AI has moved well beyond the “interesting experiment” stage in the contact center.
At Five9 CX Connect Europe, the message was pretty clear: AI is already in production, but most organizations are only scratching the surface of what it can do.
Across the sessions I attended, the conversation kept coming back to the same question: what happens when AI stops being about replacing people or cutting costs and starts being about making the entire customer experience better?
From Five9’s “Humantic CX” approach to Agentic Quality Management and a rethink of how we measure AI ROI, there was a lot less talk about AI for AI’s sake and much more about what happens when you actually put it to work.
Here are my key takeaways from the day…
Five9 wants AI to make customer service more human
Five9 opened CX Connect Europe with a simple idea:
When customer service gets easier, people interact more, not less.
That set the direction for an event where AI is clearly going to dominate the conversation.
President Andy Dignan opened by talking about Five9 being laser-focused on AI in CX. That’s a very different position from the traditional contact center vendor approach, where AI can sometimes feel like another feature added to an existing platform.
The message here is that AI is changing the way customer interactions work.
The phrase used was “smarter CX with every interaction.”
New CEO Amit Matradhas followed with an interesting observation:
“AI was supposed to shrink contact centers, yet it keeps growing.”
The answer, according to Five9, isn’t to replace humans altogether.
AI can handle more of the routine work, while humans move toward the interactions that require complex solutioning: empathy and vulnerability. It also means AI can help enable customer service around the clock, with humans brought in when they’re actually needed.
Enter “Humantic CX”
Five9 has given this approach a name: Humantic CX.
The pitch is pretty straightforward. This isn’t about using agentic AI to replace humans. It’s about AI and humans working together, with each making the other better.
The first product demo of the day brought that idea to life.
The scenario was an insurance claim following a flooded house. The AI voice agent handled the conversation from start to finish, worked through the claim and successfully reached a resolution.
Then the customer pushed back on the proposed resolution time.
That was the point where AI handed the conversation to a human.
That distinction matters.
The objective isn’t necessarily to remove humans from customer service. It’s to stop using humans for the parts of the interaction that AI can handle, while making sure a person is available when the situation needs judgment, empathy or negotiation.
It’s a much more credible proposition than simply saying AI will replace contact center agents.
What’s coming next in Europe?
Five9 also announced three major additions coming to the UK and EU on November 3:
- Voice AI
- Genius Routing
- Agentic Quality Management
The latter two should be particularly interesting as the day progresses, because they move beyond automating individual conversations and into how the wider contact center operates.
The UK needs to pay attention to the EU AI Act
Brexit doesn’t mean you can ignore the EU AI Act.
My first breakout section was What Every CX Leader Needs to Know, focusing on how large businesses can trust their AI with the strict regulation of the EU AI Act looking over its shoulder.
For CX teams, that makes transparency particularly relevant. The rules now include requirements around telling people when they’re directly interacting with AI, as well as requirements for certain AI-generated content. These transparency obligations started applying on August 2, 2026.
The broader point from the session was that the UK and EU may have different rulebooks, but the foundations for trusted AI are broadly the same:
- Transparency
- Human oversight
- AI inventory
- Auditability
For any UK contact center serving European customers, that’s worth paying attention to. There is, after all, a maximum penalty of £35m or 7% of your business.
Agentic Quality Management: putting all that QM data to work
Agentic Quality Management (QM) has become a hot topic as we try to shift away from treating quality management as a collection of disconnected tasks.
Once you’ve connected your customer interaction data, you can now automate next actions and in-call coaching to improve what happens next.
- It starts with routing: AI agents identify the customer’s intent, while Agentic QM builds a picture of human agent proficiency. The system can then match the customer with the best-fit human while still maintaining the required SLA.
- Agent Assist picks up the conversation from there: Using AI transcription to understand the interaction and provide real-time guidance to the agent.
The bigger theme was moving QM out of its silo.
That sounds obvious until you ask where coaching data actually lives. Even today, around half the room put their hands up when asked whether their coaching data is sitting in Excel and Word.
That’s a lot of useful data that isn’t doing much beyond documenting what happened.
The AQM demo showed what happens when that data is connected and made actionable. It was one of the more impressive demos of the day and it’s worth seeing for yourself rather than trying to explain it in a blog post.
There’s currently a free implementation offer valued at $14,000, along with delayed billing of up to three months. So if Agentic QM is on your roadmap, this looks like a good time to get a demo from Five9.
Proving AI ROI: from cost center to value creator
My good friend and favorite presenter, Patrick Watson, was next up to tackle what I’d argue is the most important topic when it comes to moving contact centers from cost centers to profit centers: what is AI actually worth?
Armed with a ton of data from Cavell’s CX Enterprise Research Report Q3 2026, the session put some useful numbers behind the current state of AI adoption.
🤯 96.3% of contact centers now have AI running in production in some form.
But much of that AI is still focused on relatively simple interactions, with customers escalated to a human when the situation requires it.
Sound familiar?
It’s the same “Humantic CX” theme from the opening keynote. AI handles what it can, humans step in when they add more value.
The problem is that many organizations are still measuring AI using yesterday’s metrics.
- 40.8% of contact centers see AI primarily as a cost containment tool
- 16.5% see it as a defensive necessity
- 5% think it’s over-hyped and under-delivering
That leaves around 70% of organizations that aren’t yet treating AI as the strategic value creator it could be.
Old metrics vs. new metrics
This was probably the most important conversation of the day.
We still count things like containment rate, average handle time, and first contact resolution.
But those metrics don’t necessarily tell us whether the customer actually got their problem solved.
The proposed shift is toward:
- Containment rate → Verified resolution rate: Was the issue actually solved?
- Average handle time → End-to-end time and effort: How long did the customer wait and how much effort did they have to put in?
- First contact resolution → Repeat contacts and reopens: Did the fix actually hold?
That changes the ROI conversation.
Instead of asking “How many calls did AI keep away from an agent?”, ask “Did AI resolve the customer’s problem?”
That’s a much better measure of value.
Where AI is delivering value
The data showed that AI is already delivering measurable results across contact centers.
The biggest reported benefit is:
- Faster resolution times (49.67%)
- Improved agent productivity or effectiveness (44.06%)
- Improved customer satisfaction (43.93%)
- Compliance, QA and risk management (39.92%)
- Reduced cost per interaction (35.1%)
Only 2% said AI hasn’t delivered clear value anywhere yet.
So the ROI isn’t hypothetical.
The challenge is measuring it properly.
The next wave is moving beyond AI answering questions, assisting agents and automating tasks toward AI that can own complete customer cases, anticipate issues before customers call, and orchestrate channels, people and systems.
That’s where the contact center starts looking less like a cost center and more like a business function capable of creating measurable value.
Conclusion
There was a consistent message running through the sessions at Five9 CX Connect Europe.
The future of the contact center isn’t human or AI. It’s human and AI.
AI can answer, assist and automate today. The next step is for it to own more complex customer cases, anticipate problems, and orchestrate the people, systems and channels needed to resolve them.
But that only works if you stop measuring success using metrics designed for a very different contact center.
Containment rate and average handle time tell part of the story. Verified resolution, customer effort, and genuine business outcomes tell us much more.
The technology is moving quickly. The bigger challenge now is making sure contact centers change with it.
And if Five9’s roadmap is anything to go by, November 3 is going to be another interesting date for CX teams in the UK and Europe.
