From AI to ROI: What It Means to Navigate the ‘Prove-It Era’ of Customer Engagement

Customer experience is at a turning point. Expectations are rising, budgets are tightening, and brands must prove the value of every interaction. It’s now the “Prove-It Era” of customer engagement and experience—where businesses can’t just assume that they’ll improve CX by just prioritizing it. Leaders can’t assume, either, that CX improvements are generating value for the business—like revenue growth or opex reduction.

CX programs need to show measurable impact, clear ROI and strategies that directly drive business outcomes.

In a recent webinar, “The Prove-It Era of Customer Engagement: Trends, Truths, and What’s Next,” a panel discussed current CX trends and how to navigate them. Covering everything from AI to ROI, the panel featured these experts:

• Mila D’Antonio, Principal Analyst, Customer Engagement, Omdia
• Adrian Swinscoe, CX Author and Advisor
• Dan Hartman, Director of Product Management, CSG

Discussion topics included:

• Going beyond segmentation: Why unified, personalized CX is still so hard to get right, and what it will take to do it
• Using AI wisely: A realistic look at how AI is being used today—and what organizations should avoid as they prepare for the future
• Making the case for CX investments: How to show the value of CX initiatives and connect them to business goals

You can watch the full conversation or read on to see some of the major takeaways.

Personalizing With—Not at—Customers

More brands are prioritizing personalization as a core component of their CX strategy. A recent Deloitte study found that brands expect to increase their annual personalization budgets by 29% compared to the previous year. Yet that study also found that consumers only recognized 43% of their brand experiences as personalized, whereas brands said they personalize 61% of experiences on average.

The panelists dug into this disconnect between the personalization that brand leaders believe they provide and the personalization consumers expect. Adrian Swinscoe used his phone as an example: As he’s had the same phone number for more than 25 years, it’s as unique an identifier for him as any other piece of data, he said.

“But if I call up a brand that I’m a customer of, then wouldn’t it be great if they can take that and go, ‘Hi, is this Adrian? How can I help you?’ But tell me a brand that actually does that straight off the bat.” He added that brands can confirm identity with security questions afterward, but “but first impressions matter because it shows that you’re listening and you’re leaning into it.”

Dan Hartman pointed out that most organizations have yet to break through segmentation to 1:1 personalization.

“I think if you look at how the average brand interacts with consumers, it’s definitely still at the segment level,” Dan said. He added that personalizing by segment is fine for a broad acquisition campaign, where connecting with a very small percentage of a brand’s audience is considered a success. But when the audience is an existing base that the brand has a relationship with, the engagement level should be much higher, and segmentation no longer cuts it, Dan said.

“The thing to realize here is that a segment is not a person. You have to treat a person like a person, or else it’s not personalized,” he added.

Customer interactions should not only be personalized on the individual level, but also in real time to keep pace with how quickly customers’ situations change, Dan said. He noted that personalizing interactions in real time, and at scale, is difficult to do without the right technology.

Mila D’Antonio said scaling personalization requires even more than that.

“The challenge isn’t just scaling the technology, it’s scaling the listening, the responsiveness, the choice—really shifting from pushing personalization at customers to enabling personalization with them.”

The difference is doing more than offering predictive engagement and offers—but also engaging customers in a way that is relevant and meaningful to them.

“We have to look at personalization as being a whole lifecycle strategy,” Adrian said. “It can’t just be the domain of just sales and/or marketing, because we have to think about what customers preferences are across the lifecycle. We have to respect them.”

‘Things Are Getting Real Now’ With AI

CX teams are expected to personalize every message, react in real time and keep the experience consistent. But without AI, the scale, speed, and adaptability customers demand just isn’t achievable. However, if teams can’t use it or don’t trust it, it won’t scale.

According to a survey presented in Omdia’s State of Digital CX 2025, 49% of CX professionals said difficulty integrating AI tools into workflows was a primary obstacle their organization faces in delivering personalized, unified customer experience.

