams specifically, as fewer hard-coded workflows means faster implementation, easier iteration, and AI that can genuinely adapt to what’s in front of it rather than what was anticipated when it was built.

Ask Spotlight: what an agentic analyst actually does

The main product focus of the webinar was Ask Spotlight, our newly launched agentic AI analyst within Explore by EdgeTier.

The concept is straightforward to describe, harder to replicate: you ask a question in plain language, and instead of returning a pre-built dashboard, it investigates, pulls relevant data, explores different avenues, checks its own reasoning, and returns a structured analysis with context and recommendations. The “1,000 analysts” framing comes from the fact it can do in minutes what a human analyst would take hours or days to complete, across every conversation in your data.

Caroline’s team were early beta testers. Her first reaction was honest: “I remember thinking, oh god, this is more work” but that sentiment lasted about as long as it took her to type her first question!

“When I went in and saw it, and just saw I could type in a sentence, I was like, oh my gosh. It was giving me the data on what was coming in, what the trends were. It was answering questions I hadn’t even thought of.“

One moment stood out in particular. During a period of operational disruption, Explora had announced the relocation of one of its ships. Ask Spotlight surfaced the pattern immediately: not just the volume increase, but the specific concerns guests were raising and what the team actually needed to communicate. “We didn’t have to wait for a week of data to say, okay, we’ve got this wrong, or we need to change tack. It was kind of telling us ahead of time.”

Her broader reflection on the beta experience was one that contact centre leaders without large analytics teams will understand immediately: “I wouldn’t even have known to ask an analyst for some of the stuff it’s giving me, because I’m not an analyst as such.”

What this means for the rest of the business

One thread that ran through the second half of the webinar was data accessibility: who gets to use contact centre insight, and what happens when you remove the friction.

At Explora, EdgeTier access has been extended to marketing and strategy teams. The result, as Caroline described it, has been a shift from “come and ask us what’s going on” to genuine self-service. People across the organisation can now form and test their own hypotheses before bringing them to the contact centre team.

“When you’re going to ask the expert, you’re asking the right question. You’re not fishing in a haystack.”

Asked about extending AI beyond traditional customer service, Caroline saw clear potential in finance and audit functions having rapid access to customer interaction history, potentially removing several verification steps. For the marketing team, the practical application is already in use: checking actual customer sentiment against planned campaign assumptions before going to market.

Where agentic AI goes from here

The final section addressed the future. Milo outlined proactive and predictive engagement as the next meaningful step. Most contact centres today are reactive, but with the data and AI capability that exists now, teams can start detecting issues before the customer picks up the phone and act on them first.

His practical advice for teams starting this: get your data in order before you build. Clean knowledge bases, an accessible CRM, systems with APIs. “Your AI agents are only as good as the systems and the data they can access.” He recommended a crawl-walk-run approach rather than trying to move from zero to full autonomy overnight.

Caroline’s closing message was direct: “Fight for it with your organisation. Fight to at least dip your toe in the water, because if you don’t, you’re going to be left behind.”

She put it in context. She started her career when a toll-free number was considered exotic. The teams that adopt early shape what comes next. “If we had said back then, oh, we’re not going to get an email because we’re a bit afraid of it, we’d have been stuck.”

EdgeTier’s Ask Spotlight is live now. If you want to see what an agentic analyst looks like in your contact centre, get in touch with the team.