November 17, 20264 min read

Waiting until you’re “ready” for AI? Brian Blackader on why slow adoption is the fastest way to fail

Person with long dark hair and large glasses, wearing a black turtleneck, smiles softly against a light background.

Alexa Krzyzanowski

Senior Content Marketing Manager @Parloa

Better models are coming. Platforms will mature. Once AI is finally ready, you'll start.

Brian Blackader hears versions of that plan all the time. You could wait, he says, "but the world will have moved on by then."

Brian, a partner at McKinsey & Company with two decades in operations and service, joins Parloa CMO Latané Conant in episode 6 of The Agentic Wave. Drawing on McKinsey research, he explains what separates companies realizing value from AI in customer service from those stuck in experiment mode.

Why waiting for AI to be ready backfires

AI technology will never be finished, Brian says, but it’s already good enough. Waiting for it to be ready is one way AI programs stall.

The other is the opposite extreme, which Brian calls “actionism”: 30 to 50 pilots and a “zoo of experiments” that never add up to impact.

Brian recommends a focused middle path. Start with a business problem. Use intent data, the reasons customers contact you, to find where AI can help most. Choose journeys customers will welcome. And keep a close eye on the economics.

Start with a business problem, not the technology

The companies realizing the most value start with a specific business or customer problem rather than "getting jiggy with AI," as Latané puts it. They redesign workflows instead of automating broken processes.

Redesigning workflows means changing how people work. That kind of change management is where many companies get stuck, Brian says. The ones that succeed put contact center operations, product, data, and engineering in one room, under one leader. They involve frontline employees, who know what customers call about, in workflow design. And they set measurable outcomes from day one.

Brian also stresses continuous improvement. Customers’ reasons for calling change over time, so teams need to keep spotting new patterns and adjusting their AI agents and workflows to match. He calls it a muscle every organization needs to build. That’s where intent data comes in.

Use AI to find out why customers call

Brian explains how technology has changed the work of understanding customer intent. Teams once relied on employee interviews and samples of calls. Now, AI can help them analyze conversations at a much greater scale, giving them stronger evidence for where to focus.

That scale makes it possible to look beyond broad categories from interactive voice response (IVR) menus, like "billing question," to the specificity of the problem a customer needs solved. Brian recommends digging four, five, or six levels deeper.

Brian shares an example from a consumer electronics company that makes home devices. At first glance, its calls looked like what you’d expect: mostly technical problems and billing issues. Batteries didn’t come up in the first three levels of intent at all. But at levels five and six, a battery issue accounted for roughly 20–25% of incoming calls. The finding pointed to two possible responses: address the underlying battery problem to prevent calls, or use AI agents to handle the related support inquiries.

Choosing a use case used to be a bit of a bet. Now, Brian says, "the risk is in the implementation of it, not in did I choose the right case to go after." And implementation depends on how customers respond.

Don’t assume your customers aren’t ready

Many leaders worry that customers won't accept AI agents. Brian's experience suggests that isn’t the case.

At one energy provider, customers submit meter readings quarterly, so everyone calls at once and waits can stretch to 20 minutes. When the company implemented AI agents, cost, resolution speed, waiting time, and customer satisfaction all improved for this task.

When the call gets answered right away and the issue gets resolved, customers don't really care who handles it. They "just want it to be over and done with," Brian says. As Latané points out, customers usually aren't comparing an AI agent to a great human conversation. They're comparing it to a long hold or an IVR menu.

Customers aren't the only factor in whether AI pays off. Cost matters too.

Know what you're spending, and what you could gain

In McKinsey's research, almost half of respondents said AI costs ran higher than expected.

Latané has a name for what often drives those surprise costs: vibe spending, or launching AI experiment after experiment without tracking what each one costs or having a focused plan in place. If you can't account for that spending, Brian warns, when the CFO asks about return on investment, "it'll be very difficult for you to have a good answer."

Even so, Brian expects AI to lower the cost of serving customers over time. He also sees opportunities to build loyalty and generate revenue.

He offers an imagined airline example: when first-class seats are empty, AI agents could personally call loyal customers to offer them an upgrade. Coordinating those calls before departure might previously have been impractical.

That raises a question for CX leaders: what valuable service could you offer customers if you had the capacity to deliver it?

The best time to start is now

Brian's parting advice: "Start now, but be focused." Take a few extra weeks to plan, and don't try to automate everything.

Most companies want more than advice, Brian adds. They want a solution. That's why McKinsey and Parloa have teamed up, pairing McKinsey's business planning expertise with Parloa's AI agent management platform to help enterprises bring agentic AI into their customer service operations.

Watch episode 6 of The Agentic Wave for the full conversation.

Ready to turn conversations into lasting loyalty?

Let's build your next great customer experience.