“There’s been a lot of investment in AI, but without thoughtful integration, these tools often sit on the side, unused, untouched,” Mila said. “For AI to make a real impact, it needs to be embedded into the flow of work, whether that’s guiding a contact center agent in real time or helping a manager spot trends across various teams.”

Adrian said organizations need to remember that AI isn’t a panacea or magic—it’s a tool—and they should approach the technology with that mindset.

“Rather than starting with the tools, you need to start with, ‘What is it I want to deliver?’” Adrian said. Leaders could ask themselves what they want their organizations to be, or what kinds of experiences they want their customers and employees to have. “And when you start with the end, you might then work backwards from there, then you’re probably going to have a better grasp or handle on what sort of data you need, how you need to structure it, and the AI-powered tools that are going to help you deliver that.”

Mila said she’s seen organizations apply high-impact, high-ROI use cases in call routing or self-service. Some are going beyond automating customer interactions and using AI to analyze customer conversations at scale to spot patterns, like sentiment trends, that can help improve live agent performance or even the products and services themselves.

As quickly as businesses want to move on AI implementation in CX, there are real hazards in “doing it wrong.” About three-quarters (76%) of telecom consumers in a CSG survey said a frustrating experience with their AI-powered customer service would negatively impact their loyalty to their provider.

“So with the bias that customers are bringing [to AI experiences], you may not even get a second chance if you drop the ball,” Dan said.

“Things are getting real now” in AI-powered experiences, Dan said, adding that there’s an opportunity for brands to “figure out how to cut down on those hard learnings upfront.”

To Prove the ROI, Find the KPI

What makes this the “Prove-It Era” in customer engagement? CX leaders are acknowledging that a CX initiative—from enhanced personalization to AI adoption—won’t survive budget scrutiny unless it can demonstrate real value to the business.

One of the highest hurdles in proving CX ROI is in finding the right KPIs to convey it. By 2027, 60% of service and support leaders will report failing to obtain investment in their most critical projects, due to their use of poor-fit metrics that lack credibility or fail to sufficiently demonstrate the value of the investment, according to Gartner research.

Dan said he sees some organizations trying to correlate their VoC metrics with financial KPIs by trying to assign monetary value to gains in NPS or similar metrics. But those correlation exercises “are DOA,” he explained.

“They just never translate to the bottom line because you can earn three points in NPS by sending half a million more transparent emails, or by giving out half a million free iPhones. And they carry very different costs.”

For the purpose of calculating ROI, VoC metrics aren’t the KPI, Dan added. They help the organization spot where impact was made in customer loyalty, but they don’t communicate its value.

“If you’ve ever seen somebody on a beach with a metal detector, and they find a gold ring, when they go to the pawn shop to sell the ring, they don’t bring the metal detector in with them,” Dan said. “It already did its job. And that’s kind of what I see as the problem when people are trying to shop around VoC metrics as the point.”

“When we talk about seamless customer journeys, one of the first things I ask is, what’s the KPI that you’re trying to move?” Mila said. “I think too often organizations are chasing these really broad transformations without really anchoring them to a specific outcome.” Those outcomes might be improving that first contact resolution or reducing churn.

“Starting with the KPI really helps cut through all that noise and align teams around what actually matters,” she added. “From there, I think you can really identify where the journey is breaking down and what internal barriers—whether they’re organizational or technical or process related—are really standing in the way.”

Dan added that when in doubt, leaders can choose the KPI that the CEO has identified as a priority, or any other shared metric that can “rally the troops” at the organization.

Adrian said most organizations focus on increasing customers, revenue or profitability. “The thing is you can’t do those three things—they’re outcomes.”

What organizations should focus on instead, Adrian said, are the levers for achieving the outcomes, which include:

• Conversion
• Average customer spend
• Frequency of spend
• Cost to serve

CX teams should be looking at an initiative and asking themselves how it improves one or more of those levers, Adrian said. “And if it doesn’t, you’re not speaking the CFO’s language,” he added.