Risks of Fragmented AI Implementation
Mary Beth Hamilton – 00:00:04
We’re gonna take a look at why AI doesn’t fix a fragmented customer experience. And why we’re having this conversation is that so many organizations are applying AI to potentially broken work processes without trying to understand what the problem is they’re actually trying to solve and the outcome that they’re trying to achieve.
So, there’s this assumption that you can add new capabilities, and all of them will enhance the customer experience. But in reality, you may create more fragmentation, etc.
So, my question for you is at what point does adding more CX technology actually sort of start to worsen the customer experience?
Malachi Threadgill – 00:00:36
Yeah. It’s a really good question. I think that, you know, today AI is a hammer, and everything is a nail. We’re seeing over and over again that there’s a ton of AI companies just popping up.
A lot of them are great. A lot of them are just clever, you know, OpenAI, API wrapped, tools and solutions. And I think that there’s also a lot of pressure coming from boards, leadership, etcetera, to try and solve for AI and in particular CX.
So, I think that when leaders, CX leaders, IT leaders, etcetera, are looking at how to implement AI, I think the speed by which they think that they need to bolt on additional features and not really understanding what the implication is to the data, to the insights, can it actually do the job that it’s supposed to do, can create some friction.
What we’re seeing just generally in the industry is that a lot of CX, CCaaS types of solutions that are looking for these external capabilities are often finding a lot of difficulty implementing it, points of failure, the team on how to use it.
And so oftentimes the end goal of adding AI immediately without understanding the entire ecosystem and the brevity of what’s actually going on is creating friction and actually slowing down a lot of customer experience interactions. If we think about it, today more than ever, customers are benchmarking their interactions with businesses against Amazon same-day delivery, against the Genius Bar, against the Ritz-Carlton. And so, all organizations need to think about CX and what that means. And running fast to implement a solution that’s not properly thought out can actually cause massive damage to the brand and reputation long term.
Mary Beth Hamilton – 00:02:04
Yeah, and you bring up brands. We do have a question on that a little bit later. But before we get there, so what are some of the signs that an organization and maybe from the perspective of an IT leader, someone on my team has just come to me and said, we have to implement this solution, it is going to solve x, what is the sign that you are actually just like picking a feature to solve a problem that could actually potentially create more challenge in the CX experience?
Shifting IT and CX Decision Makers
Malachi Threadgill – 00:02:27
Yeah, it’s really interesting just looking at the shift that we’re just seeing, and partners are also seeing is historically the channel has sold through the IT decision maker. And primarily when it was just voice or plain old telephone service, price was the primary driver, and the IT leader historically has worried more about price.
When you think about additional features and capabilities that are actually extending the phone systems into more CX type of solutions, and we’re layering in omnichannel capabilities, AI embedded across all of that, the conversation has to change where the IT decision maker is more of an influencer and now you have to look at the other parts of the organization, the CX leader, the CFO, CXO, etcetera. And so, when we think about just, okay, we need this capability when someone’s decision making process is primarily on is it CapEx versus OpEx, I’m not looking at the other impacts that these solutions make, it makes that sale a lot harder, especially in the partner ecosystem.
Then the second part of that bolted on feature is to say, okay, you need to look at your phone system as well as your omnichannel capabilities in its entirety and what integrations do exist.
Built In Versus Bolted On
Malachi Threadgill – 00:03:28
The phrase we use frequently is built in, not bolted on, because you could go out and find a hundred different solutions that you can bolt on to this and add it in. But once again, kind of where we started with, if it doesn’t have access to the right data and what we’re finding more important than anything is the integrations into the actual systems, whether it’s healthcare, automotive, legal, whatever these verticals have with the purpose built integrations, if it doesn’t have a proper flow through into where the CX leaders are actually doing work, then you’re gonna run into friction with these bolted on types of solutions.
Mary Beth Hamilton – 00:03:59
Yeah, and that’s one thing that we talk to clients a lot about is the idea of automating broken work, right? You really need to sort of relook at how you’re doing these processes to make sure that you’re not just adding something on to something that may already have inefficiencies, etcetera. Can you sort of “love story” time? Give us a case study or an example of where you’ve worked with a client that has done it the right way.
