Episode Transcript
[00:00:22] Speaker A: Hello everybody and welcome to this episode of metrogy's Metro site Vidcast podcast.
And we're gonna have a great discussion today because I'm super excited to be talking to Vinod Muthur Krishnan who I've known for a number of years and he's just one of the smartest people when it comes to just being able to understand and explain the value of AI in so many of the things that we do in our industry. So first of all, welcome Vinod.
[00:00:52] Speaker B: Thank you.
[00:00:53] Speaker A: Glad to have you here.
So Vinod, by the way, is the VP and chief operating officer right now for Cisco WebEx customer experience.
And I just want to start Vinod. Just give us a little bit of your background just so people know where you've been, where you come from, you know you've done a lot in this industry. So just give us, give us, you know, the high level resume here. Like where have you been?
[00:01:15] Speaker B: Absolutely. So I'll start with the relevant resume and end with the completely irrelevant part of it. So my journey to Cisco tracks it with my startup past which was I was in a fintech startup in the founding team.
I started my own startup called Cloud Cherry in the customer experience management space.
And the funny part is when we started that startup, I remember our pitch deck to investors because none of us were from the CX space. And our pitch deck said the leadership team has 200 years of experience as customers.
So we said we're building software for ourselves because our needs are unmet, our loyalty has not been owned by anybody and we're trying to build the kind of experience we deserve as customers. So that startup was called Cloud Cherry. Cisco invested in us across two rounds, acquired us in 2019 and that's how I initially came into the contact center business. I did take a two year break in between to go to an AI startup, enterprise AI company called Uniform. Incredible experience, hands on as an AI practitioner. But then when the opportunity came to bring the worlds of customer experience and AI together, there was no better way place to do it in. And so I came back to Cisco. But the irrelevant part of my background is I spent nine years in the Merchant Navy out at sea, which was when I was way younger of course.
[00:02:36] Speaker A: Yeah, that's quite a job as well, a lot of work there.
All right, so I want to get into some of the things that we hear enterprise IT and CX leaders and you know, AI leaders talking about these days.
And I'll start kind of high level and we'll drill down, but just from a Starting point.
[00:03:00] Speaker B: How.
[00:03:00] Speaker A: How do you think companies should really think about AI adoption these days? I mean, we hear about generative AI, we hear about agentics, you know, we hear predictive analytics. There's all these, everything's AI these days.
So if you had to kind of take a step back from everything right now, how should companies be thinking about AI adoption? How aggressive should they be, how strategic should they be? You know, what do you, what do you see?
[00:03:27] Speaker B: So I've always felt that before we do anything with technology, it starts with the why, why you're trying to do something. Are you going to the cloud because the cloud is available? Are you going to mobile because everyone's talking mobile? Are you doing AI because. And we've had these technology transitions, right?
And so I first take a step back and try to think about why are we doing this and what are we trying to do? And my story is very, very simple. I always say that brand interactions between brand and customer should be like engaging with your favorite human. And if you think back at who your favorite human is, it could be your spouse, it could be your best friend, it could be what have you.
And your favorite human essentially has a few attributes to that engagement which are truly special. So you know, they're always available to you on a channel of choice. They know, are you a texter, are you a phone person, are you a morning person, are you a night person? They understand, do you message rapid fire? Do you send one message in an hour?
They always know your history and context, so they never pick up the phone. Irrelevant to the context of what's happening in the rest of your life, which is very important.
Now, they don't always tell you what you want to hear, but they tell it to you. They tell it to you straight, with clarity on a channel of choice again. And if you look at brand and customer engagement, we tend to move away from that simple ask that customers have. So my ask of anyone looking at AI adoption is first not to bring AI into the conversation.
Think about the kind of experience that amplifies or exemplifies your brand and then put the solutions cape out and say, this is the kind of customer journey I want to give my customers.
And then previously there are certain things that were not possible. Which is, for example, let's say someone says, I want to be always available to my customers, always.
But it turns out your contact center can work X hours a day. You can make it 24 hours. Of course, the cost doubles or triples or what have you. And now that's where Your aspiration meets reality.
