BizTech BizTech Podcasts

81. Is AI better than a human being in QA for the contact center agents? With Ashish Nagar

August 9, 2023

Subscribe to the Next Level BizTech podcast, so you don’t miss an episode!
Amazon Music | Apple Podcasts | Listen on Spotify | Watch on YouTube

Listen in today as we discuss the hot topic of AI, specifically AI in the Contact Center, and if it’s better at Quality Assurance than a human. Ashish Nagar, CEO of Level AI, joins us today as we dive deep into their real-time QA offering and how it sits on top of other contact center offerings. We talk about Generative AI and how it solves some of the most common problems in the contact center and turns them into revenue-generating machines!

Josh Lupresto: [00:00:00] Welcome to the podcast that is designed to fuel your success in selling technology solutions. I’m your host, Josh Lupresto, SVP of Sales Engineering at Telarus, and this is Next level Biz Tech.

Hey everybody, welcome back. We got a fun episode for you today because it’s never boring to come in and talk about AI. What better time than now, right? Listen, our topic today is we ask the question here, is AI better than human being for qa? Contact center agents. And if you don’t know what QA is, I bet you can spell it.

I promise it’s not as scary as it might sound. We’re gonna go over that today. But today from a guest perspective, we’ve got the CEO of Level AI on. We’ve got Ashish Nagar Ashish, welcome to the call. I man.

Ashish Nagar: Thank you so much, Josh, for having me.

Josh Lupresto: I’m, I’m excited. You, I feel like this could be a, you know, we’re, we’re, we’re trying to make this a 30 minute podcast.

I feel like you and I could talk about this for four hours, but they, they make me keep it down. So I’ll, I’ll do my [00:01:00] best to keep us reigned in, but excited to have you on, man. Excited to kick this off. Same here. Alright, so I want you to take us back a little bit here. We want to hear a little bit about your personal background, your unique story, and how you got in this space.

Ashish Nagar: Yeah, so it’s actually very interesting Josh, and thanks again for having me. So I have been working in the space of voice AI and N L P for almost the last 10 years now. So I, I was doing a startup which is kind of c you know, Siri early version of Siri, which we sold to Amazon Amazon’s Alexa team.

And was working as a product leader in Amazon’s Alexa team for a few years when Alexa was growing up, was, was growing a lot. And the project which I was working on towards my last assistant at Alexa was making Alexa talk to any human for 20 minutes on any social topic. So imagine if you walk into a room and you find the one of those devices there and just like, Hey, how was the.

How’s the NCAA [00:02:00] tournament? Like how’s March Madness? Mm-hmm. And anything which you can talk to a human being on and, and how can we make Alexa know about everything, have a free forming conversation and be engaging and all that stuff. And on that project I was working with research groups from around the world to make this happen, and then I realized all the challenges which are there in the space to make question answering work, better machines, to understand humans and so on.

And then realized that Google, Amazon, Facebook, open AI are all working. For these mass solutions. Right? But nobody’s working in the enterprise. Mm-hmm. Who make a world class solutions. And we saw that the technologies which were there before us were all very basic. And, and so that’s how I got into. The enterprise AI space and what we first started doing, what our first product was actually a bot for frontline workers, like nurses, technicians retail sole workers.

And when we showed it to all the customers, they were like, can you first [00:03:00] apply this in the contact center? That’s where most of our problems are. That’s where we are getting all of these voice calls and other thing. And our question was, but don’t you have. Incumbent solutions there, you know? Yeah. I don’t wanna take names, but you know, and all these existing solutions there.

And they were like, yeah, but they haven’t changed in 10 years and they don’t. There are many, many problems. And as we dug deeper, we realized an opportunity to build a modern AI stack driven by large language models as some of these underlying technologies, which are getting a lot of attention now with Chad G P Ts based on that’s what we did.

That’s our journey into customer services. I was an AI guy who, you know, got into customer service over the last four years here.

Josh Lupresto: Love it. And, and so, you know, I think there’s a lot of partners that, some of our partners that listen to this might. Be already selling cx, some of them maybe are selling cloud and, and thinking about cx.

And so I wanna take us on this journey because some of them have grown up with, with some of these incumbents, right, that have been around space for a [00:04:00] while. Some have evolved, some are still evolving. But give us then, all right, you took that knowledge, you’re c e o of Level AI. For anybody that has not heard of Level AI yet, give us kind of the pitch on what your market, what your plan is.


