CX Teams That Lead With Revenue Capture Gains The Cost-Cutters Never See

The reflex when AI arrives in the customer experience function is to point it at the cost line. Customer service is expensive, often eight or nine figures a year for a large enterprise, and the temptation to automate that expense away and drop the savings to the bottom line is powerful. It's also a strategic error, because CX is not primarily a cost center. It's a revenue engine that's been tuned over decades to retain, renew, upgrade, and cross-sell, and the money flowing through the function dwarfs the money it costs to run.

Mark Meghezzi is Head of EMEA at Cresta, leading the company's go-to-market and customer work across Europe, the Middle East, and Africa. His career spans a large telco and several high-growth environments, giving him a view across dozens of industry verticals of how enterprises actually evaluate and deploy AI in their customer experience. When an enterprise weighs AI against the cost of customer service, Meghezzi points it to the far larger figure sitting right next to it.

"If you're spending $100 million a year on customer service, there's probably $1 billion of revenue flowing through. It’s not particularly smart to save a percentage of the $100 million cost if it meaningfully impacts the $1 billion, which it often does," Meghezzi says. That ratio reframes the entire business case for AI in the customer experience.

The cost lens hides the bigger number

The reason the cost-cutting instinct is so common is that the cost is easy to see and easy to measure. A CFO can point to the operating expense of a contact center down to the cent. The revenue moving through that same operation is harder to attribute, so it gets underweighted in the AI conversation even though it's far larger.

Meghezzi's point is that the two numbers sit on opposite sides of the same equation, and optimizing one while ignoring the other is how companies destroy value. Every cancellation saved, every renewal secured, every upgrade closed, and every downgrade talked back up happens in that operation. "We as companies have set those things up over the last few decades to be highly tuned revenue-generating or revenue-retaining machines. So it costs money, sure, but everyone gets more excited about making more revenue."

The inefficiency is real. There is high-cost human labor, redundant time, and legacy friction that technology should absolutely improve. But those improvements have to be made with the revenue engine intact, not by dismantling it to book a one-time saving.

Supercharge the revenue, don't shave the cost

The more valuable question, in Meghezzi's framing, is not how much cost AI can remove but how much revenue it can add. A CX team that gets meaningfully better at generating and retaining revenue produces gains that make cost savings look trivial by comparison. "Can you find a way to make your CX function 10, 20, 30 percent better at generating revenue for you as a company? Because that's just infinitely more exciting than taking a few cents off the dollar in terms of your cost base."

The math is straightforward once the revenue side is in view. A few percentage points of improvement against a billion dollars of flow-through revenue is a far larger prize than eliminating a large share of a hundred-million-dollar cost base, and it comes without the risk of damaging the customer relationships that produce the revenue in the first place.

The risk isn't hypothetical. Aggressive automation aimed purely at cost can degrade the moments where revenue is won or lost: the renewal conversation, the retention save, the upgrade nudge. Cutting cost in those moments to save cents can leak dollars, which is exactly the trade Meghezzi warns against.

Friction is where revenue leaks

Part of the revenue opportunity hides inside problems that look like service-quality issues but are really revenue-leakage issues. The most common is the loss of context across channels, which forces customers to repeat themselves and erodes the relationship at every touchpoint.

Meghezzi describes this from his own experience as a customer of a bank he considers one of the more forward-thinking ones, which is what makes it striking. "From an app to a website to a phone conversation, I had to repeat myself maybe eight times. A fairly simple inquiry, spoken to humans all the way through, and still the context was lost through every single one of those touchpoints," he shares. The friction is pervasive even at sophisticated operators, and it's eminently solvable. "It just requires some prioritization of stitching those parts together."

The revenue relevance is that every one of those repeated explanations is a moment where a customer's patience, and their willingness to renew or expand, thins a little further. Fixing the context problem protects the revenue that the friction slowly bleeds away.

Bring revenue into every conversation

For executives trying to evaluate where AI actually pays off, Meghezzi's guidance is to change the lens through which every decision is judged. Revenue and cost should both be present in every conversation, and revenue should usually lead, because it's the larger number. "There's no sense in automating something to save 5 percent of your cost base if it's going to cost you 10 percent of your revenue. So always look through both of those lenses at every turn."

The second piece of his guidance is counterintuitive: don't start with the safest, lowest-risk automation, because that's usually also the lowest-return work. The better move is to find the high-impact zone, typically where volume is high or customer unhappiness with the current process is high, and tackle a genuinely hard problem early. "There's always a sweet spot somewhere, high impact in terms of high volume or high unhappiness with today's process. Tackle a really hard problem early, because that's where you start to feel some of the tricky bits."

Confronting a difficult, high-value problem at the start sets the right expectation that this work is not always quick or easy, and it puts the effort where the return actually is. The low-risk pilot that saves a sliver of cost proves little and returns less.

The through-line of Meghezzi's case is that the customer experience function has been misfiled. Treated as a cost center, it invites AI strategies that optimize for savings but damage the revenue underneath. Treated as the revenue engine it actually is, it invites a completely different and more valuable set of questions. The companies that ask those questions, and evaluate every AI decision through both the revenue and the cost lens, will capture gains that the cost-cutters never see.