Earley AI Podcast - Episode 16: Using Data to Deliver Customer Experience That Matters with Jim Iyoob

In this episode, Seth Earley and Chris Featherstone speak with Jim Iyoob, Chief Customer Officer at Etech Global Services and 2022 inductee into the CX Hall of Fame. Jim brings 25 years of contact center experience - starting as a frontline agent at Dun and Bradstreet - to a conversation about why the biggest mistake organizations make with AI is treating it as a plug-and-play solution. He explains how Etech processes 3.5 million interactions per month using speech-to-text engines and a dedicated team of data engineers, analysts, and scientists, why there are over 420 different sentiment variations for the single word "good," how 100% AI-powered sales verification eliminates gaming, and why transparency is the only path to genuine accountability. He also makes the case that the contact center - not the boardroom, not marketing - is where organizations are sitting on their largest and most honest data asset.

Key Takeaways:

• AI is not a plug-and-play solution regardless of what vendors promise - buying speech analytics is like buying PowerPoint, because the tool still requires skilled humans to do the actual work of configuring, training, and maintaining it.
• Domain expertise overshadows all technical expertise - you must understand your customer, your products, and your contact center behaviors before any analytics tool can deliver meaningful insights from the interactions it processes.
• There are over 420 different sentiment variations for the single word "good" in a machine - context before and after a word matters enormously, meaning NLP models require continuous human curation, not a one-time setup.
• The contact center is the largest and most honest dataset in any organization - it captures what customers actually say, not the biased sample from surveys filled out only by people who are angry or were bribed with a gift card.
• Transparency drives accountability - when agents can see their effectiveness rate across hundreds of calls instead of a supervisor's judgment on ten, there is no gray area, no bias, and a clear path to coaching that actually improves behavior and income.
• Quality is a behavior, not a score - Etech's Key Performance Behavior Index identifies three specific behaviors to focus on per agent at a time, because you cannot tell an agent fifty things they did wrong and expect improvement.
• Knowledge bases at most organizations are fragmented, siloed, and outdated - companies hand over a PowerPoint or a URL with FAQs that has not been updated in five years, making it impossible to build effective virtual assistants or empower agents to resolve issues efficiently.

Insightful Quotes:

"Unfortunately these AI companies are out there saying I'm going to solve all the world's problems - it's a plug-and-play solution. Which is a lie. You still have to do the work. If you bought Microsoft PowerPoint, it's not making your presentations for you." - Jim Iyoob

"You are sitting on the largest dataset on the planet. And if you use it correctly your customers will tell you what they like about you and what they hate about you - not just the people who said I hate you, I love you, or were bribed with a Starbucks gift card to fill out your survey." - Jim Iyoob

"Domain expertise overshadows all the technical expertise you can have in this world. Unless you understand your customer, the tools at your disposal are not going to be able to help you." - Jim Iyoob

Tune in to hear Jim Iyoob walk through how Etech uses 100% AI-powered call verification to catch gaming in real time, how analyzing 10,000 calls before onboarding a new client produced better training materials than the client's own documentation, and how a spike of 185 calls in queue was traced to a broken website page - in five minutes - before the IT team had even started troubleshooting.

Links:

• Etech Website
• Jim on Instagram
• Jim on Linkedin
• 22 Influential Leaders to Watch in 2022
• The CX Hall of Fame

Thanks to our sponsors:

• Marketing AI Institute
• CMSWire
• Earley Information Science
• AI Powered Enterprise Book

Podcast Transcript: Domain Expertise, AI Myths, and Why Your Call Center Is Sitting on Your Biggest Data Asset

Transcript introduction

This transcript captures a conversation between Seth Earley, Chris Featherstone, and Jim Iyoob about what it actually takes to use AI and analytics to deliver meaningful customer experience in the contact center. Jim draws on 25 years of experience - from frontline agent to Chief Customer Officer - to dismantle the plug-and-play myth, explain why domain expertise beats technical expertise every time, and share concrete examples of how Etech uses 3.5 million monthly interactions to build knowledge frameworks, drive agent coaching, power 100% sales verification, and detect operational problems faster than any IT team.

Transcript

Seth Earley: Welcome to today's podcast. I'm Seth Earley.

Chris Featherstone: And I'm Chris Featherstone. Good to be with you.

Seth Earley: Our guest today is Chief Customer Officer at Etech Global Services. He is a 25-year contact center veteran with experience in inbound, outbound, chat operations, and social media management around the world. His work has been featured in numerous publications including CIO Review and Contact Center Pipeline, he was listed as one of the top 20 influential people to follow on Twitter, and he was just inducted into the 2022 CX Hall of Fame. Please welcome Jim Iyoob.

Jim Iyoob: Thank you so much, Seth and Chris. It's a pleasure to meet you guys. I'm humbled to be here.

Chris Featherstone: I love this because you know, what's fun is that generally I'm always in awe of the academic circles Seth runs in - and I'm always asking them: how does that thought or algorithm actually drive a business outcome like customer experience? You are that topic, so this is going to be super fun for me.