Ashish Nagar: So Level A I is a customer service intelligence platform and we analyze, automate, and assist in a agent and service representative conversations. So what does that manifest in terms of products? So, Our products in real time and post conversation help improve agent quality, agent performance. So we have an agent assist product an automated QA and a manual QA product, and we have a deep customer intelligence solution, which mines this customer intelligence for.

Agent and team performance trends, but also product service and business performance trends. So that’s why we call it customer intelligence, customer service intelligence, offering both real time and post call. [00:05:00] It’s only channel. So if you’re doing SS M Ss, email chat. Phone calls, slack team messaging. We absorb all of that into, into our mix.

We are also look at customer feedback, you know, on N P S reports and so on, thinking of social media as well. So it’s one customer service intelligence platform. For, for your listeners who are general IT and, and and CX folks? One thing I would say the way we think about the customer service world, there are four pillars.

In the customer service software stack in our minds, one is the underlying infrastructure, which is the telephony, email chat enablers, the five nines, the genesis, the Twilio’s of the word. The second is a ticketing solution, a customer, a Salesforce s. A third is an I V A or a bot solution to automate whatever you can automate on the front end.

And the fourth is a customer intelligence solution. Our view, like a Level AI, which sits on top of all of it, we sit on top [00:06:00] of C R m telephony bot. We often get requests for bot qa, for example. So I, that’s our all.

Josh Lupresto: Yeah. I love it. I love that. I love that your realtime QA product is now gonna QA bots in addition to humans.

I love it. I love technology. It’s, it’s just cool to hear out of all of that. I. It’s, there’s a real time option. We’re finally at a spot with technology where that’s there and, and just from a bi, you know, my, my engineering hat goes one way and my business hat goes another. I think opportunistically, there’s so many things that people can do to make their operations more efficient.

Upsells, you know, better responses. I mean, the, the, obviously the list is endless. So Great. Yeah. Great stuff there. And, and, and maybe that answers this, what was gonna be my next question, but I’ll let you state it your answer to it. What do you see right now if you look out, maybe you answered it with, you know, you, you thought a lot of the incumbent solutions had this, but what do you think right now is one of the biggest problems in the CX space?

Ashish Nagar: If you, it depends on sort of who we [00:07:00] talk to, right? Like, so if you talk to the C E O or the C E O or the VP of the contact center and so on. So, you know, they worry about CSAT revenue uplift Customer loyalty and, and for those situations it’s, you know, we think of the biggest brands in the world.

We think of Apple, we think of Tesla, we think of Disney. What comes to mind is incredible customer service. You know, so from a leadership perspective, from a boardroom perspective, that goal doesn’t change, right? That’s what everyone’s end goal is. But your point about what is the biggest problem with cx?

I think there are a couple of things. CX, and this is almost a cliche answer, but I. I think it’s so true. CX is often looked as a cost center, but it’s not. It’s a revenue center because again, incredible customer service leads to great brand loyalty, customer retention, which leads to people buying more and converting more and spreading the word and so on, right?

So how can we help CX leaders continue to [00:08:00] make it a revenue center, number one. Number two. How can we help CX leaders to take care of their teams better? And, and, and one of the biggest problems in CX is churned and agent retention and agent training . That’s a complicated problem. It has human resource links, it has training links, it has tooling links, technology enablement links.

But those two things I would, I would say are in my mind, how do we enable CX leaders? To take care of their teams better and support their agents better. And give them data and technology to make it a revenue center. Right? And, and by better CSAT better revenue expansion and so on.

Josh Lupresto: Yeah. Love it. That’s good stuff.

Yeah. You I’m gonna share this thing because it recently happened to me. The importance of not viewing. These agents as a cost center, as more of a brand uplift, a brand chair, a brand rebuild. My wife and I bought a treadmill just a little over two years ago. One of these treadmills that’s got it, [00:09:00] you know, it’s integrated with a 10 inch screen.

I can, you know, as I run it, it, it moves up. You know, I’m, I’m through the hills of Canada and, and all these cool places, right? So come down one morning and see this weird message on the screen of the treadmill. Try to reboot it, try to factory reset it, and if you’ve ever done a if you’ve ever, you know, hacked your old Android and you’ve seen that message, you know, or when you’ve seen an old phone, you know when it’s bricked and it’s just dead.

Yeah. And the funny situation with this thing was that even when the screen was dead and, and, and that. That piece of it was dead. You thought, well, fine. Okay, maybe I can still use the treadmill, right? Yeah, yeah, yeah. You can’t it was dead in the water. It was, you think of it like fail to close. So it was, was, that was even more frustrating.

