How CX Teams Can Keep AI Fatigue at Bay
A 2025 MIT report ‘The GenAI Divide: State of AI in Business 2025’ had some telling findings. In spite of enterprise investments of more than $30 to 40 billion in generative AI, just about 5 percent of pilot programs delivered revenue acceleration and positive P&L impact. And Gartner predictions foresee the cancellation of more than 40 percent of agentic AI projects by 2027, driven by cost escalations and opacity of business value.
Contributing significantly to this challenge is the “AI fatigue” factor. Many leaders have reported a waning enthusiasm for AI adoption. Complexity of AI systems, constant changes, and ethical concerns are leading to burnout and fatigue. Customer experience (CX), although more vulnerable to the risk of AI fatigue, seems to have pointers to beat this trend. As an active practitioner working inside enterprise CRM and CX transformations, here are my perspectives on the AI fatigue traps for CX teams and the top three disciplines that organizations can practice to demonstrate the clear value and ROI of genAI.
What AI Fatigue Can Look Like in a CX Team
The first reality is “transformation fatigue.” Often, CX teams that have adapted to one AI platform are requested to pivot again. While models change rapidly, customer context moves too slowly between them—or not at all. This leads to adoption delays, and a resigned return to manual processes. As one CTO put it, “By the time we finish procurement, there’s already a better model available.”
And then we have the “staffing trap.” In their haste to optimize costs, organizations cut staff before automation is stabilized in the organization. This overwhelms the remaining team members, leading to slower rollouts.
From the customer perspective, a trust deficit can set in when context cannot be efficiently shared between systems. A PwC 2025 CX survey showed that 58 percent of consumers were only somewhat, or not at all, comfortable using AI to engage with brands.
And herein lies a profound truth about genAI in CX. Trust is the primary casualty of AI fatigue, both with customers and employees. Remember, people use technology to make things easy and bring value. This can be achieved only through purposeful integration of systems and data across all functions of marketing, sales, commerce, operations, etc.
3 Disciplines for GenAI Outperformance in CX
It all starts with clarity of purpose with respect to the customer journey. An AI-led CX transformation model should holistically connect the core touchpoints of discovery, evaluation, purchase, use, and renewal. This clarity should then percolate to processes. It is pertinent to note that AI's earliest ROI in CX came from structured, high-volume workflows (such as intake routing, post-call summaries, and CSAT prediction), and not from replacing humans.
The required discipline to achieve this end lies in mapping friction points before choosing tools. This will reveal the interactions that have clean data, repeatable logic, and measurable outcomes. Selection of platforms, tools, and vendors can be effective only when this is accomplished. Organizations that begin with workflow clarity will be the ones that successfully move from pilot to production.
“Human-in-the-loop” governance must start from day one. Zendesk’s 2026 Customer Experience (CX) Trends report underscores the critical importance of contextual intelligence. Its research shows that 85 percent of CX leaders feel customers will drop brands over unresolved issues. How well an organization can combine AI, data, and human judgement in real time will be the differentiator between good and great customer experience. In fact, the human-in-the-loop will soon evolve to human-above-the-loop where AI agents will perform most, if not all, of the core process—and the role of the human will rest on judgment-intensive decisions.
The required discipline is to define escalation triggers before launching, not after the first customer complaint. Most importantly, the human-machine handoff must be smooth, seamless, and invisible to the customer. At the same time, they must be empowered to override AI’s actions.
Metrics must shift from operational outputs to customer experience outcomes and engagement. An Everest Group study predicts that CX ROI will become a board-level metric. How did AI and automation-led CX initiatives raise revenues, reduce churn, and improve cost-to-serve? Traditional metrics such as NPS and CSAT will be framed as leading indicators, and not as end outcomes.
The discipline will be to build CX ROI metrics that will show traceable links between experience improvements and financial value creation. Remember, if you can't show how AI changed a customer experience, you can't defend the budget when fatigue sets in.
Combined intelligently, these three disciplines will make genAI deployments stick because they treat AI as operational infrastructure and not as a one-time feature sprint or a cost-cutting exercise. Ownership, governance, measurement and continuous improvement must be built in from the start—not bolted on at the go-live stage.
One thing is clear. AI fatigue is not a technology problem but a strategy imperative. As the ones positioned closest to customer outcomes, CX leaders need to ask three important questions to address this challenge. One, where does the process break down before the technology is even involved? Two, who owns the escalation when the AI gets it wrong? And three, how will you show what it did to customer loyalty? Enterprises that answer them right will infuse the energy to dispel the fatigue.
Muthuselvan Renganathan is global head of CX at Mastek, with expertise in customer experience transformation, CRM, digital engagement, and AI-led innovation. He helps organizations build connected, scalable CX strategies that improve service, loyalty, and business performance. At Mastek, he contributes to experience-led transformation as part of the company’s broader digital engineering, AI, and cloud capabilities.