A warning from Gartner that generative AI (GenAI) resolution costs could exceed $3 per interaction by 2030 has sent ripples through the customer experience and contact centre industry. The research house suggested that, within the decade, automated AI resolutions could become more expensive than offshore human agents — a striking reversal of the cost-saving narrative that has underpinned much of the recent investment in large language models.
Yet according to Alasdair Miller, Head of Development and R&D at Sabio Group, the prediction risks drawing the wrong conclusion. The problem, he argues, is not the technology itself but the commercial structures that surround it.
“Gartner isn’t predicting a technology failure. They’re predicting a partnership failure,” Miller writes, responding to the forecast.
In his view, the research reflects long-standing flaws in the way enterprise technology is sold and monetised. Vendors, he says, continue to push consumption-based pricing models while underestimating how quickly usage — and costs — can escalate once AI tools are embedded in day-to-day operations. “Vendors flogging consumption based pricing whilst organisations ‘consume a lot more than they expect.’ Partners promising that AI will ‘make sense of your mess’ when it patently won’t. Technology providers taking their fee regardless of whether their solution actually delivers a single automated resolution.”
This dynamic, Miller argues, predates GenAI. The customer experience sector, he says, has long relied on a familiar cycle: ambitious sales promises, rapid deployment, usage-linked billing and limited accountability once systems are live. “This isn’t a GenAI problem. Instead, it’s a commercial model problem, and it’s been baked into the CX technology industry since long before large language models (LLMs) entered the chat.”
That approach, he adds, leaves organisations exposed when costs rise or automation rates fall short. “The traditional playbook is simple: sell the dream, implement the technology, invoice for consumption and then move on. If costs spiral? That’s the client’s problem. If automation rates disappoint? Well, perhaps you need more consultancy days.”
Gartner’s research highlights several drivers of rising AI expenditure, including the scarcity of specialist talent, unpredictable demand and growing infrastructure requirements. But Miller believes these factors only become damaging when suppliers have no financial incentive to control them. “When a vendor’s revenue grows as your AI consumption grows, they have precisely zero incentive to optimise your solution. When their fee arrives regardless of outcomes, why would they care if your automation rate hits 30% or 3%?”
He describes this as an “accountability gap”, where commercial risk sits almost entirely with the client. “The entire model is designed to transfer risk onto the client whilst the partner counts their consumption fees.”
Sabio, he says, has attempted to address this by adopting outcome-based commercial structures. In a recent engagement with a large telecommunications provider, the firm agreed to a “risk reward” arrangement in which payment is linked directly to successful customer interactions rather than raw system usage. “We earn a fee per completed interaction – with additional incentives only when resolutions stick and don’t generate repeat contacts.”
The implications of that model are stark. “In other words: if our AI doesn’t work, we don’t get paid!” Miller writes. He is clear that the approach is not philanthropic but practical. “This isn’t altruism; it’s accountability.”
By tying revenue to results, he argues, suppliers are forced to focus on efficient implementation, disciplined use of AI resources and genuine automation. “When your commercial success depends on your client’s operational success, you suddenly become very interested in getting the implementation right, optimising token usage, and ensuring the technology actually resolves customer queries.”
Miller’s conclusion reframes Gartner’s forecast rather than rejecting it outright. “So here’s my counter prediction to Gartner: AI costs will indeed soar for organisations who continue choosing partners with misaligned incentives.” Those costs, he adds, will “spiral for those seduced by consumption based pricing that sounds cheap until the monthly invoice arrives” and “explode for anyone who believes a technology vendor when they promise AI will magically sort their data chaos.”
For organisations willing to demand accountability, however, the outlook is different. “But for organisations who demand outcome based accountability from their partners? The economics look rather different.”
As businesses reassess their AI strategies amid tightening budgets and rising scrutiny, the debate may prove less about algorithms and more about alignment. “The future of AI in customer service isn’t about choosing between expensive automation and cheap offshore agents,” Miller concludes. “It’s about choosing partners who are willing to put their money where their mouth is.”
“Friendly zombie fanatic. Analyst. Coffee buff. Professional music specialist. Communicator.”
