Scott Kinka:
I had a note that I took down, and you could see it here in our show notes that you made during our pre-call, which I loved. You said we’re spending more time making humans better than replacing them or taking it the other way around, making it a little bit more of an active thing. We’re replacing humans versus making them better kind of things. So talk to that a little bit.
Milind Pansare:
Yeah, I mean, the analogy I give is, how many of us took a completely self-driving car from the airport? None right here from, say if you came from Palm Springs, but everyone was in a car where the human was driving better because they had Lane keep assist, they had collision avoidance. So, a lot of technologies are easier for people to adopt. So I’ll give you the example of, let’s say, Cox Automotive. All of these brands, Kelly Blue Book Autotrader, where they are right now, are trying to solve a very specific problem, which is, Hey, can you make the QA process in our contact centers automated? Because we have very few QA people in terms of ratio to the agents. We have massive contact centers. Then you go to the world’s largest food service delivery app that we all use, and they have different problems. It’s when you have an escalation, and someone’s dissatisfied. They can make an inordinate amount of noise. It can escalate all the way up to your CEO; it can go into the newscast, and it can go viral pretty much. So how do you detect those with sentiment and nip those in the bud, and you go, alright, here’s a $5 quick coupon that you can apply to your next conversation? So, there are different kinds of insights, and different companies are at a particular point in solving different workflows. As I often say, AI by itself solves no problems. It’s when you combine it with workflows and other systems of record in your organization, as you’ve rightly pointed out here, that you can really make a difference and show some ROI for the customer.
Enhancing Human Performance with AI
Scott Kinka:
A hundred percent. Actually, I’m going to take a step back from this question. I’m going to go off-script. I know you’re shocked. Part of what you just said that was super interesting to me is we spent a lot of time in the opening session yesterday talking about the changing role of the IT leader in the business, and we had this graphic around the exponential IT curve that effectively said over the last 20 years we’ve moved from kind of plumbers to strategic infrastructure people to now partners in the business. And what was really interesting about I think what all of you said is the solution that you delivered certainly had it in it, but in every one of those cases, the IT leader has to partner with someone in the business. It doesn’t own sales, they don’t own support, they don’t own the employee experience. Walk us through that journey a little bit. I mean, you all have CO relationships going back decades, right? Is this a scenario where the folks in this room, whether it’s supplier or consultant strategists, need to help the CIO potentially bridge some gaps into some other areas of the business to be successful with an AI project?
Kristy Thomas:
Absolutely. A hundred percent. One of the key value propositions that you bring into these organizations today is that no organizations have internal friction with personalities, and nobody has their own personal agenda for protecting their budget and their fiefdom. So the reality is that you bring this force because you don’t have that. You’re agnostic in terms of what’s best for the CMO or whatever. And so you can get access to on behalf of the CIO entry point into six to 12 stakeholders on average in these big initiatives, right? So what’s important to the CMO is completely different from the CFO, but they all have to be aligned from a steering committee on what the AI strategy is because it impacts the organization at a top line. And so the value that you bring is being that concierge, bringing that liaison on behalf of the entire organization, collecting all these different use cases with the stakeholders, and mapping that together to show that everybody’s benefiting in different ways. And don’t forget that you have to kiss the CFO’s ring, so do that early because what you want to do is kiss that ring early.
Scott Kinka:
That never happens here. It just never happens.
Kristy Thomas:
Oh, did I just set you up?
Scott Kinka:
You’re totally good. Bo has his face down, so you have to kiss the CFOs. I said it never happens around here.
Kristy Thomas:
Yeah, noted. Noted. But it’s real. I mean, and my experience selling over the years is that I’ve been caught off guard. The last thing we do is go to the CFO, and now your initiative’s killed because he or she wasn’t involved in it.
Scott Kinka:
Yeah. Whose budget is the AI project to improve employee experience? The best question, right? Is it an HR initiative? Is it an IT initiative? Who owns it? Even if you can get to the ROI complicated. Does anybody else have anything to add to this role, not change?
Nick Slater:
Yeah, I think it presents an opportunity. I think CIOs don’t want to be the sole decision maker, and we have an opportunity to really make champions out of the chief customer officer for their contact center for the CRO for sales intelligence. And that’s why it’s so critical to be able to offer the full end-to-end platform. Companies don’t want to have to go buy 5, 6, 7 solutions. In fact, they’re doing the opposite via digital transformation. And so if you have the opportunity to talk to these other peers of the CIO, create anchor points and champions and those folks, I think really what it does is create a buying safety net for the CIO who doesn’t want to have to stick their neck out solely. Now, they have these champions supporting them. It makes it a much easier conversation with the CFO, and that tends to be the way that you can have budgetary influences. Let’s go get a few of the CFO’s peers who are now raving fans of this platform solution, and it becomes a no-brainer.
