HITT – Beyond the Bot: Positioning AI Agents Across the Business

Speaking of AI, let’s shift our focus from where our industry is headed to how you can start having these conversations with your customers today. So please welcome Sam Nelson, VP of CX and AI, and of course Brent Wilford, CX Solution Architect for today’s presentation. Sam, Brent, thank you for being with us today. The floor is yours.

Amazing. Thank you, Cass. And on behalf of the CX and AI team, thank you all so much for joining today. We are so, so, so excited.

So in true CX and AI fashion, I’m going to share my screen. We’re gonna go through some really exciting stuff today, and then mister Brent Wilford is going to be throwing up the chat. So please, please, please go ahead and put all your comments, all your questions in the chat, and we will watch from here. Okay.

Cool. So here’s the deal. Beyond the bot, positioning AI agents across the business. We are going to talk about this really exciting, exciting shift that is happening across the entire AI market and why it really matters for you.

So for the past few years or so, a lot of the AI conversations have really centered around things like ChatGPT, Copilot, chatbots, voice assistants. And while these different technologies are really important, they matter, but they’re just one part of the opportunity itself. And the bigger opportunity is actually that AI is starting to move from just something that simply answers questions to something that can actually perform work. This is where AI agents come in.

So today’s session is focused on how you can expand the conversation beyond just voice and chat to really help customers think about AI virtually across any department, workflow, or even just a business function.

The goal today, so you’re aware, is not to turn you into AI engineers. Okay? We don’t want that. Mean, hey.

If that’s your jam, go for it by all means, but that is not the goal today. The goal is to give you a really practical way just to identify opportunities, ask them better questions, and really help your clients and prospective clients connect AI into real operational outcomes. So let’s set the stage here. The AI conversation has changed really, really quickly.

Just a couple years ago, most of your clients were asking what AI was. Should we be paying attention to things like ChatGPT? Well, then the market moved into this experimentation phase, and customers started launching pilots randomly. They were testing co pilots.

They were playing with content generation, and they really started exploring chatbot or voice bot use cases. But now we’re actually entering a new phase.

So customers are no longer just asking whether they should use AI. Now they’re asking where AI can actually create measurable value. They want to know where AI can reduce a lot of the manual work, where they can improve productivity, where they can lower costs, maybe speed up some of the mundane processes, and then, of course, overall improve the customer or employee experiences. This is a really, really important shift to recognize.

Because when the conversation was just about primarily experimentation where customers were just trying tools, the conversation now moves to execution. In other words, customers need strategy. They need architecture, governance, integrations. They need people who understand the broader technology ecosystem.

And that is where you have the opportunity to expand from there. Traditionally, what has just been about voice and chatbots. Now one of the biggest challenges I see with most AI conversations is that they’re still very, very narrow.

They typically stop at the chatbot. They stop at the copilot. We all talk about meeting summaries now. We talk about content generation.

And those are still legitimate use cases, but they are not the entire story. So if we think about where work actually happens in the business, most of it does not actually happen in a chatbot. Now, Graeme said, it does happen when it comes to communicating with customers externally or maybe the chatbot deflecting a lot of the interactions. That way humans can focus on more complex interactions. But we forget about the work that happens in departments like HR or finance, even IT. Yes.

Or how about operations or procurement or there’s compliance? How about how about this one? How about sales and even customer success? We hold some money bags. Right? Let’s be real. It’s typically sales and marketing, these revenue generating departments.

And across all these, you’re gonna find a lot of repetitive tasks. You’re gonna see manual processes, sort of disconnected systems, employees spending time searching for a lot of information. And this is actually the bigger AI opportunity.

So the message for customers is not just, hey. AI can help your employee or customer experience. The message is actually AI can help your organization rethink how work gets done, And that’s a huge conversation. Now I’m looking in the chat real quick.

Wanna take a quick pause. So customers, yes, from mister Thomas Cross. Customers are also focused on ROI, TCO, and token cost. I’m so glad you brought that up.

Like, give me your Venmo, and I’ll I’ll shoot you a Venmo later, because we’re actually gonna talk about that, towards the end of this conversation. So thank you. Thank you. Thank you for bringing that up.

Okay. So let’s start from the beginning. What is an AI agent? Let’s define AI agents in simple terms.

Now before you all start asking, can I get access to the slides? Absolutely. Don’t you worry. We’ll have a copy of the recording.

I will send out a copy of this deck. You can use it to your heart’s content. Because I actually think that these slides help in describing what these things are to customers because a lot of times, they do not understand what this stuff is, and they’re depending on you to explain it to them.

