[00:00:00] Is AI actually making customer experience better, or are we just talking ourselves into believing that because it sounds cool and futuristic. Yeah, that's the question we're answering on today's episode of CX Without the bs. Hi, I'm Brian Nichols. And real question for you. If customers keep saying, Hey, can I please just, you know, talk to a human, why are we doubling down on AI and bots that just don't listen?

And today we're gonna get very real about ai. Human intelligence versus ai, artificial intelligence, and how the fight shows up in he, which I call the human experience because that's the only scoreboard that really matters if the experience gets worse. I don't care how many AI badges are on your slide deck, you're losing.

So let's start with why AI is everywhere. Two words, hype and fear. Hype means attention. 10 years ago it was cloud. Say cloud enough times. And people [00:01:00] nodded like you were a genius. And now that word is ai. Same playbook. And it works because we're all afraid of missing the next wave. Fomo, right? Fear of missing out.

No one wants to look like they're behind. So companies don't ask the right questions. They ask the easy questions, not does this help the customer, but does this make our marketing sound smarter? And then fear kicks in. Your competitor announces something AI powered, and the board says, Hey, where's ours?

And suddenly product teams rush sellers start saying AI in every call. Buyers nod so they don't look clueless. And meanwhile, the poor customer, they just want their bill fixed or their shipment tracked without running a maze. And that's the gap. The gap is where trust dies. So from today's episode, if you remember nothing else, remember that the customer doesn't care how you do it.

They care that you do it well. And if AI helps, awesome. If AI adds friction out it [00:02:00] goes, period. So let's talk about real life. And you've lived this, right? You go to a website, a chat bubble jumps out, all smiley. Hi, I'm Ava, your AI assistant, how can I help? And you type, my bill is wrong. And Ava says, here's our FAQ on billing.

And you say, no, I don't need an FAQ. My bill was wrong and I was double charged. Then Ava comes back like, would you like to upgrade your plan? No, Ava, no. I would like to talk to a real human being who understands English and urgency, and the fact that I'm staring at a number that's twice what it should be.

AI misses three things a real person brings without even trying. First is context. My bill is wrong. Doesn't mean teach me your building policy. It means I feel cheated. Fix it now. Two emotion. Humans hear the heat in your voice. AI sees characters and tries to map them to canned answers. And then [00:03:00] third judgment.

A human can bend a rule, it can wave a fee, it can escalate fast. AI usually can't, and when it tries, it makes messes. Another pain spot is intelligent call routing that makes you press eight buttons before connecting you to the wrong department in the first place. Or a voice spot that hears cancel and then decides you really want to change plans because you really wanna spend more money, right?

No, that's not intelligence, that's manipulation. Dressed up with fancy buzzwords. So here's a simple test. If your AI creates more work for your customers than it removes. That's not intelligence, that's expensive friction. Now let's do the flip side. Hi. Human intelligence, and I don't mean PhDs. I mean the everyday intelligence that comes from listening, from noticing, from solving real problems.

The intelligence that lives in a veteran support rep who has heard every weird edge case [00:04:00] and still shows up calm, cool, and collected. The intelligence that lives in a technology advisor who can translate a business owner's messy description of phones are a disaster. People keep missing calls into actual, you know, plans that reduce chaos.

Human intelligence does a few things machines struggle with. First, it can read the room. Tone of voice matters, hesitation matters. A sigh means something different than a pause. It fuses facts with context. Yes. The ticket says billing, but the human hears stress and says, time out. Let's fix the money first and then figure out the root cause.

It exercises. Judgment rules are guidelines. Humans can make exceptions to rules that keep customers loyal for years, and the customer can feel that. They feel it when in the first 10 seconds they hear a person say, Hey, I got you. Customers don't need perfect. What they want and need is ownership. And that is the essence of hi that [00:05:00] human intelligence, specifically in cx, taking ownership about the outcome.

And no, I'm not here to dunk on AI and pretend it's useless 'cause it's not. There are places absolutely where AI is the legit superpower when it's placed in the right role, transcription and summarization, turning a 45 minute support call into a searchable summary. That's actually helpful. Signal from noise flagging trends across thousands of tickets.

