The debate around AI replacing contact centre agents is getting louder. Cresta’s answer, backed by its latest workforce report, is clear: that’s the wrong question.
Banzon told CX Today that
“76% of interactions today are being handled by both humans and AI already working in tandem.”
For Cresta, the future isn’t AI versus humans – it’s a unified workforce where both operate side by side. AI handles high-volume, lower-complexity conversations. Human agents own the high-stakes moments that require empathy, judgement, and trust.
“If you’re selling a $20,000 cruise package, you probably want that closed by a human agent.”
The logic is sound. Not every automation saves money when the cost of a broken relationship is factored in.
The Platform Behind the Experience
One of the report’s sharpest insights is the value of a unified data platform – one that consolidates AI-led, human-led, and digital interactions into a single intelligence layer.
The goal is continuity. Banzon gave the example of if a customer mentions their son needs an exit row seat on a flight. A year later, they call back. A unified system surfaces that context instantly, turning a routine interaction into a loyalty-building moment.
Fragmented tools produce fragmented experiences. Cresta’s thesis is simple: the infrastructure behind the interaction matters as much as the interaction itself.
Getting Automation Right
For enterprises balancing cost reduction with customer value, Banzon’s advice is to let strategy lead – not cost savings.
Start by identifying call drivers with low variance and clear end-to-end workflows. Check API readiness. Pilot before scaling. Automating the wrong interactions, or before the back-end is ready, creates broken experiences that erode the trust enterprises are working to build.
Synthetic Customers – the Next Big Thing in CX
The most significant announcement from Cresta is synthetic customers – AI-generated personas built from real customer interaction data. The applications are wide-reaching.
Before deploying an AI agent into production, enterprises can test it at scale against every customer type in their portfolio. Training human agents becomes more realistic, with synthetic personas recreating difficult caller types so staff can practise before those conversations happen live. And in marketing, teams can test campaigns against specific customer segments before any spend goes out the door.
The trajectory is compelling. As AI systems accumulate richer interaction data, synthetic models will become increasingly accurate – less a guess about customer behaviour, more a probabilistic model built on millions of real conversations.
The enterprises that move early won’t just reduce risk. They’ll iterate faster and deploy with greater confidence than competitors still learning from live feedback alone.
Synthetic customers aren’t just a feature. They’re the next layer of customer intelligence – and a signal of where the whole industry is heading.
Watch the full interview above. For more CX leadership insights from CCW Vegas, follow CX Today on LinkedIn.