So the easiest way to think about an AI agent is more like a digital employee or a digital worker, so to speak. A chatbot will typically answer questions, but an AI agent will take action. And it can understand a request. It can retrieve information.

It can start making decisions based on rules that you implement. It can trigger workflows. It can update systems. It can really coordinate kind of activity across multiple applications, which is why it’s really, really cool.

So, for example, a chatbot on one hand might say, here’s the process for submitting an expense report. Go do it. Now the AI agent, on the other hand, might actually pull up the expense policy. It might check the receipt that was submitted.

It could validate the category that it belongs to. It can route it for approval, and it can update the system. That’s the difference. And this is, again, one of the biggest shifts in AI right now is we’re moving from AI that just assists people to kind of AI that participates in the actual business process.

And that opens up a huge, huge set of use cases for all of you to explore with your customers. So AI, right, where you can expand the conversation, AI is not just limited to customer service or even just to sales. In the front office, AI can actually help with things like sales follow-up, marketing personalization, think of staying on brand, customer service automation, customer success insights, right, looking at churn rates, things of that nature, account research. Then there’s this middle office where AI can do things like support operations, think supply chain, logistics, procurement, project management.

Thanks for the Venmo, Tom.

And in the back office, AI can assist with things like HR, finance, IT compliance. We talked a little bit about that. But think of all of these administrative functions. Now the common thread is not the department.

The common thread is actually the work. So every single department has repetitive work. You all have it. We all have it.

And every department has information trapped in multiple systems, and every department has processes that are typically slower than what they really should be. And this is where AI agents become super, super relevant. And for you as tech advisers, the opportunity is to stop thinking about AI as just a product category. Think about AI as more of a business capability that can show up across the entire enterprise.

Yep. And I love all the comments I’m seeing in here. I’m seeing AI from marketing all day long. Absolutely.

Mister Wilford made a great point. Outbound sales is a growing segment for us. Absolutely. And, yeah, a great point, Harry.

Like, the problem with most AI chatbots is poor implementation. A hundred percent that we gotta he said it. Shut up and listen. So, anyhow, what do you look for?

So when you’re talking with customers, there are kind of four signals that should immediately get your attention. The first one is repetitive tasks. If employees are doing the same thing over and over again, that is a potential AI opportunity.

The second is around manual processes. So if the work is being moved, for example, through spreadsheets, email threads, there’s a lot of copying and pasting, or maybe there’s a lot of manual approvals, Those are all really great signals to address. The third one is information searching. Now if if someone is anybody, right, an executive, a frontline worker, whoever, if anybody is spending way too much time looking for answers across multiple systems or folders or applications or just knowledge bases, AI may be able to help with that. And then last but certainly not least is workflow delays. So if a process is sitting anywhere, whether that’s waiting for approval or maybe it’s waiting on routing or a response or there’s a handoff or there’s just some kind of delay within a workflow, that’s potentially something AI can help address.

So these four particular signals are really easy to listen for in a customer conversation. Again, repetition, manual work, information searching, and delays. When you hear any of these four, you likely have an area of the business to dig deeper into. Now I know you’re gonna ask me, what do I ask? So one mistake that a lot of advisers are asking that I’ve been witnessing is simply, what are you doing with AI?

And a lot of customers don’t know how to address that. Because once AI becomes normalized in a business, a lot of them don’t actually think of it as AI. They think of it as a normal piece of the business process.

It’s not always the best question because sometimes they don’t even know what’s possible just quite yet. And a much better approach is to start with business friction. Ask them what’s taking a long time? Where are employees frustrated? What work is repetitive? You get the idea.

And while these questions are really, really simple, they can be really, really powerful. And they’re actually helping your clients talk more about the business problems rather than the products that they’ve seen or the products that potentially might talk about. Right? Have them talk about the business problems.

And this really matters because the problems are what create the urgency, the budget, the executive attention. And it it basically uncovers what matters right now. And your role then is to help translate those business problems into technology opportunities. So instead of leading with AI, lead with the workflow.

Now we’re gonna go into this slide where I’m basically just reinforcing that AI is not one department. It’s essentially an AI I’m sorry. It’s an enterprise capability. So the opportunity might start in one place, but it’s going to expand everywhere. So for example, it might start in just customer service or in CX or ECX, employee or customer experience.

And it can quickly expand into things like sales, marketing, HR, finance, IT, like, you name it.

Think about how many processes cross all of these different departments. A customer issue might turn into a support billing operations, account management situation. Maybe an employee onboarding process involves HR, certainly involves IT. Maybe it involves finance, security, facilities, right, access.