Scott Kinka:
Interesting to think of it as not only a bridge that needs to be made but a way to share the responsibility. Absolutely. Right here. Shawn, anything to add?
Shawn O’Connell:
Yeah, it’s a cross-functional sale, right? It’s land and expand. You have to have a champion somewhere in the business that’s got a set of use cases that are tied to a problem that they’re willing to spend money on. They all have the budget for it. It may start there, but if you don’t have that champion, they’re not working, and you can’t build that relationship cross-functionally. It’s a very difficult sale.
Scott Kinka:
And I think an opportunity for the consultants in the room to expand their influence inside of the organizations, right?
Kristy Thomas:
The one thing, too, you got it, Scott. I’ll also add that while you’re figuring out your mic if we’re talking about land and expanding wider stakeholders, let’s also remember that we are in an era of an informed buyer. All of these individuals are informed. Your value that brings is when they go to make a decision. You’re validating that decision that they’re making across all of them because you’re externally saying, I know this technology. I can validate it because I have seen it deployed in X, Y, and Z as well. So your credibility with this engagement is one plus one, which equals three for them. So you’re getting the liaison stakeholders, but you’re also validating them from the outside, saying this is a great decision that you’re making regarding your purchase.
Scott Kinka:
Alright, great points. And again, a sidebar. I like that. Let’s talk about the way customers want to engage. I think one of the big, almost oxymoronic things that’s occurring right now as we’re talking about AI is we’re coming out of the era of high touch in COVID. Our expectations of how we engage with an organization increased. We wanted the agent to have better information. We wanted to have a more personal relationship as we went through there. And now we’re kind of entering the era of AI. And a lot of times, the first inclination is to gut that process and slap some bots in there. That makes a lot of sense, right? Particularly when the customer has been in a position of wanting, knowing and expecting a deeper relationship. So, I think you shared this slide with us, right? I’m going to let you speak to this first. The idea here is presence versus preferences or personalization over human touch. What is the end user really expecting? And do these things need to necessarily be in conflict with each other, I should say?
Customer Expectations: Personalization vs. Automation
Nick Slater:
Yeah, absolutely. So this is a study by McKinsey, and I think more and more companies are starting to come to the same conclusion, which is buyers’ customers are more than willing to do asynchronous work, watch a video, read a handout to get informed and educated on how to use a solution. But when they run into trouble, they want to speak to a person, not a virtual agent, not a chat chatbot, an actual person. And what this data shows for those that have trouble seeing is actually true all the way through millennials and Gen Z. And so, at Dialpad, what we have really strived to do is see these agents as heroes rather than cogs in a wheel. How do we supercharge their ability knowing that their prospects and their customers prefer to problem solve by speaking human to human? It’s now our job to make them more enabled and more informed to make that interaction more delightful. We’re not here to replace individuals; we’re not here to see them as cogs in a wheel. We’re here to see these agents, these support teams, for the value they provide and make it a much more effective and pleasant interaction. When do folks run into these problems?
Scott Kinka:
How often are you working with the customer, though back from that expectation? So take that one. We’re going to do AI. How many people can we whack? Are you walking customers back from that expectation into the explanation that Chris just gave you guys? Do you guys run into that kind of idea of going for the firing first? I hate to be blunt about it, but I would prefer to go for the staff reduction first versus focusing on the effective improvement of humanity. Are you walking people back from that expectation?
Kristy Thomas:
Absolutely. Was this a question for all, or just me?
Scott Kinka:
Fire away.
Kristy Thomas:
Okay. I would say absolutely. I think now what we’re seeing is that although that was the initial conversation about starting it, and there have been debates around it in terms of AI replacing humans, it’s absolutely not the use case for AI. But what I would say to piggyback on this personalization thing, too, is that when we step back and look at the marketplace today, there are five generations of buyers, five generations that want to buy differently. And so when they do need to reach out, they want to reach out differently. But the reality is when any of us have to pick up the phone, what is the one thing we want to do? Is our objective? Get off the phone as fast as possible. Correct, correct. Do I get a heck? Yeah. Okay. So that’s what we want. And then you have the other person on the other line that wants to do that too, but they don’t have the tools to do that. It’s not that they don’t want to help you, it’s not that they’re a mean bitch, it’s because they don’t have the tools to do their job. So there’s this inflection that happens where technology powers the human to be able to do that, to be able to connect the two of ’em, to get Brian off the phone as fast as possible because he’s got stuff to do and to help Mike do his job because he doesn’t want to quit. And you can try and churn, too. So there’s all of these different components. It’s not just eliminating humans.
