The unspoken truth about AI in CX: Bad experiences are baked into the model
Charlie Besecker, CEO of Gladly, explains why deflection-driven AI leads to bad experiences—and why measuring resolutions and assists creates efficiency with lasting customer trust.
When a major tech CEO boasted that AI let the company slash thousands of support jobs, reactions were split. Some applauded the efficiency gains. Others pointed to companies that over-automated, only to backpedal as customer frustration forced them to reassign employees into service.
The uncomfortable truth: Measuring AI by short-term savings alone isn’t progress. It’s financial engineering disguised as innovation. What looks like efficiency often becomes long-term revenue leaks when customer experience deteriorates. Quick wins may grab headlines, but they don’t build sustainable businesses or create the competitive moats that drive lasting growth.
The customer experience crisis demands a different approach
Customer experience (CX) is deteriorating at an alarming pace. Forrester’s 2025 Customer Experience Index found CX at its lowest point ever, with 25% of brands declining and only 7% improving. Meanwhile, PwC research shows 59% of U.S. consumers will walk away after several bad experiences, and 17% after just one.
When cost-cutting undermines service—whether through headcount cuts, reduced hours, fewer channels, or AI applied as blunt-force deflection—customers wait longer, struggle to get answers, and lose the human touch that builds lasting relationships.
The real metric separating leaders from laggards is outcomes velocity: how quickly customers get what they came for. Don’t just track how fast an agent responds; track how fast a customer’s problem is solved. Don’t measure interaction time; measure time to resolution.
Smarter conversations, measurable outcomes
The companies that thrive with AI create conversations enriched with context and empathy, measured by two key outcomes that drive real customer value:
- Resolutions: AI fully solves an issue end-to-end, eliminating the need for escalation while boosting satisfaction. These complete solutions free human teams to focus on complex, high-value conversations that require creativity and emotional intelligence.
- Assists: AI accelerates human agents by gathering context, conducting initial troubleshooting, and preparing insights before handoff. When customers want to talk to people, AI ensures those conversations are better. AI becomes an intelligent copilot, helping agents work faster while ensuring customers feel heard and known.
This distinction matters because smarter conversations combine human qualities (recognition, empathy, continuity) with measurable results. When both are present, customers get what they need faster, agents feel empowered with better tools and information, and businesses see stronger loyalty across the board.
When incentives sabotage experience
“Outcomes-based pricing” sounds customer-centric, but in practice, it usually means vendors only get paid when AI “resolves.” The unspoken truth is that this rewards deflection, not customer solutions. Revenue tied solely to resolution drives AI to block escalations, trap customers in loops, and leave agents without context. Said differently, poor CX isn’t a bug; it’s a feature.
Gladly takes a different stance. Outcomes should include resolution and assistance. Our model values both, rewarding smooth handoffs with rich context and creating a feedback loop where every assist trains the AI to resolve more in the future. Optimal CX means both are valued, designed for, and priced in. That is how you align incentives with great service and deliver results that last.
Proven AI results across industries
Gladly customers are proving the impact of outcomes-based AI every day. Outdoor brand KÜHL achieved a 59% resolution rate with AI while seeing a 120% increase in revenue per conversation. Breeze Airways now has 71% of conversations assisted by AI, with customer satisfaction scores consistently high. Rothy’s reduced average conversation times by 34%, while Crate & Barrel unlocked more than $10 million in new revenue through chat payments—strong evidence that the right AI approach saves money and drives loyalty.
The leadership imperative
For executives responsible for customer experience, the litmus test is simple: Ask your AI vendors how they make money on handoffs to humans.
If the answer is “we don’t” or “we minimize those,” you’ve found the source of your customer experience problems. Vendors that only profit from resolutions will underinvest in seamless transitions—because every successful handoff costs them money.
The truth is that most AI pricing models make bad customer experiences inevitable. When deflection becomes the business model, customer frustration becomes the predictable result.
Discover how outcomes-based AI creates proven impact at gladly.ai.
Note: This article was created by Gladly.