Malachi Threadgill – 00:04:20
Yeah. Absolutely.
Healthcare and Automotive AI Use Cases
Malachi Threadgill – 00:04:22
And I think that it’s really fascinating to look at the role that these types of solutions are starting to play. And there’s a lot of fear and uncertainty and doubt as it relates to AI. Obviously, the most common denominator is job replacement. But one of the things that we’re seeing in particular, the two areas that we’re seeing massive growth right now with AI adoption, selfish plug is because we’ve done the integrations into Athena and other healthcare systems and an automotive Tekion One and other systems as well.
But both of these, although they’re very different industries, they have similar patterns. And the easiest thing to say is if you’re an automotive principal or a dealer owner, you know exactly how much a missed call costs. You know exactly how much call costs in terms of service, terms of uptime, etcetera. And AI, in its most basic form, in that scenario, with an AI receptionist, means no more missed calls.
We can automate a lot of those processes. But I think what’s more important is that, when we look at the healthcare industry in particular, CX employees, or customer-patient employees, they have high turnover; they’re incredibly stressful jobs.
If you’re thinking about when a patient’s calling in because they have chest pain, they’re stressed out; maybe they have insurance issues, which can be incredibly stressful, which causes massive turnover when we think about the customer-patient service reps in this industry.
And so, we’ve seen over and over again: not only if I need to reschedule a call, reschedule an appointment with my doctor, my primary care physician, or if I need to check on a prescription, I don’t necessarily need to talk to a human, right? So, AI receptionists can solve those basic frontline items for me. I may not wanna wait five minutes to talk to the front desk, but when we think about the second and third layers of the actual CX employees, when they think about sentiment analysis, the during call transcription, knowledge base, the coaching that can come in while they’re having those conversations, it actually helps those employees navigate difficult conversations.
“Hey, here’s how you should respond.” “This is going this way.”, and it actually increases the quality of life of those employees and solves those difficult outcomes. So that leads to higher employee scores; it leads to longer tenure.
You don’t have to invest as heavily in training because you’re turning over employees over and over. And then finally, with the analytics layer, it allows the leaders to understand, okay, when are these issues happening most frequently? Can I coach my employees on how to better handle these types of things? And so, the net is that you have happier patients or customers, which ultimately drives loyalty, which ultimately drives retention, if you will.
Eliminating Swivel Chair Inefficiencies
Malachi Threadgill – 00:06:42
It’s really powerful to just look at the positive side of how a lot of these solutions really work. And the final thing is, to kind of your earlier question, when I say built-in, not bolted on, the term we used over and over again in the industry is swivel chair. If I have Salesforce over here, I’ve got my ERP system over here, and I’ve got my phone system over here, and I’m constantly having to copy and paste data into each of those. We also start running into error issues, quality issues, and time issues.
When we think about a phone system that unifies communications properly integrated into Athena Health, etcetera, and AI and analytics flows through, it just makes it seamless for the patient and the CX leaders.
Mary Beth Hamilton – 00:07:19
So, the swivel chair is an interesting idea, and you sort of hit on: what does it solve beyond just the cost, how it increases by addressing it, the employee experience, etc. And then touching on the unified aspect of it, and you’ve touched on like, rip and replace, where do you start? For a CX leader, what is the path forward to achieving this? Is it a rip and replace?
How do you get to that unified CX environment that achieves the outcomes that you just outlined?
Navigating Rip and Replace Decisions
Malachi Threadgill – 00:07:44
Yeah, I think it depends on a lot of factors.
Right? It depends on what industry you’re in.
It also depends on the size of the business and the features and capabilities that you need, and the capabilities that you think you need, more importantly.
A lot of enterprises go a certain direction because this is the enterprise play, when in reality, a lot of the vendors in the space today have similar capabilities that are described differently, that solve the same things at a fraction of the price.
We see a wide gamut from people still using, there was- I’ve been in this industry a long time, and a long time ago they said PSTN was gonna be gone by 2020. It’s still here, there’s a lot of people that still have on-prem equipment that haven’t even gone to the cloud yet. So, in those instances, absolutely, probably rip and replace. There are areas of efficiency just generally that they can enjoy.