Now, AI can solve for this. Always availability.
The next is, I don't want it ever have anyone wait. Like your favorite human never makes you wait, they always make time for you. So if you say, I never want to make anyone wait, which means what? I should have no call queues, no call wait. That's not possible.
Because, you know, what are you going to do? Hire 10x the number of agents or what have you, AI can come in. So I think if we define the kind of experience we want and have an understanding of why we want to do that, then we should look at how AI, automation, and a ton of other technologies can help us achieve that. So what to do with AI we can come to in the next step, perhaps. But defining that North Star experience is probably the first step, and AI should probably not be part of that definition first. And then we should be looking at how AI can help us achieve that goal. So we'll come to that. But I just want to pause with that statement because I honestly think it is incredibly important before you go down the rabbit hole of I have 10 pounds of AI, let me see, where do I put this?
[00:06:23] Speaker A: Well, and there's so many ITCS leaders who will say, well, our CEO said we've got to have AI, so we got to do something, let's do something. You know, it gave us some budget and they rushed to it. And I found that in our research, so many companies try and boil the ocean. They're trying to do too much at once.
They're trying to just roll out AI for AI sake, rather than identifying what their business problem or opportunity is that they're trying to address. Like, you know, what are you trying to do?
Sort of the why before the what.
[00:06:53] Speaker B: So I think that brings me to, I think the other question you asked within your question, which I think is very important, which is how do you go about it? Now, I think we all agree and understand that AI is here to stay and we've got to use it. So that's undeniable. And hence that's not the focal point, sort of everything that we're doing. But now that you've defined the North Star customer experience, the things you want to do, there is a precondition to AI, which I think we cannot get away from, which is, as an enterprise which deals with customer information, customer data, customer experience, we need to first completely take care of data privacy, security, sovereignty, all of that. That's table stakes. That should never be part of the active call. It's not a feature you give someone who pays 10 bucks more, right? So the incredible thing with being in a company where security is not a feature, but it's woven into the fabric because we also have massive, massive security business, right? It's not something we think of as a bolt on is important because armed with the comfort of knowing that security is non negotiable, you can then truly experiment and go and build AI use cases. So let's assume that's the pre step.
But you asked a question which I think was very important, which is how do you go about it? If you have your security sort of taken care of, you have your North Star defined, then the best way I think of using AI is to look at use cases where AI at the very least keeps your customer experience the same or makes it better.
And then do that at an empirical level and see what the impact is on efficiency, customer experience, outcomes, all of that before you scale it. So simple example, let's say you have a and 8 to 5 contacts and at 5pm Eastern you're basically telling your customers, you know, you can't call me. Or you have these seasonal ups and downs, lots of people calling about the same thing. Or we have a very simple feature called topic analysis which uses AI to go through all your call flows and tells you why are people calling. And then you identify that look, 20% of my callers are calling about the exact same thing, which is a fairly objective, it's a number or a yes, no answer or it's a check from a system.
Why don't we look to automating that? Because when someone calls up to find out is my flight still running on time? I'm just making this up, okay? Or when is my refund due? And you have that information in the system, it's their speed of delivery is more important than the entire full fledged human experience. I wait for five minutes, then you come, you ask me about my day and then you don't need the full fledged human experience, you want a very, very optimized, efficient experience. So, so when we look and put the experimentation hat on and look at these as experiments at an empirical level and then measure the impact on customer experience from an efficiency perspective, from an impact perspective, from a CSAT perspective, then you sort to sort of start to prove out the value of AI, the ROI of AI. And when you've done enough of that, you'll realize that there's an old axiom which says go slow to go fast.
So when you've gone through and you've done the right things, then you'll realize you can rapidly scale up the adoption of AI.
And that's the approach we tell all our customers to take.