Billing complaints spiked after the last invoice cycle. That's helpful. Proactive alerts. Let's spot a shipping delay before the customer notices, and then we can go ahead and queue for a real human to reach out. That's helpful. How about assisted workflows, suggesting next steps to human reps who still make the final call?

Again, that's helpful. Notice a pattern here. AI is assistance not authority. AI is a copilot, not the captain. The human [00:06:00] retains judgment. The human is accountable. The tool helps the person move faster with better information, and that's the right balance. AI leads. AI supports that human experience. It improves.

And all this ladders up to the same place again. E what I said earlier, that human experience, this is the scoreboard. This is the win-loss column. Not AI badges, not demo sizzles and and sparks. This is the experience. So ask three questions after every interaction, did we reduce effort or do we increase it?

Did we leave the customer feeling respected or dismissed? Did trust go up or trust Go down. If effort went down, respect went up, and trust climbed, yeah, great. We're winning. But if the metrics on your dashboard look great, but customers are tweeting, screenshots of your chat bot failing basic questions, you're losing and you're [00:07:00] losing revenue that you can't measure until it's too late.

So here's one case study. A national provider rolled out an AI billing assistant and it handled the top 10 questions. Great. And then number 11. Complete chaos because the edge, the edge case is there. They spiraled. People got stuck in loops. Call volumes went up because customers kept on mashing zero to escape.

Social media lit up with your bot, won't let me talk to a human. And the company claimed deflection success. Yeah. The customers felt deflected and there's a big difference there. Compare that to this other story, A local credit union that actually listened, not with fancy ai. But just a team trained to take ownership.

A member calls frustrated and the rep says, I can see this is urgent. Give me two minutes to look and I'll, I'll stay with you. Fee waived. Confusion cleared, member stays their family joins. Generational loyalty. [00:08:00] That is hi at work. Now let's blend these together. Here's a third case study, an online retailer who uses AI to detect at-risk shipments.

And instead of blasting an auto email that sounds robotic and impersonal, they trigger a human outreach and they say, Hey, quick heads up. We saw your package might be hitting a snag here and here's what we're going to do. Do you want it rerouted? Do you want split? And, and that right there, that's AI feeding hi and the outcome better ag.

So I like simple frameworks. You can repeat them until your team can recite them. So let's look at AI versus AI and how it specifically impacts he. So start every design, every purchase, every playbook. With this question, how will this choice show up in the human experience the he, and then decide the role of the tool?

Will AI assist a human or will it gatekeep a human? If it gate keeps you better, give it a very good reason and a very clean [00:09:00] escape hatch. Measure the outcomes that matter. So effort, respect, trust, and not just ticket volume or handle times. And then print it on a card, put it on your wall, make it your product meeting opener.

And by the way, let's talk to the advisors for a minute. You technology advisors, you are stuck in the middle. Vendors are pushing AI and clients are asking, Hey, do you have ai? Your job is to move the conversation from buzzwords to tangible outcomes. So here's a simple talk track that you can use literally tomorrow.

Hey, look, ai, it's a tool and we use it where it speeds up real work like transcription, analytics, proactive alerts. But your customers don't buy ai. They are buying outcomes. They buy shorter wait times. They buy first contact resolution. They're buying fewer transfers or people who are, I don't know, taking ownership revolutionary.

I know. This is human intelligence leading. [00:10:00] So we designed the stack. So AI supports your team and your customer feels the difference, and then back it up with proof. Offer a before and after storyboard. Hey, here's your current call path. Here's the new path. Two fewer steps. A human is engaged earlier and there are clearer options.

Show a playbook. Snippet. Hey, if sentiment is hot or the topic is billing or cancellation, we can short circuit right to a person. Then share a pilot plan. Hey, let's test this with 10% of the call volume. We can measure and see how it affects effort and trust, and then we can scale. 

See, you're shifting the sale away from features to experience math. I like that executives understand effort and trust. Tie it to renewal and referral. And now you're selling money, not AI modules. So here's a little advisor toolkit. You can literally copy and paste and use any [00:11:00] time. Um, start out with the only scoreboard that matters is human experience, and then have three bullets, right, lower effort, higher respect, stronger trust, and then in your little proof box, you know, outline what it looks like.