But I think it kind of goes back full circle to the first question: what are the problems that you’re trying to solve? And are these problems worth solving? So, if you think you need AI to solve a problem, the common thing we say is ask why five times, right? I need this to solve X.
Well, what do you really need it for? Why do you need this? Why do you need this, etcetera? And that really helps you get to the core base and foundation of what you’re looking for.
When you think about ripping out a system and replacing it, understand what are those second- and third-order impacts that are going to happen.
Just saying AI receptionists is easy, but IVR and IVAs are pretty complex. And so, understanding all of that technical debt that you’ve invested, it may not make sense to actually do a rip-and-replace. In that instance, bolting something on may make sense.
So just really understanding the entirety of the ecosystem is really important and looking ahead to what you’re actually trying to solve.
Mary Beth Hamilton – 00:09:17
No. That’s a great sort of way to look at it.
And you mentioned, like, ask why five times, I think, sort of to the point of understanding what you are trying to solve for. I think bringing it back to something you touched on earlier is that the buying committee or those impacted by CX is broader than just the IT leader.
So, it’s sort of that importance of having that conversation with multiple leaders within the organization: the CXO, the CMO, the whole host of other individuals within the organization.
Are there other, sort of, when you think about who you should be, organizations should be having the conversation with internally to sort of get at the heart of what they’re trying to solve that you recommend they think about, like all of those individuals in the conversation?
Understanding the B2B Buying Committee
Malachi Threadgill – 00:09:54
Yeah, absolutely.
As I alluded to earlier, I’ve been in this industry a very long time. I did take a brief detour for about a half decade and I actually worked at Forrester doing some research for a little bit of time and while I was there, we talked quite a bit about the idea that we called it the buying group and our research showed us that 85% of B2B decisions are made, above 5k ARR, which is small in our space, are made by a committee of four or more.
So, really understanding, what is the understand the CIO, they are worried about CapEx versus OpEx.
So is the CFO, they’re worried about primarily that phone line. As you start to think about other categories, whether it be the CMO or the chief sales officer, how can I collect these insights so that I understand how to train my sales team on outbound dialing, handling phone calls, etcetera?
So, really understanding those pain points. And I think of this as an opportunity, not just for Go to as a vendor, but for partners like you and just the industry at large, is really empowering our partners to understand the shift that has occurred as we’ve layered in CX, AI and analytics, and really helping to solve those pain points.
Because if I can show the sales leader that they’re not gonna miss any more calls, or I can show them the common use case that I use is that if I work at a tire shop or even if I have a barbershop, 80% of my missed calls happen thirty minutes before I open, thirty minutes after I close, and the hour during lunch.
If I can go in and show that business leader or small business owner or even medium-sized business those types of insights and how my solutions can answer or solve those problems, once again, now I’ve moved a back-office solution into a front office solution.
And that evolution, I think, is a really powerful thing that all of the channel and partner community should be considering to really change the narrative of how they can be successful.
Mary Beth Hamilton – 00:11:34
Excellent. So, flipping back to AI and the employee experience, what should AI take off of an employee’s plate and what should say distinctively human?
Human Ingenuity Versus Automated Tasks
Malachi Threadgill – 00:11:44
I think personally, like I alluded to, one of the things that I look at is there’s stuff that AI does incredibly well.
If I need to reschedule a call, a voice agent can do that for me. If I’m worried about my chest, I should have a fast lane into talking to a human being. And I think that AI is really good at solving those simple tasks. It’s also a very good support agent in those difficult conversations.
The things that do need to remain human is, I think, the ingenuity and the ability of humans to have those connections and actually solve real-world problems will remain human for a very, very long time.
When we look at the kind of problems that we can solve once again- after-hours dialing, those types of things- it increases the efficiency of those employees, and it provides more meaning for employees as well. If they don’t have to handle these simple, mundane tasks, the repetitive tasks that are happening over and over, it allows them to focus on more important work.
They don’t need to necessarily sit at a cubicle; they can be out helping in the service areas; they can be out helping patients and those types of things. And so, I think, yeah, the repetitive tasks AI has got, but there is no replacement for the complexity that we have.