[00:10:08] Speaker A: And you kind of mentioned some of the analytics that you can look at within calls. I think that's one of the hottest areas right now of anything we're looking at in AI and cx. Just being able to just pull out those hidden gems within conversations and interactions with customers and be able to use that information in so many ways. And I see right now, now a lot of companies using them within the contact center and to improve agent performance and things like that. But I see that as something so much bigger, you know, like, why is this data not shared with the CEO of the company who's making decisions on a daily basis that impact so many people? What do you, what do you think? Where do you see the future of this whole area? I mean, everyone calls it something different. Interaction analytics, conversation intelligence. You know, just. Just being able to use AI to pull out all sorts of data and statistics from conversations and interactions.
[00:10:58] Speaker B: So acknowledging that a majority of stats are made up on the spot, I'll tell you some stats. So I think it's fair to assume that if you look at what we know in a CRM about a customer, we know your name, we know your address, we know four other things, we know some purchase history, but all of that is the past.
And there is so much we learn in active conversation. When a customer is messaging you, calling you, for example, yeah, I was looking at that iPhone, I was thinking about this, I was thinking about that. Or there are certain cues that aggravate a customer. There are certain things they like. So I'll give you a simple example. Let's say you offered a discount to somebody. There may be customers who love the discount, they'll take it and then leave you. And there'll be customers who love the discount and say, this is a great deal. I'm going to come back again. Knowing the difference between the two is incredibly important.
What does certain treatment do to a customer? So the point is, so much of customer insight is hidden in unstructured verbatims, voice, text, what have you.
It could also be subtle things like when you said this, the customer actually got angry, did not like it. For example, certain kinds of customers or certain kinds of customers, when coupled with certain types of call types, halfway through start screaming, just, agent, representative, just get me out of here. Right?
So to know that is so incredibly important and I think the greatest treasure trove of information that we are only scratching the surface of is conversational data and being able to track this interaction data across the journey of the customer, extract the metadata that matters, use that to power your actions. It could be next best action, it could be the right offer, it could be a marketing outreach. Subsequently, I think is going to become one of the most important areas of investment for companies because you'll realize what you knew about the customer is the past.
The only thing that is present, continuous and real is what the customer's live action telling you and being able to. And this information has a shelf life.
It is not in perpetuity. So for example, when I say, yeah, something happens and you know, someone says, my flight's delayed, yeah, but I was going for this graduation, that graduation is not going to happen for the next three years. Right. I think day after tomorrow.
So anything you want to do to make that an incredible experience that's relevant to that graduation is relevant for the next 48 hours. Right. So how do you learn that on the fly?
How do you act upon it?
How do you allow that, to use that to deliver superlative customer experience very profitably?
That is where I think the puck is at or where it's headed.
And the greatest source of that information is conversational data.
[00:13:34] Speaker A: Yeah, absolutely, for sure. And the other thing that I have seen quite a bit of is just the whole idea of proactive outreach to customers. I think that that is another big area that interaction analytics is going to help quite a bit in that area for sure.
But what our research is showing is that most consumers, this is our consumer research, Most consumers actually say they want you to reach out to them proactively.
A lot of businesses say, oh, I don't want to be annoying. Well, no, that's what they want.
They just want you to know, here's how I want to be contacted, here's how often. By the way, the average number is no more than 3.67 texts per week, for example, is where people say it's reasonable, but just catering to what they want. Maybe I want to be notified of a sale, but not new products. Or maybe I want to know about new products but not any events that you're having. You know, things like that. And, and AI plays such a big role in everything with proactive. So I, I wanted to get your perspective on that. You know, like when you think of proactive outreach. And so certainly using proactive for customers, but even internally you can use proactive in how you use your data. You know, from those conversations you can, you can be proactive in how supervisors are Interacting with agents. There's so many ways you can use sort of that whole proactive mindset. I think the biggest one though, is with customers and just knowing what to do there and just having the right communications platforms. Do we do this over CPAs? Do we do it over CCAs? You know, how are we, Are we sending emails out of CRM? Like, what are we doing?
So talk a little bit about your perspectives on the importance of proactive from a business standpoint and what companies should be doing.
[00:15:17] Speaker B: So, as you know, we have a ucaas, CCAAS and a CPAAS platform. And the reason we feel all three are part of the same continuum is we've gotta understand that, you know, marketing or some team like that or sales owns the outbound. Okay?