Average whole time down, 32% in 30 days, escalations down 18%. NPS comments mentioned. Finally talked to a real human being. So your call to action then would be run a 30 day hi, test first, and if effort doesn't drop, where will we work it for free? That right there, that's a clean pitch and you're selling a result.

Now, let's talk to business owners and IT leaders. You don't need to be a CX nerd, I promise. What you need is a checklist that you can throw at a vendor claiming AI magic, and here's what that five question checklist might look like. First, show me the journey. So when does a human appear? And how fast can we actually have a customer get to them?

Show me the escape hatch. So if a bot fails, what [00:12:00] happens in under 30 seconds? Show me the outcomes, not demos. What changed for the customer effort? The customer time. That customer trust Show me. Control. Can I tune the bot? Can I change thresholds or do I have to open a ticket and wait two weeks? Of course, show me the cost of wrong.

If the AI routes it wrong, who catches it? How fast do they catch it, and what do we say to the customer? And if a vendor can't answer those cleanly, they're selling a science project and they want you to fund it with your reputation. Alright, so let's break a great interaction into five beats that you can hear when it's working.

First is recognition. Oh, I see the issue, not, I'm sorry. You feel that way. Human name's the real thing. Ownership, I've got you. I'm staying on this until it, it's done. Customers can feel that difference. How about clarity? Hey, here's what I'm doing now and here's [00:13:00] what happens next. No mystery, and of course resolution.

The thing gets fixed, not fiddled with. So money is reversed, that shipment is rerouted, the feature is enabled. And of course we wrap up with closure. Hey. Anything else shows up. Here's my direct line. And that's where loyalty starts. See, AI can support steps two and three with data, but it cannot replace steps one, four, or five without risk to trust.

So let's go to another topic here, and these are the five most common AI traps and how we can avoid them. First is the bot is a bouncer. So making the ai, the door guy that refuses entry to humans. No, no bueno. Fix it. Bought as a greeter, not a gatekeeper. How about no sentiment override. Heated language should short circuit to a real person, so you fix that by adding a heat switch or an escape patch.

How about menu mazes? Too many options, not enough exits. [00:14:00] Fix it with fewer choices, clearer language. Press zero for a person that actually works. How about edge case amnesia? Because bots are trained on the top 10 issues, but they're dumb in the next 100. So to fix it, you route anything uncertain to a real life human.

And then I know this is wild, have the AI learn from that interaction. How about metric theater declaring victory because deflection wet up while complaints exploded. Nope. Fix. You measure it with effort, respect, and trust. And if you fix those five things, your AI has a shot at helping the human experience instead of strangling it.

Now, let's talk about some scripted moments that you can copy. Literally word for word here. And I want you to use these lines 'cause they're simple and they work First opening with ownership. Hey, thanks for hanging in. I've read your note. I've got you. Here's what I'm gonna do [00:15:00] First. Next resetting a heated call.

Hey, you're right, I, and I get it, you're frustrated. If I was in your shoes empathetically, I'd feel the same. And here's the path to fix it. And let me close this loop. Hey, I'm staying on this right, and if anything weird pops up, reply to this message. It comes straight to me. That's short. It's human and it's clear.

Alright, Brian, so how do we add ai? Without breaking the he that human experience. Well, there's three rules. Number one, always, always add a human earlier than you think. When in doubt, connect, also, always give a bypass. Give a visible talk to a person that is not fake. Always start small. Pilot with a slice of volume, measure the effort and trust, and then scale after proof.

If you can't follow those rules because of budget or architecture, you don't have an AI [00:16:00] problem, you have a design problem. And let's talk about the money math, right? Because this is why human intelligence pays off. And I hear this all the time. Executives will ask, well, what's the ROI of adding more humans or letting people bypass a bot?

And that's fair. That's a very fair question. And here's the math that you can repeat back. Hey, Mr. Executive, every 1% drop in a customer effort correlates to lower churn and lower churn. That's compounding interest faster. First contact resolutions. They reduce repeat contacts, and that is labor savings right there, or trust.