Mary Beth Hamilton – 00:12:51
Excellent. So, where that also leads into is this idea that each one of those different touch points from a brand perspective, a customer experience perspective, there is the potential for it to become fragmented and not feel like a unified, I will use unified from the aspect of a unified customer experience and unified brand for that client.
I guess, does fragmented AI become a brand problem and not just a technology problem? And then what is the sort of solve for that?
Preventing Data Silos Through Unification
Malachi Threadgill – 00:13:18
Yeah, it’s funny. You know, I think GoTo is one of the, not just in terms of the AI solutions we’re bolting on, but even internally, we’re ahead of the curve in terms of how we’re using AI just in our daily lives, from sales to marketing to others. And I think part of what we’ve done well is centralizing AI so that we don’t have all of these dark investments into someone using ChatGPT over here, someone’s using Claude over there, someone’s using Gemini, and they’re all paying for it themselves, and they’re feeding data that is proprietary and confidential to third parties. There’s a lot of governance that people should be thinking about.
So first and foremost, I think having a philosophy in terms of how the organization is going to use AI internally, but also how is the organization going to implement AI from a customer experience standpoint, is crucially important. Just to give one example, in today’s day and age, we don’t get to choose how people want to interact with us. Some people may prefer SMS, some people may prefer WhatsApp, some people may prefer email, phone, or fax. Mean, fax is still around.
I think that there’s an AI solution for all of those, and then there’s AI that combines all of that.
The main goal, I think, regardless of what, so fragmented AI creates fragmented data silos. Ultimately, if you think about the omnichannel capabilities and you think about how AI and analytics and the integrated systems can give you one unified view of the customer, I think that’s the most important thing that CX leaders should be considering.
If they can think about how to unify that entire journey over the life cycle of the customer and how AI can implement it, it might be multiple solutions but ultimately making sure that it’s fed into that unified vision.
That’s where we’re going to see the success that CX leaders need to focus on.
Mary Beth Hamilton – 00:14:54
Excellent.
So, we touched on all this drives back to improving the experience, driving the business forward outcomes, etc.
Around the AI topic, we hear a lot about the various pilots and everything that people are putting into place. And then it always comes down to how are you measuring it? How do you know what success looks like? What’s the ROI?
Before we go into that conversation, if we take it back to like, how do you tell that what you’re putting into place and how do you measure is improving the customer experience?
Measuring AI Success and ROI
Malachi Threadgill – 00:15:19
Yeah, there’s a number of ways to look at it. I mean, the most simple is NPS. You know, if you ask your customers what they think, they’re gonna tell you. And that’s actually, you know, the philosophy that I think we’ve done really well.
We saw natural pockets where we were successful, education, automotive, healthcare, and we asked them what they wanted and then we built products to support them. But then to your point, how do we show them that the solutions we’re building for them are actually driving success?
So really, understanding what matters is that output. So, as you think about the evolution of the buying group, knowing that the principal cares about X, Y, and Z, and being able to solve for those outcomes and being able to show them through analytics.
I think analytics is the underlying piece here that a lot of people don’t consider because of the focus right now on AI, is that you can show people how many missed calls they had before. You can show how many fewer missed calls they’ve had today, and then you can start to quantify that monetarily for them.
And I think oftentimes, this industry, because it’s evolved so much so quickly over the last three or four years, it used to be you didn’t have those insights, so you just needed a phone system. So, what is the lowest common denominator?
Price, free phones, SPIFFs, etcetera.
And just that shift that we’ve seen over three years, it has evolved where what matters to the barbershop owner may matter; something way different may matter to the retail store, and you have to ask them, and then you have to build analytics and solutions that help them answer: is this actually solving my problems, and is it worth the cost of replacement?
Mary Beth Hamilton – 00:16:41
Yeah, and I think I love the starting with NPS, but so many things can claim that they were the ones that impacted that NPS, right?
There’s a lot of moving parts that can impact that score. So, I think as an overall organization, having that as a key metric that’s being tracked, but then, to your point, figuring out what are the other items, what are the other components that need to be tracked from the analytical perspective to know that you’re moving the needle in those different respects and prioritize them for the organization.
So, you can see NPS, but then you also have visibility further on to understand which levers may or may not be impacting that client experience.