And then of course, there's a contact center, then there's another team, there's another team that sends you appointment reminders for the customer. It is part of the same continuum, which is an engagement between me and my brand.
So the more we siloise this, the harder we make it for our customers.
So if you recognize that they are part of the same continuum, you'll treat them as the same and be very cognizant of how many times you reach out. What do you communicate? Do you allow them to start to converse with you when you send them something?
So for me, if you look at Outbound, you said something very interesting, which I think is very important. So when I look at Pro Outbound or someone reaching outbound to me, again, putting my customer hat on, they need to understand my channel of choice. I hate being called, for example, many people love being called. I love messaging. But to your point, 3.7, there's a number, correct? Don't message more than that.
So once you know, channel of choice, time of day, all of that propensity, the point really comes down to what am I messaging you about? If I send you 3.7 texts of something completely irrelevant, okay, then it's pointless, right? You got the number right, you got the message wrong. So that message comes from, as I said, conversational data. So I think the greatest thing about Outbound is being obviously cognizant of channel, understanding the context of all prior engagement between brand and customer. Know that whether you call them and message them. So you can't make six calls to them and 3.7 texts and say, oh, I only sent you 3.7 texts. You call them 9.7 times. That's what's essentially happened. So be able to thread seamlessly engagement between brand and Customer across all channels, recognizing that do not disturb on one channel means do not disturb on the other. Being able to manage.
[00:17:15] Speaker A: Yeah, that's a good point.
[00:17:16] Speaker B: That is incredibly important. You, I block you on text, and then you start to WhatsApp me, you know, so that's not fair. Now, whether it's legitimate or not is a different story. From an experience perspective, that's wrong. Right. I should not have to have my guard up against the brand on every channel.
So the ability to manage your interaction with customer across all channels, voice being a channel, being able to orchestrate in one place, cohesively manage compliance, cohesively dedupe any interaction between brand and customer, and then ultimately use conversational data to feed the right message. My gripe always has been, I think the last superlative innovation in personalization was mail merge, where, you know, we got the same mail, but it said dear Robin and it said dear Vinod. And that's not personalization. That personalization is knowing that you've got this seminal event happening, knowing that you're getting a new job. You may have learned it through some other way, or someone would have said, yeah, look, I'm moving to Denver for this new job. Such a great place to say, hey, dress yourself up for your new gig. Right. It's such a relevant message because your mind's thinking all of these things, right?
Is my office bag relevant for my new job? You're thinking all these things.
And so any message relevant to that life event or that what you said is important. So along with all the things I said around managing your omnichannel engagement with the customer from one place, which is outbound and inbound, and then messaging out the right message at the right time on the right channel to the customer, you'll realize you said exactly right. People want to know that there's a deal out there for them or the right product for them, but they get very.
How do I say this? Like, you, you lose all the signal in the noise because I've just gotten 16 messages from you last week. The 17th may be relevant, but I've stopped reading your messages. If I'm not dnd you, I've stopped reading your messages. And then you lose that channel and that customer on that channel forever.
[00:19:08] Speaker A: Yeah. That's like whatever we look at from a consumer, consumer standpoint, what they care most about is time. It's like, respect my time. When you save me time, I'll buy more from you. That's like the top thing that makes them buy more. When you save me time I will use an AI agen versus a human agent, or vice versa. So time is so important to your point. Like, if you're sending somebody 16 messages, that's not respecting their time because it takes time to read every one of those.
[00:19:33] Speaker B: And you said something important, which I want to amplify again. I said, we have the CCPAAS and the CCAAS platform together. The reason it is together is because we truly, truly believe every interaction between brand and customer in the customer's mind is part of the same continuum. So let's say, for example, I send you what was in the realm of alerts, which is, hey, Rogan, your payment is due on the 19th. Okay? Historically, you get that on the 19th. Now you're like, oh, my God, let me check if I have money in the bank. What's the story? What about this?
And now you call them and say, look, I've got this. Imagine if you get that message and you have two questions, which is, hold on, can I break this up into parts? Can I? You know, if I pay two days late, I'm out. I can't. You know, whatever. You have a question.