Jim Iyoob: I think we're both going to have some fun. I'm very passionate about the experience, and I think that's what brought me some of those accolades over the years. At the end of the day, it's about that experience. And if you're not doing something about it in a global marketplace, your customers can buy it anywhere.

Seth Earley: Tell us about your background and how you got into this space.

Jim Iyoob: Fortunately for me, I've had the same boss pretty much my whole career. When I came out of college I went to Dun and Bradstreet and I met this guy named Matt Rocco, who is now Etech's President and CEO. I was just an agent doing the basic stuff everybody does. I walked into his office and he asked me what I wanted to do. I looked at him and said: I want your job. For whatever reason, instead of firing me, he took me under his wing and became my unofficial mentor. Nine months later I had been promoted and was working in bigger parts of Dun and Bradstreet.

One night at dinner he said: we're moving to a place in Texas. I said, where is that? He said, you'll find out. So he recruited me to what was then called Barry Direct - one little call center, 400 seats, one customer - and hired me as an account manager. Two years in, BellSouth divested the assets and he took the company private. Fast forward to today: Etech operates 10 call centers in three countries, with no mergers, no acquisitions, no venture capital, and about 3,600 employees globally.

Matt Rocco has done a great job instilling servant leadership. When you instill a people-first culture - technology enables, we get it - but when people are actually first, people will be happy to work for you. And everything I learned as an agent, as a coach, as a frontline leader brought me to this: I have to give the best experience not only to our employees but to the customer. Some of these companies come in with three-piece suits and beautiful PowerPoints and they're selling tech stacks that are not the holy grail. I think it's the people and the culture.

As far as the CX recognition - one of my goals throughout my career has always been to add value and be a great person. What I look at in customer experience, I'm looking at it from the customer's lens - the end customer, not the brand. And I'm software agnostic, because there is no best solution. First you have to identify the opportunities your customer is trying to solve for. Once you understand that, and only then, do you earn the right to give them a recommendation.

Chris Featherstone: There's 60 to 160% attrition year over year in call center environments, and yet organizations don't think about the psychology of what they're doing to their most valuable asset - their human capital. These people are paid some of the worst wages yet held to some of the highest standards in the organization. What are we missing?

Jim Iyoob: For the record, not everybody is like that. If you think about analytics and tech stacks and how you intersect them with culture - technology, specifically performance management, is an enabler. Speech to text is an enabler. It's a driver toward a positive culture, if deployed correctly.

When you think of a positive culture, a few things come to mind. Transparency - transparent performance evaluations, communicating key metrics consistently. And reduced agent effort - how hard is it for the agent to get feedback and do their job? When you look at speech analytics, it hits all those points. There's no gray area in performance management when you're analyzing all calls, not making bias decisions based on 10 or 12 calls. Agents and supervisors have clear sight to what their KPIs are holistically.

An agent interacts with four to ten different tool sets just to get the data they need. You want them to take care of the customer and also watch 16 different windows? And then these "agent enablement" solutions want to tell the agent in real time when the customer is getting frustrated - on top of the 19 things they already have open. You want to give them something else in their headset?

What's missing is the data science team that has domain expertise - leaders who understand the product, the solution, and the CX indicators and behaviors we want to drive. No disrespect to the IT guys, but the IT person sitting in the back is solving problems we don't even need solved, because he's probably never taken a call in his life.

Chris Featherstone: How do you start when you go into an organization?

Jim Iyoob: The first thing I tell people is domain expertise overshadows all the technical expertise you can have in this world. Unless you understand your customer, the tools at your disposal are not going to be able to help you.

At Etech, we have a data engineer, a data analyst, and a data scientist - there's a reason for three different buckets. We have industry-specific leaders with domain expertise who came from the call center. That is what ensures that when you implement a tech stack, you become an expert not just on the product but on the behaviors you are trying to drive. We run 3.5 million interactions a month through our speech-to-text engines.

And unfortunately, these AI companies are out there saying: I'm going to solve all the world's problems, it's a plug-and-play solution. Which is a lie. Did you buy Microsoft PowerPoint thinking it was going to make your presentations for you? You still have to do the work. The sales guy doesn't know how to use the tool he's selling. He's just trying to get a sale.

Seth Earley: Talk about what goes into setting up controlled vocabularies and an architecture for this - how well prepared are organizations typically?

Jim Iyoob: There's so much confusion in the marketplace because software companies are telling organizations it's plug and play and it's not. They say "I'll set it all up free in 90 days" - which they do, but then nobody knows how to run it.

I talk about our engineering knowledge framework from the start because we've already built it. You create the categories and the knowledge base - and here's a good analogy: there are over 420 different ways to express sentiment for the single word "good" in a machine. Good is not always good. If I'm Charlie Brown and I say "good grief" - that's not good. So your engineer has to build a knowledge base, look at natural language processing, look at sentiment, and look at what is happening before and after the word that was said.