Shawn O’Connell:
You want to arm the agent with the information they need so they can better serve the person requesting it. If you can weaponize that agent, you’re going to make the experience much better. So give ’em access to all the data they need, so they can respond accordingly. And to your point, get on and off the call quickly. The longer they’re on the call, the more it costs that company anyway. And frankly, the more upset the customer is, right?
Milind Pansare:
Yeah, absolutely. I think, Kristy, you hit on a very important point, which is people just want to resolve their issue unless they’re very lonely, in which case they want to talk to you for a long time. But one of the issues that’s happened is that we’ve had many failed technologies. So we’ve had these horrendous IVR technologies in the past, and for all of us who have probably created some such technology, we owe the world an apology because we messed it up in the past. These things were frustration additions and a step to stop the customer from actually reaching a human as opposed to solving their problem. So what’s really important now is you have the data in all the conversations so you can deflect the right calls. It’s like, is my package there? How far away is it? Well, that might be something I can just see on my chatbot. Hey, I don’t like the texture of this; I want to return it for this reason. That might be a quick call with a natural-sounding voice bot. And then maybe a more complex thing, which is I’m not quite sure which specialist I should be seeing for this condition that my grandma has now that’s an assisted human call where all the information is coming out and is on the agent’s screen in real-time. And then there might be other things that you solve after the call, which is, Hey, let me coach my agents on how this cohort is performing better at selling than that cohort of people in this contact center.
Scott Kinka:
So, it is right for the right job at the end of the day. Alright, let’s take a look at this. There’s a lot of data on here. I would just want to focus on the right. Only 34% of IT decision-makers feel that their organization currently has the right data and technology in place to enable effective AI. So we touched on this a little bit earlier, and I think most people have heard the Air Canada story as an example, but very regularly, we meet with clients who say, I would love to automate this process. And the first question is, can we see the process? Oh, well, we train over the shoulder. Well, that’s going to be a little bit of a problem. How often do you have that you are not ready, or do we need to help you get ready before we can engage in part of the conversation? Shawn, I’ll go to you first.
Shawn O’Connell:
Yeah, I think we would like to start with whether it is a people process or a technology problem. Oftentimes, it’s a process. It’s not necessarily a people problem, so you’ve got to dig into that process. But if you’re leveraging AI properly, you’re tying into multiple disparate data sources, pardon me, so that you can give access to get information to better serve the customer and circumvent some of those process issues that may exist.
Scott Kinka:
Gotcha. Anybody else wants to add to this one?
Nick Slater:
I would also say that beyond having the right data structured in the right way when we talk about things like chatbots, you can have the best chatbot in the world, but it is pointed at a knowledge repository. And so when we talk about you being ready for it, it’s not just governance, it’s not just your data, but really making sure your help center that is powering your agents is up to date with the most recent product names and prices and responses. AI is only as good as the information that you pointed out. And I think that’s kind of a panic moment for some customers that realize, oh, we’ve got some work to do to really get our house in order as far as our knowledge base goes, but that is a must before you plug in this kind of powerful AI solution.
Shawn O’Connell:
When you go through the testing and piloting phase, it really is eye-opening. So, depending on what the use cases are, it could be something as simple as your hours of operation. It starts to run against that, and it surfaces at incorrect hours. And what inevitably comes back down to is somebody on the marketing side or somebody somewhere hasn’t updated that, and they’re like, oh, we’ve got the wrong information, we better go fix that. And it’s shocking just how much information sits out there isn’t accurate, how quickly you surface that, and how quickly they need to be able to go and repair it.
Scott Kinka:
Am I thinking completely outlandishly to think that we could potentially use AI for the businesses that are operating over the shoulder, right, could they begin to develop the content for the knowledge base that informs the AI so long as it’s checked right and filled in their gaps through their call recordings, their call logs, all of that? Why don’t you talk to that a little bit?
Governance Challenges and Creating AI Policies
Milind Pansare:
Yeah, I would say the answer to that is yes and no in the sense that, let’s take the example of working with a very large BPO and one of their customers making coffee pods and coffee machines. So yes, your knowledge base has the six-step way to describe the machine, but when your agents are on the phone, can you summarize each of those six steps? Are you going to use something like a chat GPT and just say, oh yeah, just go ahead and clean it. It’s like a Shakespearean sonnet, which doesn’t work, right? You need those six steps, and now you have to summarize them. So it’s very nuanced on how much generative AI can solve your problem and how much structure you need to put around it to ensure that you get precise actionable information.