So what was a notification message sent by a random team becomes a conversation.
Okay. And I come in and say, you know what? No problem. You can do it. I actually have an offer for you. I have a zero interest. You can break this up on six months or have a low apr, whatever, Right? That becomes a marketing offer, and then that person does the transaction. It's actually a sales event. The point I'm making is we are breaking this into messaging and sales and marketing and support for the customer. It's one thing. You send me a payment reminder, I needed something, you presented me options, I transacted upon it. Correct. And suddenly I'm paying you 2% interest more. So the point is that if you recognize that's what we do, obviously, with our Connect platform, we see all of this outbound, inbound AI in the channel, all of them as one continuum. And the more seamlessly we look at these technologies and not get too caught up in platform A, platform B platform, you'll realize you'll create the kind of experience the customer wants. And funny enough, when you create the right customer experience, it's almost always more profitable. So if you put experience first and the business value second, more times than not, you'll get it right?
[00:21:26] Speaker A: Yeah, for sure.
So I want to shift gears a little bit and talk about the underlying data that supports everything we've been talking about.
You know, the more I talk to companies, the more I hear them saying, well, we started doing AI and we, we just really realized our data wasn't ready. And you know, when we look at our research right now, less than half of companies say that their data is ready for AI. Yet 90 some percent of companies are doing something right now with AI. So there's a disconnect there. So what's your recommendation for companies when it comes to, you know, comes to data? You talked a little bit about security, which is another important, you know, ancillary yet important component of AI. But what about data data readiness?
[00:22:05] Speaker B: So I think you know this industry as well as anyone. So I mean, without having a data strategy in place, it's almost moat thinking about an AI strategy. Because you know, I always say in AI nobody's got a moat anymore. Anyone can make an agent or an assistant or what have you. The real moat is a few things.
One is access to first party data. Next is a deep, deep, deep understanding of the workflows that are very kind of domain specific. And third is execution, which is how well you productize what you build. And data is obviously central to all of that. So a few interesting things I think. First, it's incredibly important first to have an active way to consume real time data, which is what we call the journey data service, to be able to learn on the fly. There's obviously a huge data prep part which is if you need your data to become knowledge and to be able to run a rag on it, it's very, very important for your data to be vectorized, for it to be structured, to be, be, it needs to be cleaned up, right? So that's an entire stream of activity that needs to be done. But you need to also have a way for this active data. So there's passive data stuff you always knew it's data at rest. Then there's active data stuff you're learning that is enriching your data that you're learning on a daily basis. How do the active and the passive data kind of meet? How does that continue to enrich your understanding of your customers, what works, what doesn't, and improve the treatment? How does that help your model get better?
That's an extremely important investment that everybody essentially needs to make. But in doing that, I think there's a couple of things that we need to be very, very mindful of. The first, I think is when you look at the vast amounts of data that you have, you'll realize that every model, every application, you're never going to be in a place where one model runs through all your data.
So to give the right model, the right agent access to the right data makes it cheaper.
Because imagine if you have like this 100 petabyte data store and you just tell the model to crawl it. Every time someone asks a question, I spend $100 serving a $1 query. So you've got to be very smart about what models have access to what data and you got to make these very efficient. That is where there are two things that are happening which I think are very relevant to our space. The first is frameworks that allow multiple agents to coexist, coordinate, communicate and collaborate with each other.
We obviously part of agency where we are doing exactly that. What is a multi agent orchestration framework? Because that leads me to the second thing that I think is the most important thing that's happening with data.
Historically, what's happened is the biggest problem with enterprise software has been enterprise data integration, bringing everything in one place. How to have one sort of ETL layer that normalizes data.
But what's changing now with frameworks like A2A and MCP is each system owner is able to build an agentic application. And this multi agent orchestration frameworks allow these agents to talk to each other. Which means if I'm going to, let's say Robin Inc's software, I don't need to access your database or your stores. I need to speak to your agent. Your agent has a sort of a JSON saying, hey, this is how you talk to me. This is my readme. This is what I can do. This is what I respond to. This is what I can do. So you own that agent, you train it, you guardrail it, you do all of that and the agent becomes an interface to the world. And my agent essentially talks to your agent. That I think is a very transformational change.