Malachi Threadgill – 00:17:15
Yeah, couldn’t agree more. I think one of the challenges that we have is whether it’s NPS, CSAT, and even as a marketer, we see things that we can pat ourselves on the back for as well. And it’s fascinating that if we focus on one metric, the metric becomes, we can say, “Oh, we have a high NPS”, but when you start to dig into the organization, is it really what it is?
Did we actually pull all of our customers?
There’s various levers that people can pull even in marketing: oh, we hit our MQL goal, but if sales doesn’t sell, then the number of MQLs we’ve created doesn’t actually matter.
And so, really going kind of back to that fundamental buying group: what are the four or five key areas, and what are the things that matter most to them?
IT cares about costs, as I said; sales cares about X, Y, and Z, and it’s not one metric; it’s a myriad of metrics. And that’s just the big shift that we’re seeing with this technology and how fast it’s all advancing.
Mary Beth Hamilton – 00:18:05
Excellent. So, before we get some of our final questions, what are some of the risks that leaders underestimate as part of a move or transformation in this area?
Data Governance and Implementation Risks
Malachi Threadgill – 00:18:15
I think the key risk is not knowing what you’re getting into, thinking that you need it just because you need it. You know, what problem are you trying to solve?
I think the actual risks in terms of implementing it are turning on features and capabilities that you maybe shouldn’t turn on. HIPAA is a really great example: data retention policies- where is the data being stored?
Because of what AI is able to do today and turning unstructured data into structured data, whether it’s transcription, sentiment, email, whatever, understanding how your data is stored needs to be stored and the policies there are typically the biggest red flag, where if an organization isn’t prepared for what that means, it can cause down funnel issues for them that they’re not even thinking about.
The other parts of that are, as you start to build out knowledge bases, chatbots, etcetera, leveraging AI and those tools is the accuracy, what you put in and what comes out.
Oftentimes, we call it hallucinations, and there are fewer hallucinations today, but the bad data that we could use in the knowledge base, whatever it may be, AI receptionist, compounds very, very quickly, and if you think about the most common issues around that is those errors aren’t found once they’re implemented for weeks, months, or even a quarter.
And if you look at the micro impact that that can make over time, it can be massive for organizations.
And then there’s just the practical pieces, making sure that everything’s connected properly, making sure that we’ve properly tested and embedded it. Often, people just wanna move quick and turn things on, and not thinking properly about how it all works together can create friction and problems down the road.
Mary Beth Hamilton – 00:19:47
Yeah, and I think that data cleanliness, getting that wrong, and like you said, it’s not immediate. But as customers start to experience wrong answers, or things that just don’t feel accurate, it quickly degrades trust, right?
All these steps forward that you may have made now can be impacted by little things that a human never would have picked up, or a human would have known to correct. And so going back to the importance of the foundation being correct, and then learning as you go, and then having gates for being able to identify those challenges and address them.
Malachi Threadgill – 00:20:19
Absolutely, keep it simple, start small. Don’t try to build the whole thing all at once and make sure you do have those gates of iteration to make sure you’re getting everything right because could have set step one, two, and four up correctly, but if three is wrong, you may not see it now, everything’s broken.
Mary Beth Hamilton – 00:20:34
Absolutely.
Okay, so obviously this is more than a ninety-day solve, but the idea of the theme of this conversation is sort of the next move that a leader should make. What would you advise is the most practical next move a leader should make?
I’m gonna say in the next ninety days, when considering buying an AI solution.
Practical Ninety-Day AI Strategy
Malachi Threadgill – 00:20:52
Wow.
It’s tough. Yeah, if you’re not thinking about AI, you’re gonna be left behind.
And I think anyone who’s probably seeing this does understand the importance of AI and they’re already thinking about it. I do think that today when you look more broadly at everything that’s happening is AI is doing two things simultaneously.
One, it’s become a commodity very quickly.
Years ago, there was massive investment in data scientists by all the companies and then OpenAI released their API and boom, just like that, it’s a commodity. So, what that means is that there’s massive competition out there today.
If I were to make a recommendation to a CX leader, I would say, look at where I am today: what is going well in my organization? What are areas of optimization that I need to make? What are those friction points that I have?
