Your CX Stack is About to Start Taking Calls
Plus: Why “agentic” is suddenly everyone’s favorite word (and what to do about it)
To opt-out of receiving DCX AI Today go here and select Decoding Customer Experience in your subscriptions list.
📅 February 13, 2026 | ⏱️ 4-min read
Good Morning!
Today we’re talking about AI that doesn’t just help… it actually does the work, including on phone calls. We’ve also got new numbers on what customers now expect from AI-driven service (spoiler: it’s faster, and they want explanations).
The Executive Hook:
Right now, everyone’s acting like “agentic AI” is magic. It’s not magic. It’s AI that takes actions, not just answers questions. That can be awesome for customers, or a total mess. The big problem is this: speed is great, but if you lose control, you lose trust. And trust is the only thing your customers remember.
🧠 THE DEEP DIVE: Zendesk Puts Agentic AI On The Phone
The Big Picture: Zendesk is letting some customers try voice AI agents that can handle phone calls and pass the full context to a human when needed.
What’s happening:
Zendesk is pushing its AI agents beyond chat and email, into live phone calls where customers talk normally.
The voice agent can handle the basics end-to-end: greet the customer, figure out what they want, help solve it, then wrap up the call.
If the AI needs to hand off to a human, it can send the transcript and summary so the agent doesn’t start with “Can you repeat that?”
Why it matters: Phone calls are where emotions show up. If your AI sounds confident but gets it wrong, customers won’t say “the AI failed.” They’ll say “your company failed.”
The takeaway: Start small. Pick a few simple call reasons. Set “stop rules” for anything sensitive (payments, identity checks, medical, legal, angry customers). Track the handoff quality, not just how many calls the bot avoided.
Source: Zendesk Support
📊 CX BY THE NUMBERS: Customers Want Faster Service, And They Want You To Explain Your AI
Data Source: Zendesk CX Trends 2026
74% of consumers say AI makes them expect customer service 24/7. Meaning: waiting until “Monday morning” now feels old.
88% of customers expect faster replies than they did last year. Meaning: your “normal response time” might feel slow already.
95% of consumers expect an explanation when AI makes a decision. Meaning: “the system says no” is not a good answer anymore.
The Insight: Customers want speed and clarity at the same time. If you only go fast, customers will assume you’re being shady when the AI makes a call they don’t like.
🧰 THE AI TOOLBOX: Medallia Insights Assistant And Smart Topic Builder
The Tool: Medallia is adding new GenAI features to help teams spot customer problems faster and actually do something about them.
What it does:
It lets people ask simple questions about customer feedback (like “What are customers mad about this week?”) and helps find themes automatically, with humans reviewing the results.
CX Use Case:
Catch problems sooner: If login failures spike or billing confusion pops up, teams can see it faster and jump on it before it blows up.
Turn “insight” into action: Finding a theme is nice. Fixing the theme is what matters. Use this to assign owners and track progress, not just make reports.
Trust:
AI can “find patterns,” but it can also get things wrong, or overreact to noisy data. Put rules in place: who can publish themes, who can change labels, and how you confirm it’s real. Also watch for bias. If the AI labels certain customers as “difficult,” that’s a brand risk waiting to happen.
Source: Medallia
⚡ SPEED ROUND: Quick Hits
Bellagent Launches AI Agent Platform to Remove Barriers to Enterprise AI — More proof the market is moving from “chatbots” to “AI that actually completes tasks.”
Proofpoint Acquires Acuvity to Deliver AI Security and Governance Across the Agentic Workspace — Security teams are trying to put seatbelts on AI agents before things get weird.
Genesys Advances Virtual Agents to Task Completion — Virtual agents aren’t being judged on how “human” they sound. They’re being judged on whether they finish the job.
📡 THE SIGNAL: Trust Is The Feature Now
Agentic AI is going to make it super easy to chase quick wins: fewer calls, fewer tickets, lower costs. Cool. But the brands customers stick with will be the ones that build AI with brakes, not just an engine. Clear handoffs. Clear explanations. Clear “nope, a human should handle this” rules. Because customers don’t care how advanced your AI is. They care if your company feels responsible when something goes wrong.
So here’s the question for today: are you building AI to move faster, or to serve better when it counts?
See you Monday.
👥 Share This Issue
If this issue sharpened your thinking about AI in CX, share it with a colleague in customer service, digital operations, or transformation. Alignment builds advantage.
📬 Feedback & Ideas
What’s the biggest AI friction point inside your CX organization right now? Reply in one sentence — I’ll pull real-world examples into future issues.