And the number one trigger to true multi system agent tech applications is going to happen through that wherein you'll see that I don't need to access all your data so long as that system cleans up its data and gives its agent the ability to interface with a third party.
And we build frameworks which allow for multi agent orchestration. The customer still gets one interface, you still have one point of interface. But this sort of concierge agent, or this master agent is essentially coordinating the action of these multiple agents and giving you one seamless interface. What they don't realize is you're talking to hundreds of systems in the back. So these are the transformational things I think that are happening in data base, data cleanup, Readiness for AI I think is a non negotiable. I think every CIA on the planet is well underway. But this multi agent orchestration frameworks and the ability for these to work in mesh I think is the most dramatic sort of enhancement in, in this area.
[00:26:37] Speaker A: I agree. And it's moving quickly too.
I'd like you to expand a little bit on that though and talk about, you know, how are companies today using AI to drive these like real business outcomes, things like revenue growth and improve retention. How do you see it happening today but then moving forward, where do you see this agentic, this like hierarchical agentic sort of framework changing things and you know, kind of improving even more than what we can do today?
[00:27:07] Speaker B: I think so the very, very obvious areas that everyone knows is obviously what we call call containment. I hate that term because no customer wants to be contained or deflected. But I think proactive customer engagement across multiple channels and the ROI is very, very, very clear. Right. Which is, but I want to mention I would never ever see call containment rate in isolation.
Even if you use the term, it has to be seen in tandem with csat. If your CSAT is at the very least neutral. Right. Before and after. And now you're taking a $6 cost per serve to one. Fantastic. Right? Customer lost nothing. In some cases the experience got better or they were available to you at a time that historically you weren't and then you've dramatically lowered. So that's a very simple slam dunk use case. But what you'll see is what customers are doing is they're looking at using prescription refill.
It's a simple workflow or a stream checking account balance. Right. So a combination of what digitization is supposed to do with hey, look at your mobile app. It has all this information to conversational automation will allow us to take the most common use cases and you'll realize extreme end to end automation can be done including fulfillment. Fulfillment meaning as I said, prescription refill, ordering something, changing something, getting a status, getting a claim process, what have you. Right. So I think those are the simplest in each industry. You'll see if you draw a pyramid of use cases. There's a baseline number of use cases that everyone is working to automate any. And that's the interesting part. AI is a tool to automate, whereas the worlds of AI and automation were different primarily because of the fulfillment aspect of the whole thing. But you asked me a part two question around where is this headed with agent tech? And I think the. So let's put straight line use Cases which is you're able to execute upon a transaction which is can I order this? Can I change this? Can I change my line? Can I add my, my wife to my phone line? So you can do some basic use cases end to end, fully automated. But where AgentIQ truly is taking us is where the agent is able to own the workflow because oftentimes something changes. What about this, can I add this one thing? You need to check to a certain check with a certain system you need to orchestrate action in a system that action yields an outcome and with armed with this new knowledge you need to on a dynamic basis change the flow which is I want this, it can be done. Turns out I know you wanted 10,000, your credit limit is only 7.
Here's two options for you, right, I can do this. So the ability to do what a human would do, that is you interface with the system, get some information, it's not straight line, it now needs you to kind of decode that and then take the customer down a certain flow. That is where agentic is going and that's where what I said earlier really becomes important.
As multiple systems using multi agent protocols are able to talk to each other, you will be able to automate these actions.
That I think is the, how do I say it? I think that's the magic with AI wherein you'll see not just an extreme amount of automation but the kind of trade offs, the kind of. If you can't do this, let me suggest this to you. The very humanness of the interaction you'll realize AI will be able to do because it has access to information, it has the ability to rationalize what's going on and within obviously guardrail set by you, for example, if you're not allowed to give a refund, it doesn't give a refund. That's easy, right? You code the guardrail in, you prompt it that way but it's able to make those trade offs and take the customer through the entire journey navigation even when the flows aren't predefined. And I think that is where we are headed. That's present continuous as we speak. But in the next year or two or three we'll see an extreme amount of that happening including the involvement of multiple systems talking to each other. And that's what I'm really excited about.
