Average handle time (AHT) went up... but nobody minded.
That is the surprising result from one Cresta customer that spent a few years rethinking its contact center.
It cut inbound volume by two-thirds; yet total handle minutes fell by only 50%, because the conversations that remained were far longer.
“AHT is up, but nobody is upset by this result,” said Devon Mychal, VP of Product Marketing at Cresta, in an interview with CX Today.
Over the same period, the customer’s NPS rose by 50 points.
But what does all of this really mean?
When AI Takes the Easy Calls, the Hard Ones Take Longer
For leaders who have been told for years to shorten calls, this will feel more like vindication than shock. Given AI's ability to handle more of the simple requests, it stands to reason that human agents will be left with more complicated, time-consuming customer interactions.
Mychal explained how those remaining calls are also “the most valuable or high-leverage conversations.” They decide whether revenue is won or lost, and whether a customer stays or looks elsewhere.
But if AHT is no longer the holy grail contact center metric, what should organizations be measuring instead?
Mychal argues that "we should judge these conversations based on those outcomes, not how long that they take.”
AHT still drives cost and staffing, but a Cresta-sponsored CCMA report suggests leaders are moving on.
Indeed, in the report, Rhian O’Hare, Head of Customer Service at Marks & Spencer, said:
“If we take time to resolve an issue and protect a customer who will buy with us ten more times within a year, that matters far more than whether our AHT was within target.”
Count the Calls You Prevent
The Cresta customer did not stop at automation. It used insight from its conversations to fix the products, policies, digital experiences, and confusing communications that were driving calls in the first place.
“Instead of thinking about containment, they actually think about conversations that they are able to prevent,” Mychal explained.
Those minutes are “saved for the end customer, not just their business.”
The Contact Center Hears It First, If Someone Listens
The CCMA report referenced earlier claims that the contact center is “uniquely positioned as an early warning signal of public sentiment,” yet many companies still treat it as the last stop in the customer journey
This point was put to Mychal, who first pointed to the technology, detailing how legacy speech analytics lacked depth and early LLM tools suffered from reliability problems, including “bad sampling, hallucination.”
“That all has changed in the last 18 months,” he said.
But the bigger barrier is organizational. The companies seeing the most impact have built “dedicated insights to action functions” rather than expecting contact center teams to raise the alarm without a remit.
For example, at United Airlines, a dedicated team uses Cresta Insights and AI Analyst daily to find fixes, many outside the contact center, and takes quantified evidence to the teams that own them.
Why AI Projects Stumble
When it comes to the widely discussed issue of AI projects failing to live up to their promise, Mychal sees two common patterns:
First, AI “cannot operate effectively in a vacuum.” It needs connected systems, real-time APIs, and clean knowledge. “Knowledge that is written for human consumption is often not optimized for AI systems,” he said.
Second, teams throw point solutions at problems, especially conversational AI agents.
They start on intuition or a vendor’s pitch and design around SOPs and golden flows, which take “a very sanitary view of what’s happening in these conversations rather than the messy, complex reality of it.”
“It’s easy to get to a great demo,” he quips.
Reinvest the Savings
Customer experience and ROI have traditionally been two warring factions, with boards often failing to see the value in any CX programs that do not involve reducing headcounts or increasing efficiency.
However, Mychal believes the tide is starting to turn on this thinking. He cites a Fortune 50 telco that uses its contact center to spot missed revenue opportunities, and argues that “AI is significantly disrupting the cost basis of providing great customer experiences.
"But the organizations that are really differentiating themselves are going to be those that reinvest those cost savings in ways that are creating value for their customers, and building on that to accelerate their revenue growth.”