I know trust drives referrals and reviews, but those reviews, they reduce acquisition cost, and that's real dollars and cents. And of course, the silent killer, not just heart attacks, but reputation. One viral clip of your bot being some smug smart ass costs more than a quarter of headcount savings. So yes, there's a balance, but if you only look at deflection, you're gonna miss the [00:17:00] slow leaks.

Loyalty leaving while you congratulate yourself for answering fewer tickets. Alright, Brian, this is great. So what's the The human intelligence first blueprint? Well, here is what it looks like, and you can actually deploy this this quarter. So phase one roadmap, you're gonna sit there and whiteboard your top five customer journeys, sales, uh, new orders, billing fixes, cancellations, outages, and then you're gonna redline every single friction point and count steps to a real life human.

Number two, you're gonna rewire. We're gonna move a real person into the equation earlier. So if that's billing, cancellations, anything with high heat and the, the, the heat switch if language spikes route to a human in under 30 seconds. And of course, give agents a forgiveness budget that they can use without having to ask permission from a manager.

Phase three assistance. So [00:18:00] add AI where it speed things up without blocking real life people. So that goes back to what we were talking about earlier, summaries, suggestions, alerts, and then send the transcripts to the customer automatically. Hey, Mr. Customer, here's what we did and here's your case number, phase four, measure.

Track the effort. Track respect, track trust. Count transfers, watch, hold times and repeat calls. And then share three stories per week with a rep using judgment to save a customer. And then you celebrate that. And phase five, scale. Roll the working pattern to other journeys. Keep the bypass. Keep that forgiveness.

Budget. Keep the stories. You will feel the culture change in 60 days. People stop hiding behind tools, they start owning outcomes. And that's hi taking the wheel. Now, one thing that's also important is to have language that makes CX simple for [00:19:00] everyone. So let's just say, you know,

talk to people like people, make it easy to reach a human. Fix the problem, not the blame. Explain what's happening and what's happening next, and say, Hey, I've got you and actually mean it. If you do those things, you will outplay fancier systems that do none of that.

So here are some examples of what this could look like in action. I call this one the 92nd save

a small HVAC company put a call, a human now button on their site. During a heatwave, abandoned carts plummeted, bookings went up 14%. Or how about a software startup that gave reps a $50? Make it right. Wallet refund. Fights plummeted online. Reviews mentioned they owned it. Or how about A DTC brand using AI to catch late FedEx scans that had humans send short, honest texts.

Hey, we see the delay, we're on it. Guess what happens? Returns drop, repeat [00:20:00] orders are up. And none of those things required. Fancy AI doing any of the talking. The human did the talking, the tool did the spotting and the experience got better. So. Here's some words that I want you to get rid of from your vocabulary.

So we need to retire due to high call volume. Customers hear that and they hear, we didn't staff, and we don't care per our policy. Ugh, no customers hear. We don't care. We apologize for the inconvenience. Customers here, oh, that's a copy and pasted response. So here's what I want you to start embracing. You are right.

We were wrong. Hey, I can fix this and I got you. Hey, here's what happens next, Mr. Customer. I'm staying with you until this is resolved. Do you feel it? Do you feel it, Mr. Krabs? You can feel the [00:21:00] difference and your customers will. By the way, here's a scorecard for human experience that you can print off. Um, just go ahead, copy and paste this, and you can go ahead and print off or just email me, uh, brian@briannicholsconsulting.com and I'll send this over to you myself.

Grade each month, one to five effort. How many steps do a human, how many transfers, respect do customers feel heard in the first 30 seconds? Trust. Do they believe us more or less after we talk than they did before? If you're not improving those, your AI project is not a CX project.

And by the way, you're gonna get objections here, right? So when you hear we can't afford more humans, I want you to remember to say, Hey, you can't afford churn. You can't afford reputation damage. You can't afford repeat contacts caused by bad first connections. Or our AI vendor says deflection is success.

You say, Hey, success is lower effort and [00:22:00] higher trust deflection. Without those, that's just hiding in the pain. Or how about customers who say things like, um, I actually like self-service. Okay, they like good self service and the second it fails, they want a person fast. So. Give them both. And uh, by the way, I want you to teach your team the three sentence cx.

