A New Mindset for Customer Experience in the Age of AI

A prediction for 2026 strikes home for me. According to Sharyn Leaver, CRO at Forrester, 

“In 2026, a third of companies will harm experiences with frustrating AI self-service. The pressure to cut costs will cause companies to deploy customer-facing genAI chatbots and virtual agents prematurely—and in contexts where they're unlikely to succeed.”

This prediction speaks to the challenge of how we’re applying GenAI and suggests a path towards success. I’m an engineer, so I decompose things… let’s decompose this prediction. 

The prediction highlights four key mindset mistakes companies are making. Imagine an alternative mindset.

Shift from “self-service” deflection to real issue resolution

With any new technology, people first see a way to make the old way better. When the iPhone first appeared, the most exciting feature was visual voicemail, not the ability to change transportation (e.g., Uber).

The same is true with generative AI. Ooooh! Now, we can turbocharge self-service by dropping our knowledge base into genAI and get friendlier answers with better search.

The mistake is thinking that the problem with “self-service” is a user interface problem. Self-service frustrates because service can be really complex and unlikely to boil down to a few faq’s or knowledge articles. Moreover, what makes service complex is often out of your hands – it’s the customer’s own situation or downstream elements of your service that go beyond a simple update.

GenAI is amazing - let’s ask more of it. Instead of “harm experiences with frustrating AI self-service” what if we “deliver great experiences with AI that resolve issues.”

Let’s set the goal to get AI to actually resolve every customer question, issue, or change.

Shift from cut costs to lift loyalty

All the AI hype is around using AI to replace humans and drive down labor costs.

What a missed opportunity. If we’ve learned anything in the last two decades, it’s that customer loyalty drives market share and valuation, not cost-cutting. Has everyone forgotten the power of NPS?

Let’s reframe the value of AI. Instead of “pressure to cut costs will cause companies to deploy…” how about “pressure to lift customer loyalty will cause companies to deploy …”

This puts the customer at the center of why you deploy AI. What if instead of focusing on cost-cutting through enhanced deflection, you focused on loyalty through customer-led journeys?

Imagine a 3rd way - AI Customer Advocacy

The magic of generative AI now seems to come in only two flavors: chatbots that tap into knowledge and virtual agents that automate tasks.

What happens when the knowledge the customer needs isn’t available or the task the customer requires isn’t automated? Either hallucination (make up the answer) or escalation (send the frustrated customer to a human agent).

Both are scary alternatives that lead companies to either a) postpone deployment of AI until they have organized all their knowledge and automated all their tasks or b) narrow the deployment to simple self-service contexts that they’re sure AI can handle.

AI Customer Advocacy is a third flavor: AI that engages your team when it doesn’t know the answer or can’t automate the task. AI that uses its understanding of your organization to orchestrate humans in the flow.

AI advocates can deploy on day one because if they don’t know, they know who to ask.

Let’s imagine a new way to deploy AI. Instead of “customer-facing genAI chatbots and virtual agents prematurely,” we can deploy “customer-facing genAI advocates ready on day one.”

Serve in any context

To truly lift customer loyalty with customer-facing AI, you must be ready to handle anything the customer throws at you. If you can’t, you’re back to frustrating self-service.

With AI customer advocacy, AI can now resolve every issue because your AI either knows the answer or how to automate the task, or your AI knows who to ask. Now, AI can orchestrate your organization to follow your customer’s lead on their journey.

Instead of “in contexts where they’re unlikely to succeed”, we deliver AI “in any context the customer needs.”

Have your cake and eat it too

Generative AI is a once-in-a-generation technology that deserves its hype. The real limitations of genAI are limitations of imagination.

With a simple reframing, the Forrester prediction can show the way.

Instead of “... companies will harm experiences with frustrating AI self-service. The pressure to cut costs will cause companies to deploy customer-facing genAI chatbots and virtual agents prematurely—and in contexts where they're unlikely to succeed.”

What if “... companies will deliver great experiences with AI that resolve issues. The pressure to lift customer loyalty will cause companies to deploy customer-facing genAI advocates ready on day one—and in any context the customer needs.”

It’s your choice. What kind of company do you want to be?