The AI Race Has a Customer Experience Problem

RocketPhone.ai CEO and former Salesforce Director Muj Choudhry discusses the limits of AI-led CX, from fragmented data and rushed deployments to the people still needed to make automation work

AI & Automation in CXInterview

Published: August 10, 2026

Rhys Fisher

Enterprise AI investment is under greater scrutiny, with Salesforce among the vendors facing questions over how quickly ambitious AI strategies can translate into dependable commercial results.

Speaking with CX Today Associate Editor Rhys Fisher, Muj Choudhry, CEO of RocketPhone.ai and former Salesforce Director, said the challenge is not simply choosing the right AI platform. Organizations must first confront disconnected data, cumbersome processes, and the operational realities behind customer service.

Choudhry pointed to Salesforce’s long acquisition history as a factor complicating its path forward. The company has brought data, integration, analytics, and AI assets into its portfolio, but joining them into a coherent customer and employee experience is another matter.

For CX teams, the more immediate issue is readiness. Customer data often sits across CRM platforms, contact centers, field-service tools, and phone conversations. Without that context, AI can only provide a generic answer.

“AI isn’t everything. It’s never the panacea that’s going to cure all your ills,” Choudhry said. “You’ve got to just go right back to basics and say, right, what is it that we’re trying to improve?”

He shared an example involving a boiler manufacturer. By analyzing calls between customers and field engineers in real time, AI can flag a likely fault before an appointment, helping ensure the engineer arrives with the right replacement part. The result is fewer wasted visits and a better customer outcome.

However, Choudhry warned against treating AI as a shortcut to lower service costs. Token prices may rise, while poorly governed bots can create compliance, safety, and customer trust risks.

AI can support coaching by detecting risky language or identifying improvement areas, he said. But experienced managers remain essential when nuance, empathy, and judgment are required.

AI check: This was reviewed against the requested test. The initial draft was tightened to remove repetitive phrasing, formulaic contrasts, and generic AI language.

Agentic AIAgentic AI in Customer Service​Artificial IntelligenceCall & Contact Center SoftwareCCaaSCRM
Featured

Share This Post