Something in procurement, it might involve operations. I’m thinking probably legal, finance, probably a compliance component there. So all of these workflows, you have to understand are cross functional. So I’m trying to make sure that you see the bigger picture here is that AI agents can become really, really valuable because they help orchestrate activity across all of these different departments and their associated systems.

So that means for you is that one AI conversation can actually open the door into multiple parts of the business, which is also why we’re starting to see more and more decision makers. So also important to understand that customers do not actually buy AI. This is one of the most important points of this presentation. So there’s one thing you take, folks, take this.

They do not buy artificial intelligence. They do not buy language models. They do not buy machine learning. They buy outcomes.

They buy these things, faster processes, higher productivity, lower cost. They buy better overall decisions. AI is just the mechanism for these outcomes, for achieving these outcomes. Now if the conversation becomes way too technical too quickly, you’re gonna lose the business audience.

I will tell you that immediately. I’ve seen it time and time again. We witness we witness this weekly, almost daily. So when you’re positioning AI, focus less on the model and the features and the capabilities and more on the measurable business results.

In other words, what changes for the business? What gets faster? What gets cheaper? What gets easier?

What improves specifically for the employees and for their customers? And that is truly where the value lives. Let us handle some of that techie stuff. Right?

We’ve got suppliers for all that stuff. Don’t worry about it. Focus on the business value.

Now let’s talk about this new challenge that I am so glad that mister Tom brought brought up earlier around tokens. So let’s imagine the customer is really successful. They deploy HR everywhere. They deploy they deploy AI in HR and finance and sales and IT and offer like, you name it.

Anywhere. It’s fantastic. But we now have this new challenge in the business, in the industry, and every single AI interaction consumes resources. So for those of you who are not familiar, every prompt, every workflow, every action that an AI agent takes, every you think of every AI API call, you think a retrieval step, quite literally every generated response consumes what we call tokens.

Yes. Absolutely. For those of you, like to throw in a little story here, but you go to Chuck E. Cheese, right, as a kid and you’re using tokens to play games.

Folks, it is the same type of idea. Right? And then you turn in your tickets to get prizes, and there you go. It’s like exchange, exchange, exchange.

So at a small scale, this yes. Like an arcade in the eighties. Thank you. At a small scale, this may not feel like a major issue.

But as the business grows, as it starts using more artificial intelligence capabilities with hundreds or thousands of employees with many AI agents operating across multiple departments, token consumption can be a real cost management issue. We have seen the articles come out. So as AI adoption grows, the customers have to think not only about how to deploy AI, but how to manage it. So that actually creates a brand new conversation around cost, performance, governance, even visibility, as well as control.

So I am very, very excited to talk to you now about how has a brand new meaning.

Most of you know TEM as technology expense management or telecom expense management. Now I am. You heard it here on this call, and it is being recorded. I am recoining the word now, the term, the the acronym right now. You heard it here first. AI introduces this new version of TEM, which I am coining as token expense management.

So, again, tokens are essentially this unit of consumption in many AI models. And every time a user, again, submits a prompt and the AI model processes information, it generates a response, Tokens are being consumed, and those tokens have a cost. Now as AI becomes more embedded into workflows, organizations have to manage the spend. In other words, who is tracking token usage, which maybe which departments are using the most?

Which workflows are driving the highest cost? Are they using the right model for the task? Which I’ll get into in a bit. And maybe they’re overpaying for really simple requests.

This is a very natural evolution of your conversations with your customers. You’ve helped them think about telecom spend, cloud spend, software spend, but now AI spend is becoming part of that same kind of operational discipline.

And great questions in the chat. There are so many, so I’m gonna pause here.

Right. So lots of studies comparing the cost of an employee to the cost of a digital employee. Absolutely. We actually have a lot of suppliers folks, and also can who who can get into that.

Estimate the average cost of a token in the client’s environment. That is a tough one, because you have to look at overall consumption, and then you have to break it down by what exactly like, work your way backwards and break it down to what exactly they’re doing and then associate costs there. And different things may cost a different number of tokens based on the complexity of the particular task, and that can be really overwhelming. Now the good news is we do have suppliers who play in this space now.

Yes. You will see some announcements over the next couple of weeks around actual AI suppliers who offer exactly this token expense management. In other words, it will assess the current environment. It will show token consumption across multiple tasks, multiple departments, and help companies further optimize token usage.

Now one thing out there that you will encounter is that, yes, we have Claude or we have ChatGPT, and we have a limit that we set to that. Well, that’s great that you have a limit for token expenses. However, you still don’t know if you’re actually optimizing all of your token usage. Right?

It’s not a matter of how much. Now it’s a matter of how are you actually allocating that token usage across the business. But look at it this way. Okay?