Scott Kinka:
Makes sense. I’m just going to, in the interest of time, jump into these last couple of topics we’ve talked a lot about, and we’ll continue to talk a lot about some of the issues around AI governance and some of that here. Accuracy, bias, privacy, the things that are there. There are risks across the same CIO survey we were looking at earlier. One in three CIOs says there are no governance steps in place in the business today around AI, and they’re turning on Copilot and Office 365 or they’re turning on these tools. Obviously, governance generally is first. So the question here is, as you’re engaging with customers, we’ve talked about the idea of getting after ROI about going across the aisle to the business leaders. How often, though, are you stepping into a business that wants to use your tools and literally does not even have an AI policy yet? Have you had to walk customers back on that topic?
Shawn O’Connell:
All the time. It’s the number one thing that happens. Legal gets involved, it isn’t a technology issue, it’s a fear of the unknown. They’ve all heard the horror stories, they’ve all seen the stuff in the press, they look at it, and they begin to peel the onion back and go, we have no policy or procedure in place. What if something goes wrong? And so you end up having to walk them back through the process that you’re operating with inside their firewall. So if there is data that exists inside their firewall that could be offensive, they have a bigger problem that they need to address. So, we spend an inordinate amount of time educating not only the technical teams and security teams but also the actual legal teams. We get on calls, and we sometimes have to have two or three calls just to help them understand how the technology operates.
Scott Kinka:
Just simply stated for our team members in the room, what are the kinds of things that are addressed in an AI governance policy? What should they have? Even if it’s one sheet of paper, what should be on that sheet of paper?
Nick Slater:
Yeah, I would say, do we want this to be an all or a solution? So there are some solutions that are either on or off for everybody. There are other solutions that allow you to throttle. Some groups may have no access to AI, and others may have it fully, and you can tailor that by region and function. That gives a lot of security. Another one is whether we are okay with data leaving our instance, going out to a third party, and coming back with solutions. That’s for some companies, that is a full-stop security risk. Others may be okay with it for certain groups. So, coaching them through what data are you comfortable with, leaving your platform and coming back? Do you want a solution where it stays fully contained in these four walls? Those kinds of security things, I think, are headlines in that one-pager.
Scott Kinka:
Did Chris miss anything in there?
Kristy Thomas:
And I think we’re going to see the constant evolution of the technologies in today’s marketplace around this data discussion wherein they’re stuck to morphing themselves in vertically specific ways. I think that’s important. And then there’ll be governors or lanes on what you have access to: the knowledge base for X and open API for Y.
Scott Kinka:
I mean, certainly, it’s not a one-page document if it’s a compliance organization. I threw down the one page, but that’s the key. When I talk to customers, I just say, at the very least, put your intentions down on paper and make sure everybody has it. You can’t control behaviour all the time, but in many ways, it’s very much like a corporate information security policy. You don’t intend to release this information. You intend them to use their PC in a specific way. At the very least, that’s got to be on paper someplace.
Industry-Specific AI Applications and Practical Use Cases
Shawn O’Connell:
Industry plays a big factor, right? Healthcare and banking are highly regulated. Security is obviously of utmost importance. And so they’re going to have different requirements than, say, a traditional e-retail or something of that nature where you want to give access to contracts, things like that. You’re just not going to be able to do that in the healthcare space.
Scott Kinka:
I mean, Milind, you’re at the core of the call recording and transcription business. You’ve got customers who aren’t even able to store data in some cases that are in those recordings at rest. Talk a little bit about that.
Milind Pansare:
You have to redact stuff. You have to get rid of personally identifiable information. So all of those things are going on. Those are your traditional trust and security things. Those haven’t gone away. But where we are seeing a change, even in the year and a half since LLMs became a household name, is that our InfoSec process with it is far more mature now in that we are not just talking about where my data sits at rest. How is it being transported? Is it going over to AI servers? Where are your servers located? Do you have a specific cluster for HIPAA customers? So, where are you doing the computing? What measures do you have in place to check for AI bias and drift? That’s a real thing, as you train it with bad data. What do you have in the process to check for that? What do you have in the process of checking for ethics? So basically, in the past, you had an InfoSec document. Now, we have to get ahead of that with an entirely sophisticated trust console. Depending on the size of the organization, IT organizations are now doing this. And they’ve settled down. About a year ago, there were these ad hoc AI committees because nobody knew what to check for. Fast forward to today, and it knows how to do this, they are running through a checklist. So it is fascinating to see that. And then, as vendors and consultants in the sales process, you have to be ahead of that and bring them in early because otherwise, it’s just going to shut the door on that sale.