And this is again, the rule of three. Train everyone to hit these lines. I see it. I've got it. Here's what's next. That's the spine of every great interaction. , if you only change that. Your he score skyrockets. And by the way, this requires a culture shift.

We have to go from metrics to meaning and yes, metrics matter. We're not data here, but we're anti metric theater where pretty charts while customers are pissed off. . The culture shift is when [00:23:00] leaders start reading these out loud in meetings, not just on dashboards.

When a CEO calls a customer back when the VP sits on a support shift for two hours every Friday, that's when the organization remembers what business it's in, the human experience business. And by the way, AI versus AI in the wild. Think of it like this grocery store. Self checkouts great for two items, bad for a full cart, and that one random barcode from 1997.

The human clerk saves the day. Or how about airline apps? Great for boarding passes. Really bad for rebooking a family of five. During a storm, the human agent becomes the hero or doctor. Portals great for getting your lab results really bad for scary symptoms at 2:00 AM that nurse line matters

tools help until they don't. Humans are the safety net and design your infrastructure with that in [00:24:00] mind. So here is your 30 day challenge. Remove one step between your customer and a human. Add one forgiveness lever for your reps. Add one proactive human touch in your highest friction journey. Kill one robotic BS phrase and read five outcomes out loud every Monday.

If your inbox doesn't feel calmer in a month, I'll go ahead and I'll just rip up my show notes and I'll toss 'em into the fire because I can guarantee you will feel the difference because your customer will feel the difference. And by the way, there are times when not to automate, when not to ai, just in general.

And that could be anything re regarding grief, your finances, health kids, if the topic touches any of those, just default to a human cancellations and complaints. I mean, if you try to save a customer with a bot, you'll save a dollar, but you're gonna lose a year. Revenue or [00:25:00] promises. Never let a bot promise a timeline that it cannot keep and is not held to humans.

Own commitments, automate checklists. Automate updates, automate lookups, but do not try to automate empathy. And where to automate first. Where's my stuff? Great for automation? As long as the answer is real and the dms are open, how do I do whatever knowledge bases with short human videos beats a paragraph that a bot will read to you.

Hey, I need a copy of invoices, receipts, policies, automate retrieval. With a human escape patch, again, start with low risk tasks that save time without risking trust. That's how you build momentum. And by the way, if you're using a bot upfront, teach it. Humility. I can help with lookups and quick answers

if I can't fix it in one try, I'll get you a real live [00:26:00] person. Wanna jump to a human now? See that right there when a bot admits limits? Customers forgive when a bot pretends to be all knowing and fails. Customers rage and rightfully so. Alright, so here's where we land and where this episode's gonna wrap up.

The future of customer experience. It's not anti-technology, it's anti-friction. It's Antifa, it's pro truth, pro ownership, pro results, dare I say, pro outcomes. Human intelligence is the differentiator because people solve messy problems and make judgment calls. Artificial intelligence becomes powerful when it supports those people and the human experience, the feeling your customer walks away with, that's the only scoreboard that really matters.

So the next time someone says in a meeting, can we AI this? Ask the real question. Why [00:27:00] will this improve the human experience? And if the answer is yes, go for it. Build it up. But if the answer is, maybe pilot it. And if the answer is no, just stop abort. Put a person there and give them tools. Let them own it.

Watch the metrics you actually care about. Start moving in the right direction. Alright folks, if this episode of CX without the BS help, please do me a favor, share it with a teammate who keeps getting dragged into the AI theater. I call this shiny object syndrome. It's everywhere today. And by the way, again, if you enjoy today's episode, please drop a comment.

With your best bot story or your worst bot story, and then maybe your best human save. And if you want that simple one pager I mentioned earlier, go ahead, email me Brian at Brian Nichols Consulting and I'll send that over. And if you're building something that actually reduces effort and raises trust, I wanna hear about it.[00:28:00] 

Remember, hi versus AI impacts E. Put the human experience first and the success we'll follow. Alright folks. This is Brian Nichol signing off for CX without the bs. We'll see you next week.