I’m gonna move on. But, again, once you start managing token costs, the next question here then becomes, how are you making sure that every single AI request is actually using the right model in the first place? And this is where the idea of an AI gateway or, like, data AI orchestration, this layer, this orchestration layer comes in. So instead of connecting every single AI workflow directly to one model provider, organizations can actually use an intelligent orchestration layer.

Now that layer can help decide which model should handle which requests. So it can optimize for cost, speed, accuracy. It can support governance requirements. It can provide flexibility as you have new models emerging.

By the way, you never hear about when they’re going to emerge. They just show up. So this is really important because the AI market is changing incredibly fast. And the best model today may not be the best model six months from now.

In fact, it may not be the next model, the next best model in the next hour. So customers do not want to have to rebuild their entire AI strategy every time the market makes a move. So the value of orchestration is actually flexibility, and it gives customers this way to manage AI just more intelligently as the landscape continues to evolve.

Now what does successful AI really require? And there’s a reason I have one graphic on here and one graphic only.

Successful enterprise artificial intelligence requires an ecosystem. It requires all of these different components. Now look at it this way, folks. Again, the data garbage in is garbage out. It requires quality data because AI is only as useful as the information it has access to.

Yes. It also requires integrations because AI has to interact with the systems where the work actually happens. And it does require workflow automation because the goal is not just to answer questions, but to actually complete the tasks. And then don’t forget folks, there’s the security and governance component because organizations need to control access. They have to protect the data and they have to manage the risk.

And then the AI orchestration piece because customers are likely going to use multiple models over time as they evolve. And then the last component, but certainly not the least important, is the token expense management piece, ten. Because AI consumption needs to be visible and controlled. So this is a bigger sort of strategic message. It’s it works great in front of clients because AI is becoming this sort of enterprise architecture conversation. That’s why you all have a very, very important role to play.

So why does this all matter? Because AI has the potential to expand almost every single technology conversation that you’re having today. It might lead with just automation, but here’s the deal. It can lead to things like contact center or UC modernization.

It can lead to cybersecurity discussions or even cloud infrastructure or networking or data governance or, heck, even managed services, professional services, you name it. AI is not the product. In fact, AI is your catalyst. And if you are using this smart, you will have a great reason to reengage customers around business outcomes and overall technology strategy.

Show them this, and it will spark all kinds of conversations. I promise. Because AI touches the systems, the data, the security, the workflows, the communications. So you are very uniquely positioned today to help your clients and prospects think through a broader picture.

And this is where you all move from solution selling to being more of a strategic guide. So here’s the deal. I’m gonna wrap up with a few key takeaways, and then I’m gonna catch up in the chat.

But first, right, AI is moving beyond just the chatbots and the voice assistants. Two, agents are becoming more digital workers that actually perform tasks. They retrieve information. They coordinate the workflows.

Third, every department contains AI opportunities. Don’t leave anyone out.

Fourth, customers buy outcomes. They do not buy models. They do not buy features, capabilities. Think of the business value, the business outcomes.

Fifth, AI requires orchestration, governance, integrations, you name it. Don’t forget about the security. And, of course, TAM, the new cost management piece. And then finally, I’d put this one in here, but tomorrow’s advisers are not just going to sell AI.

They have to figure out how to help customers actually operationalize it, and that’s a future opportunity, folks. It’s not just AI for customer service. It’s not just AI for chat or for voice. The future opportunity is essentially AI for anything.

So, I’m going to go in the chat here. There’s a lot going on, a lot of great questions. And don’t worry. I will make sure that this deck makes it out to all of you.

In fact, I’m actually gonna put this deck into the AI Launchpad. If you do not have that link, we’ll pop it into the chat for you. But the AI Launchpad, we’ve got a CX Launchpad. I will make sure that this deck is available to you on that page.

So if you don’t get the email, at least you’ll get the landing page where this will live. So lots and lots of exciting things to talk about.

Let’s see here. I’m looking through the chat. Mister Brent Wolfer’s got my back. He is chiming in.

So, yeah, of course, if you have any additional questions, feel free to reach out to us directly. We are here to help guide these conversations. The beauty of something like artificial intelligence, and working in it directly here from Telarus is that we all have had to go really deep into this topic in all of our respective practice areas. So when you reach out to anyone on our team, whether that’s my team, Advanced Solutions, or you reach out to solution engineering, we all have an understanding of generally how to approach these AI conversations.

And then we can engage the appropriate resources to go deeper and, of course, include the appropriate suppliers as needed. So with that, I am going to stop sharing. Chandler’s going to take over the screen share again. And, Cass, I am going to bring you back on and pass off to you.