Another example: I have a Texas accent. If I say I'm getting my oil changed, the transcript might come back with "a-l-l" instead of "oil." But a data scientist who sees "I'm going to get my a-l-l changed" understands from context that it's oil. That kind of fine-tuning is a full-time job. We have full-time engineers whose only job is to listen to subsets of calls and make sure the engines are hitting correctly - for example, that AT&T comes through as "AT&T" and not "AT 18." That is how we get to 85% accuracy. Yes, some companies will tell you they're at 92 or 93% - and I've seen it - but the only way that happens is with constant ongoing maintenance.

Chris Featherstone: What's the right approach to getting from strategic goals through to the right agent incentives?

Jim Iyoob: Sales queues are a good starting point because 100% sales verification is now achievable with AI, and it actually prevents people from gaming the system. Every time an agent sells something, I'm recording the call cradle to grave. We've coded our machine to verify that the price was stated correctly, that the terms were explained. If the sale is rejected, it goes back to the agent and their supervisor within seven to ten minutes so they can recall the customer and potentially resave the sale.

Now, the deeper incentive question: transparency drives accountability. When I go to an agent and say - hey, we listened to 150 phone calls and you're only effective at probing 62% of the time - that is a completely different conversation than the supervisor saying "I listened to a call and you didn't do well." There is no gray area when accuracy comes from hundreds of calls, not three.

And when you deploy this correctly, agents actually love it, because now you're being transparent, you're being servant-hearted, and you're actually showing them how to earn more money. We say: want to make more money? Let me show you the behaviors the person in the top tier is doing consistently.

The key insight we document is that quality is not a score, it's a behavior. Our Quality book makes that point - and our Key Performance Behavior Index identifies three specific behaviors to focus on per agent. You can't tell an agent 50 things they did wrong. Give them three. Get those moving. Then work on others.

Seth Earley: What are you seeing when you go into organizations and look at their knowledge bases?

Jim Iyoob: It's an eye-opener. If you look at what most organizations have, your client hired you, gave you a PowerPoint presentation if you're lucky. If you're extremely lucky, you got a facilitator guide. And if they're really a well-versed operation, they gave you a URL that hasn't been updated in five years with their FAQs. The reason it's siloed is because different people own different pieces of it.

The Holy Grail is combining call center metrics, quality, and knowledge into one platform, so the agent does not have to go to 50 different places. That is on our product roadmap. But to answer the broader question - the call center should be able to get the most information, because that is where the data is actually generated. The problem is the person in ops thinks their job is just to coach the agent. That's why it remains siloed.

The real opportunity is having a dedicated person come in and look at it from a different perspective. If you do a true knowledge pull - you can do it with APIs - it might take 30 to 60 days. It's not the rest of their career. You need someone full time to do it and approach it fresh, because if you're too close to your own data, you stop seeing what's in it.

Marketing sometimes doesn't like to talk to sales. Everything coming out of marketing's knowledge base sounds legal and scripted, which is terrible for the agent. If you talk to the CFO, all they want to know is how much does a knowledge base cost and how many minutes can we shave off average handle time. You have to get all of them in a room, and - as my boss taught me a long time ago - the first rule is: none of you have titles. We're all the same in this room. Then you can drive the right leadership, put a plan of action together, and assign it to an individual - not a team. When everybody is responsible, nobody is responsible.

Seth Earley: How do you use that data to drive upstream fixes - not just call center improvements?

Jim Iyoob: As a trusted advisor, if our call volume drops because we fixed something upstream - I add value and we win more business. Here's a live example. We had a bunch of calls and chats coming in asking about certain things that should have been self-serve. My agents just thought: we're getting a lot of these calls. When I sent in an analyst, we found that 22% of our interactions in a given month were people asking for a copy of their receipt. And when I went to the website to check - it reminded me of the Tootsie Pop commercial: how many licks does it take? It took me nine clicks to find my receipt. That is not effortless. When you give them the data quantified like that, things get done.

And here's another one. We had a TELCO client with 185 calls suddenly spiking in queue. IT was standing up all their incident bridges thinking it was a switch problem. I went into the speech analytics tool, did a browse-and-search, and within five minutes we were able to tell them: one of your website pages is down - here is the exact page. We gave them that information before the IT team had even started tracing their problems.

Seth Earley: Final thoughts - what do you want to leave people with?

Jim Iyoob: Three things. First, be a lifelong learner. I started as an agent. I read an article a day - that is my motto. I am a self-learner. I didn't know AI when I started in it. Now I'm considered an expert. You don't have to go to 50 different schools. Read one article a day and you will become a master in a year.

Second, take what you do seriously but don't take yourself too seriously. I think in the corporate world we've lost personality. Be passionate about what you do but keep your humanity.

Third, deliver customer experience that matters. Stop looking to solve all the world's problems. People make mistakes, and you learn from those mistakes. My boss will tell you I make hundreds of mistakes a week. That's great - I'm still here after 30 years. There is no losing, only winning or learning. Don't be afraid to fall down and get up.

Seth Earley: Jim, this has been fantastic. Congratulations again on the CX Hall of Fame, and thank you so much for joining us today.

Jim Iyoob: My pleasure. If you want to follow me on LinkedIn, I put out a CX article every month. And if you want the sarcastic stuff, find me on Instagram at jim_iyoob. Have a blessed weekend.