Scott Kinka:
One of the comments I just want to pull out of that great answer, there was a slide I had in here that we removed because the data was a year old, but it was a survey of CIOs around who owns the AI governance document in the business. At the time, this was a year ago, and it’s since moved, but two out of three CIOs said they weren’t or they didn’t know. Suppose they were responsible for the AI governance policy inside the business, which is really interesting. I want to get to one of the last macro-level topics. I’m going to skip ahead. We’ve talked a lot about helping customers outside of just consuming technology, needing to get around governance, needing to find the use case, and needing to help the IT leader bridge across the aisle into other peers in the business. And let’s be very clear, you are all platform companies up here. So while you’re happy to engage in those conversations respectfully, you’d probably be thrilled if you didn’t have to also, right at the end of the day, so this slide talks a little bit about kind of this channel, the consulting channel, and there is quite a bit of skepticism, at least if you believe this data from Canales on here, that there’s a monetization opportunity for the IT consulting industry around ai, right? So we’ve just talked about all the challenges and all the pre-work. So what I would love for all of you to just take a quick shot at doing is share with us what we can do to help you make good customers at the end of the day, talk a little bit about integrating there, and I’ll start, and we’ll just go right along the road. Fire away, Chris.
Nick Slater:
Yeah, Nick. Chris is my full-head-of-hair alter ego.
Scott Kinka:
Why did I do that? My apologies.
Nick Slater:
No, all good. So I think one thing is to give us the opportunity to work with you before we’re even talking to prospects and customers. Let us help you learn how we are tackling AI at Dialpad. We have almost 20 PhDs in AI. We have 70 employees, and their full-time job is nothing but to focus on AI. And we really want to be a resource to you all. So you can come in and feel confident and informed about what’s real and what’s not about AI and how we can help our customers drive their business outcomes that are tied to their stated strategic initiatives. I think all too often, we get engaged when there’s an opportunity. There’s a prospect, there’s a customer, there’s an opportunity for us to work one-to-one in preparation for a lot of really meaningful prospects and customers. We are happy to help advise, consult and teach what we’re seeing in the AI landscape and how we can partner together to be really effective.
Scott Kinka:
Great answer. Milind.
Milind Pansare:
Hey, I’ll spell it down in three steps. One, the money’s there. If you look at contact centers, there are 17 million agents. What did we spend money on per agent per month before? It was CRM and workforce management. What money is going to go in? It’s an AI, so let’s be ready for it. It’s there, it’s happening. Two, the customers are here. This is not a tomorrow and day after technology. We have 350 customers all the way from mid-enterprise to large enterprises. So make sure you’re talking in the language of the customer in the industry they’re in. And three partners are with us, just like you said. I’ll give a shout-out to two of my colleagues who are here. Jim 10 is our VP of channels. Go find him, talk to him, talk to Scott Eastman, who’s here as well, and let’s work together and solve these problems.
Scott Kinka:
Shawn?
Shawn O’Connell:
Yeah, I echo both Nick and Milan just said, right, we’re here to partner with you. We’re here to help educate you. Let us spend some time helping you understand what the opportunity looks like, why it’s important for you and your customers, and how you will capture some of that market share. We can arm you. We’ve got an army of resources ourselves, and I do think that we can partner together and drive a nice chunk of revenue.
Scott Kinka:
Love it. Kristy.
Kristy Thomas:
I’d say we don’t want to be everything to everyone. So what I would love to do is collaborate with you and understand what the ideal customer profile is for our solution. Then, once we tier down from there, what are the use cases specifically around that ideal customer profile? Learn the use cases because that’s the best way to start a conversation with a customer. And then also a couple of our win wires, those three things, 1, 2, 3, get you into a conversation for the right fit for us, and it doesn’t waste time mutually and in partnership.
Scott Kinka:
I love it. So we’re going to wrap up. First off, before we get into the wrap-up here, can we thank these guys for bearing with me and my voice here? Make sure you see them when we’re around today. Nick, Milind, Shawn and Kristy, great job by these guys today. And in terms of the customers who are watching this at home, we’ve got a great framework around getting there from here around AI. We’re sort of out of time on the panel, so I won’t read all this, but there are some good tips and tricks here. We are to help you define and execute that strategy, whether it’s consulting, running workshops at your business, supporting you on advisory or engaging in guided implementation along with these great suppliers that we have here and AI innovators, frankly, we have here on the panel. So we’re thrilled. Please give our panel one more round of applause, and for those of you at home, we will see you soon on another episode of the Bridgecast. Thanks.
About the Author