[00:31:06] Speaker A: All right, so we've talked a lot about a lot of stuff today so let's close out by.
So when you, you've talked about what we can be doing moving forward, what's going to happen in the future. What can we be looking at? What do I need to do now though? So like what's next? So what do I need to be thinking about today to prepare me for the next two years, let's say, or two years from now?
[00:31:32] Speaker B: So the question itself has the answer, and I love the question you're asking because we've got to understand that AI adoption is a journey.
So the first thing, obviously we've spoken about everything else around what we need to do. If we recognize the fact that it's a journey, then if you're going on a journey, you choose the right partner. And the partners you need to choose, surround yourself with are the ones who also recognize it's a journey, who align with you on North Star, but align with you on what's the first step, what are the intermediate milestones? And how do we basically build a game plan that has checkpoints where we assess what's going on? So what you should be doing now, obviously, as I said, I will start with get your data in order to get your security posture like do it now. Don't discover a day before launch of applications that oh my God, this is not allowed. So do the hygiene first does happen.
Exactly. In fact, I think there was a, I read a certain report which said majority of agent tech AI projects will fail by 2027 or whatever. So I don't know, I don't have an opinion on that. But I think the point is if you just plunge headlong without thinking through the hygiene stuff, you'll more likely than not fail. But once that is done, and given that we all believe that AI will have a transformational role, solve things at an empirical, pick the most common use cases that are absolute slam dunk and don't get besotted with AI just in the front line, the amount of automation and efficiency AI can drive inside. I'll give you a simple example.
When someone calls a help desk, the AI actually goes through even code snippets and is able to bring out and say okay, this is what's happening, this is the issue.
So you invisible below the waterline. So much AI can be used, AI can be used so much for optimization. Pick those kind of experiments and become conversant with the technologies that you're going to use to scale. Because when the training wheels come off, there's enough people who build smaller projects that seen success, understand how the systems work and they're then ready to scale. So long as you have a partner who believes in that, you have the same a platform that allows you to experiment and you have that mindset and methodology and then you fix your data, get your security in order, then the sky is the limit. I mean, we're going to see some incredible gains from AI. And if you get our hygiene right, you'll realize that this will be truly probably the most transformational technology in the history of enterprise tech.
[00:34:01] Speaker A: Sure, for sure. I think companies need to also, you know, stand up as an AI center of excellence of some sort. You know, like where they have a cross disciplinary group of people who are meeting regularly, setting the rules of engagement and just, and just, you know, you're not just having technology people look at this, you're also having people in sales and marketing and hr, you know, throughout the company and really trying to bring in all those groups and you know, perhaps even have one strategic leader like a chief AI officer who kind of runs a strategy at the company. Not to say that they're going to do everything, but we are seeing that starting to emerge quite a bit in our research as well.
[00:34:37] Speaker B: I think that is great insight, Robin. AI truly is a team sport. And I loved what you said, cross functional teams.
The great thing about AI is in the way it's coming to enterprises is you don't need to be an AI expert to harness the poverty.
What is the great thing Cloud did? Cloud said, I'm going to give you all these services on demand. Go build the best application you could. Don't become an infrastructure expert because you need to build an application. Right. What AI is now saying is you don't need to know to code. You should know your domain really well. So what the great power it's giving us is know your customer, know your domain, know your North Star, understand what you want to achieve, okay? And find the best way. We're going to give you all the tools to achieve that.
And that is the true sort of democratization of access to tech, wherein everybody in the organization who knows HR knows the people's stuff best. You know, engineers know engineering, give them the tools. That's what I would do.
[00:35:36] Speaker A: All right, well, Vinod, thank you. This time went by fast. I appreciate all of your wisdom and insights and I'm sure the audience does too. So until the next time, Vinod, appreciate again your time. And that concludes today's episode.
[00:35:51] Speaker B: Robin, thank you so much for your time. It's always a pleasure speaking with you.
[00:35:54] Speaker A: Likewise. Thank you.
