[00:00:00] 

Everybody keeps saying AI is going to save customer experience. It's gonna make things faster, it's gonna make things smarter, more personalized, less frustrating, like the Holy Grail of cx. Right? Wrong, because here's the thing, AI isn't fixing cx and actually in a lot of cases it's making things worse.

And I know that sounds crazy, but hang with me. 'cause today we're gonna break down the top five reasons that AI is hurting CX more than it's helping. And this episode is really for three different types of people. First the tech buyers who are trying to figure out if this AI powered tool is actually worth buying.

You tech advisors out there who are trying to help clients avoid stepping into a big steaming pile of buzzword Bs. And of course, real customers who are like, why do I keep getting stuck in a loop with a robot that doesn't understand me? 

Alright, let's go lie number one, [00:01:00] AI makes support faster. I mean, yeah, it, it can, but.

It usually doesn't. 'cause you know what? AI often makes support slower. Yeah, like way slower. Let me paint a picture. You call into a business because you have a very, very simple question, and then you get stuck in an AI voice loop that sounds like Siri had a stroke. Please tell me in a few words what you're calling about billing issues.

I think you said killing tissues. Is that right? No. No, it's not. It's not right. So you try again, you get nowhere and eventually you yell representative or some four letter word like you're being held hostage by a smart toaster, and by the time you get to a real life human being, you're already pissed.

The rep has to calm you down before they can even help solve your problem, and AI didn't speed things up, right? What AI did here is add [00:02:00] friction. Now, when AI is done right, I mean sure it, it can route calls faster, it can surface relevant info, it can trigger workflows, but that only works when the AI actually understands natural language.

It needs the data behind the scenes to make sure it's clean and the humans still need to have the power to override. And by the way, most companies don't have that. They bought some plug and play AI tool, they slapped it on a bunch of mess and hoped it would actually fix things magically, but spoiler alert it didn't.



Line number two, AI makes things feel more personalized. All right. This one's my favorite kinda lie because this is the marketing deck kind. You have the vendor, they walk in and they say, with our AI powered recommendation engine, every customer gets a uniquely tailored journey. Sounds awesome, right? But okay.

Then what happens? You get an email that says, hi, first name. Based on [00:03:00] your recent activity and the fact you went to college at a BC University, we think you'd love. Nothing related to what you actually want. And God forbid you bought something as a gift for someone else. Now your whole experience is screwed up.

You bought dog food one time. Guess what buddy? You're a dog owner forever. Now, AI isn't personalizing. It's pattern matching. It doesn't know context, it doesn't understand intent, it doesn't think like you and I do. Unless your customer journey is super repetitive, that AI is gonna do more harm than good, you want to actually feel personalized.

Then let humans train the AI with real world exceptions in mind. Build logic that accounts for the messy middle, and then test it. Test it, please with actual users who can tell you, yeah, this feels like me. Line number three, AI reduces the need for human beings. Let's get one thing [00:04:00] straight. 

AI is not replacing your support team.

It's replacing the first 30 seconds of a conversation and that's it. But what happens when leadership believes this hype? Well, they downsize, they cut support in half, and then you're sitting there with a gold mine. Ai, fuel fueled efficiency, right? Well, that's what they think. And the reality is what they've got is burned out agents.

Angry customers, lots of bad reviews, all because some exec saw a demo somewhere where a bot could reset a password. AI is a force multiplier. It doesn't eliminate the need for real life humans, it just helps them focus on more valuable stuff like, I don't know, actually solving tough problems, escalating the weird use cases, talking to the angry customer who don't trust bots to begin with, but when you cut humans out too soon.

You're creating a support system that looks good in theory, but it breaks in [00:05:00] reality. 

Lie number four, AI will give you better insights. Data is not insight and ai, it's not magic. Most of these AI power dashboards, they're just automated reports at best. They're glorified Excel spreadsheets with fancy ui, and you know what doesn't help?

Dashboards that show sentiment scores, but don't explain why. Charts that track customer satisfaction scores, but not how to improve it. Summaries that, highlight calls, but mis nuance. You need a human brain to look at the data and say, Hey, 83% of people are complaining about the same thing. Maybe we should fix that.

AI can help surface those patterns faster. Sure. But if your team isn't trained on how to use that info, it just sits there useless. So if you're a tech buyer, ask questions like, well, hey, what kind of reports does the AI [00:06:00] actually give me? And then who interprets that? And what actions do we take from these reports?

And if you can't get answers to these three questions, that means that AI isn't giving insight. It's just giving noise. Line number five, 

AI improves the overarching customer experience. And here's the brutal truth, no BS 'cause of course this is CX. Without the bs, nobody cares how fancy your tech is if the customer still has to walk away.

Pissed CX is about how people feel. AI is about logic and language and large data sets. So what happens? AI tries to automate empathy. And it sucks at it. So what you get is fake apologies. You get this weird, monotonous robotic tone, you get this cold formulaic response, and all the while, what customers actually want is someone who listens, someone who will take ownership, someone who can [00:07:00] empathize, you know, a real connection you can't automate, or AI that you can support it.

You can augment it. You can streamline the background stuff, but real CX still lives and dies by real authentic human moments. If your AI is just there to deflect calls, that's not improving cx, that's protecting your budget, and customers can smell that from a mile away. So what do you do instead? 

Well, let's wrap this with something useful.

If you want AI to actually help cx, here's what you do. Number one, use AI to support, not replace. Let it handle the repetitive BS and let it tee up info, but never pretend it's the star of the show. Number two, test with real customers. Don't deploy AI in a vacuum. Run pilots collect feedback, adjust fast.

Third, teach it with real world messes. [00:08:00] So give it that weird edge case. Train it on nuance. Teach it exceptions to rules. Fourth, always, always leave a human escape hatch. Every AI interaction should have a clear path to a real life human. Zero exceptions, none. And finally, number five, track outcomes, not just activity.

Measure real impact are tickets resolving faster? Our customers happier Are agents less stressed? If not, your AI is failing. 

So there you go. AI isn't evil, but it's not magic either. It's a tool, another tool in the tool belt and tools only work when the people using them know what the hell they're doing.

Use AI to simplify and amplify your CX strategy, not replace it. And if you do that. You'll actually stand out because most companies are out there playing this buzzword bingo, while their customers are still screaming representative [00:09:00] or F word, F word. F word into the phone. Alright folks, that's it for today.

If this episode hit home with you, please do me a favor, share it with your friends in the CX space, share it with a client, share it with your boss who thinks that chat GPT is the future of support. And as for me, your host, Brian Nichols. I will talk to you next week on next week's episode of CX without the bs.

We'll see you then.