Containment is dead: Five takeaways from the AI ROI in Contact Center Summit
SiliconANGLE Zeus Kerravala
Contact center vendors want to kill “containment” and “deflection.” They say the real win is resolving problems across systems, not trapping callers.
Based on reporting by SiliconANGLE, Zeus Kerravala — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
At theCUBE Research’s AI ROI in Contact Center Summit, the sharpest argument wasn’t about a new product. It was about throwing out the old scorecard. Vendors from Cisco, Five9, Zoom, Talkdesk and an implementation partner spent the event pushing back on the way contact centers have measured AI over the last two years.
The big shift is away from treating AI as a fancy call-flow tool. Vinod Muthukrishnan of Cisco said “agentic” is being used too loosely, and that a real AI agent is more than a slot-filler that completes a transaction. Pedro Andrade of Talkdesk made a similar point, calling CXA an operating model shift rather than a technology purchase. Five9’s Amit Mathradas drew the line even more bluntly: the older wave was basically a decision tree with voice on top; the newer one is probabilistic, which changes what the system can safely do.
That change also breaks the old obsession with containment. Cisco said nobody wants to be “contained” or “deflected,” and Zoom’s Ram Rajagopalan said a frustrated caller who drops off cannot count as a resolved interaction. Talkdesk backed that up with survey data: its CXA leaders reported about 22% net promoter score gains, compared with 5% for the next tier, while cost per contact improved by 57% versus 48%. Andrade’s point was simple: savings alone misses the bigger value.
The more interesting metric is end-to-end resolution. Andrade said the clock should run from the opening of a problem until it’s closed, back office included. That matters because the handoff gap is still everywhere. Talkdesk said 98% of companies have deployed AI somewhere in the customer journey, but only 15% combine agentic AI with cross-department orchestration.
And that gap is not about model quality. It’s about enterprise readiness: compliance, security, disconnected systems and legacy infrastructure. The summit’s message was that the technology is moving faster than the org chart, which is usually how these things go wrong. The winning teams won’t be the ones automating the most calls. They’ll be the ones that connect the systems, keep humans where judgment matters, and measure whether the customer actually got help.
My take — AI-written commentary, not fact-checked reporting
The industry is finally admitting that “contained” is a terrible compliment. Contact centers don’t need more decorative automation; they need fewer handoffs, better routing, and a human escape hatch that actually works. The funny part is that the expensive part was never the model — it was the organization pretending the spreadsheet counts as orchestration.
Read more about this at: SiliconANGLE