And I would actually go out there and do the right research into what can actually help me solve those problems. In our space in particular, there are many good organizations that solve these problems. I’m a fan of many; I’m a fan of ours, but oftentimes the vertical focus is surface level. People say, I’ve got, oh, we focus on this vertical and you look at it, and they say, ” Oh, our integration is Salesforce.com or HubSpot.” They don’t actually go into the systems.
I do think that the most important thing that CX leader needs to think about today is how is my data going to flow in and out of whatever AI I’m going to bolt on because that’s going to have the most immediate impact on your employee base.
And really understanding if someone says that it does X or integrates with Y, make sure that it actually does and understand what those integrations actually mean, because that’s gonna have the biggest impact on your artificial intelligence.
Mary Beth Hamilton – 00:22:20
Excellent. Two more for you. Similar concept: what should a CX leader stop doing? And then I’m gonna say start doing. So, stop.
Malachi Threadgill – 00:22:27
If I were a CX leader trying to think about what I should stop doing, I would try, and I don’t want it to be philosophical, but it’s, you know, everything has an opportunity cost.
And what we’ve found over and over again is that a lot of the CX leaders are still using the tools that they have in place because they’ve always had them in place, and that causes some discomfort.
And we’re in a period right now where discomfort is a good thing because the landscape’s changing so quickly that I may have used this enterprise phone system for the last ten years and it’s all set up and it’s perfect.
However, the advantage of being able to implement a new solution may give you a compounded value to the business and your employees. And by being so myopic and focused on what’s working, you know, stop and pause and say, what is the opportunity? What if this actually went well? If I were to make this change and implement something new would be tremendously valuable.
Mary Beth Hamilton – 00:23:17
I love it. And it sort of leads into the: what should they start?
You may have covered it all with that, but if you have something different, I’ll take it.
Maintaining Intellectual Curiosity in AI
Malachi Threadgill – 00:23:24
No, it’s a two-parter, but I think that the intellectual curiosity that you have to have right now in this uncertain period- things are changing so quickly that having a firm understanding of where the puck is and where the puck may be headed is incredibly important.
If I’m a CX leader, one of the things that happens is: what’s the quality of the voice, text-to-speech, that’s rapidly evolving. So not just saying, okay, this vendor has X, Y, and Z, but really understanding what is 11 Labs doing? What are the frontier models doing?
What are these other things doing? And really being able to understand the landscape more broadly, what’s a local LLM versus the frontier models.
I think that’s incredibly powerful, not just for a CX leader, but anyone that is passionate about the role that AI can make because understanding the individual components can help you understand the art of what is possible moving forward.
And it’s not just for UCaaS, CCaaS, CX, AI; we’re seeing massive disruption in basically all parts of technology, software as a service.
Why am I spending so much money on Salesforce.com? It may make sense to do that, but the disruption that’s happening right now with our ability to not just bring in AI solutions, AI receptionists, etcetera, but also how do we augment the other parts of our DSSOSS, our CRM, our marketing automation platforms, and really understanding the entire ecosystem is incredibly important.
Mary Beth Hamilton – 00:24:41
Excellent, perfect. Well, we’ve hit all my questions. Anything else you would like to share that we didn’t cover?
Malachi Threadgill – 00:24:47
No, think that’s it, I appreciate the time today. I think that it’s a unique period, both of us being in marketing. I think that it’s really fascinating to see how rapidly everything is evolving on the AI front.
But what I will say is that today, this complexity is going to drive, fundamentally, I think the role of the channel partner into a new era.
I’m more excited now for this industry in particular than I’ve ever been because with this increased complexity, the partners that do embrace AI and understand the buying group and all of the implications, it actually increases the importance and the role of the trusted advisor.
Because if you can actually show the value that you add and actually help them solve and navigate this very difficult space, I think that will make the trusted advisor more sticky, more loyal, and actually help them build out the solution in a way that makes sense for them.
I think that’s the hardest thing that a lot of, in particular, small and medium businesses have is that they know they need to make a change.
They don’t know how to make the change, and that’s the opportunity that you all fulfill. And I think that this industry is going to thrive in this new era, and I’m really excited for it.
Mary Beth Hamilton – 00:25:50
Excellent. Well, we are too, and we appreciate partnerships like the one we have with GoTo.
So, thank you.
