AI is transforming customer engagement—but most businesses are still missing the mark. In fact, according to MIT, nearly 95% of enterprise AI deployments fail to deliver the outcomes they promise.
The reason?
Too many AI initiatives focus on automation and efficiency rather than what truly drives success: customer trust and delight.
In a recent webinar, “Beyond the Zero Button: Building AI Agents That Customers Love,” Cassie Kozyrkov (CEO of COSR and former Chief Decision Scientist at Google) and Cresta’s Devon Michal (VP of Product Marketing) unpacked what it really takes to build AI systems that both customers and businesses love.
Here’s what every enterprise leader should take away from their conversation.
AI Should Serve Humans — Not Replace Them
AI should enhance human connection, not erase it. Cassie challenged companies to move beyond efficiency metrics like call deflection or containment and instead ask: “Does this experience make our customer feel understood?”
AI agents that feel dismissive or overly scripted don’t just fail, they damage trust. Instead, successful systems are built around empathy, transparency, and designing for delight. The best AI doesn’t sound robotic, it feels helpful and seamless.
Redefine What “Success” Looks Like in AI Deployments
Too many enterprises still measure AI performance with outdated KPIs. As Devon emphasized, automation rates alone don’t measure success. Customer satisfaction does.
Real innovation comes when organizations move from containment metrics to experience metrics, understanding not just whether the issue was resolved, but how the customer felt about it. That shift requires real-time data, feedback loops, and ongoing learning from every interaction.
Read our latest Benchmarking Guide for an in-depth look at how to approach measurement in 2025
Learn (and Laugh) From Mistakes
Cassie shared the now infamous story of a dealership chatbot accidentally selling a car for $1, a funny but truly powerful reminder that AI must be built with guardrails, observability, and constant iteration.
In enterprise environments, those safeguards aren’t optional. They’re foundational. Whether it’s hallucination handling or escalation logic, resilient AI systems are those that evolve with both data and context.
Empower Customers Through Personalization
The future of AI in customer service isn’t about replacing people—it’s about giving customers control and personalized experiences that feel natural.
Cassie described a near future where AI agents act as trusted copilots, helping customers make informed decisions quickly while maintaining human oversight. The most successful companies will be those that strike the balance between autonomy and accountability.
The Takeaway for Enterprise Leaders
The message from the webinar was clear:
It’s time to stop automating for efficiency and start designing for empathy.
The next generation of AI agents will be measured not by how fast they resolve a ticket, but by how deeply they connect with the customer on the other side of it.
This session goes beyond buzzwords and dives deep into the real-world strategies that separate AI that frustrates from AI that builds loyalty.
Watch the full webinar recording here to learn how to:
- Identify the right success metrics for AI in customer experience
- Design empathetic, customer-first interactions
- Build AI systems that earn trust—at scale