Like, oh my gosh, I can’t even use this. So looked up the warranty. It was about a month or two months outside of the the two year warranty. And so I’m thinking, geez, we just dropped 1800 bucks on this. How does a software update brick? I don’t think I did anything. It was a very unique situation where I thought, I feel like I [00:10:00] have an angle to complain here.

’cause this is, it’s a little off the wall. Yeah. Tried tried calling, no support, tried email, no support. Yeah. Went into the, the, tried to dmm them on Twitter. Got an instant response. Worked with them over a 30, 45 day period. The product was backlogged because it was me and thousands of other people that this had happened to, and evolved itself into a class action.

But the way they handled it, The way that that brand handled it and got me all of this was done on Twitter dms. The last place I would’ve thought of, you know, getting support from a company like this, just that you, you can’t, everything that what you said, you can’t underscore the importance of that brand enough.

Whereas I was ready to share my bad experience with a lot of people. Yeah. But the way that they handled it, I’m not dissat Telarusfied. I’m bumed that it died, but I’m, I’m so happy in the way they handled it.

Ashish Nagar: Will you buy another product from them? Absolutely. That’s it, you know? That’s it. They’re gonna launch more.

They, and do you and your friends will know about it. Like, I [00:11:00] have one of those machines at home and my friends follow me. I’m like, dead last. But still they follow me about it, but it’s okay. But you know, that’s where it comes to customer service is revenue and, and one of the things is, you know, better analytics can help with that, right?

So if I’m the VP of contact center and I can say, What, how many Joshes are there, you know, with good trust on that data around, you know, software failing and then being in a queue for two days and, right. Yeah. And can I build a tiger team to address that and train agents on it? So if you trust the data, you can do something about it.

You can, you know, build it back into the program. So, There’s a lot we can do there. Yeah. I

Josh Lupresto: love it. Let’s, let’s get back to maybe partners like to understand, all right. Now that I know where it fits, I know it fits in qa and so we understand that a, a as an add-on to some existing contact centers. Walk to me, walk me through maybe some differentiators that you [00:12:00] have.

Right. You. I think you, you mentioned a key one of them in the beginning, but I would love for you to kind of restate if there’s other products that are out there, what do you feel that some of the differentiators are that you bring to the space?

Ashish Nagar: Yeah, so I, I’ll, I’ll talk about it very simply, George.

So AI in the contact center, I. Is stuck. And I don’t mean it in a cl I don’t mean it like hyperbolically, but it’s stuck in the two thousands, if not 2000 tens. You know, why do I say that? Most of the incumbent solutions, and when you also talk to customers, they say like, oh, do you do transcription? What is your transcription accuracy?

Or would you do sentiment and analytics? And we’re like, oh yeah, but that is 10 year old technology. We don’t even talk about, we do that. It’s base case. It’s like, you know, so, so what, what the, what Level AI does is we do the transcription, we do the sentiment. It’s all base case, but we can get to true intent.

True meaning of why somebody’s calling in, why there is a, why there is a refusal, or why there is an escalation. Right? [00:13:00] And I’ll give you a very simple example. Someone calls in and says, What’s the problem with existing systems? I want to cancel my account, right? So in existing speech analytics systems, conversational analytics and auto QA systems, you have to go in and program every way.

Someone would say, cancel my account, cancel I want to cancel, and so on. And then the speech analytics system recognizes that that is term, that’s not machine learning. The problem with that is because someone says, I do not want to cancel my account. Just want it paused, you know, still pick it up as, cancel my account, because those words were there, right?

Yeah. So Level AI is fundamentally different. We would tag it accurately. We would say first one was cancellation, the second one was account pausing. So we understand the whole meaning of it. That’s a fundamental difference. And wow. And how do we set it up in existing systems? It takes days and sometimes weeks to set, set that up by giving all the keywords and phrases and every variation of it.[00:14:00]

Our system, it’s a machine learning based system. It takes less than a minute to set up. You can observe AI performance, you can see the model performance. You can be, you just need a, you know, a middle school degree to set up, right?

Josh Lupresto: Yes, I passed, I made it.

Ashish Nagar: Right, right. Good. Right to set it up. And the results are sometimes a hundred times better.

On data which you can trust for analytics, conversational analytics for order, QA for agent assist. It’s a fundamental difference on setting up the system and the time it takes. That’s one part, so that you can actually trust the data and actually, you know, as a VP director, QA leader, make, that’s one part.

The second part is on, so we call a detection. You know, the other part is synthesis. We use generated a AI, so we had a, so we have multiple generative AI solutions, which is kind of what chat G P T is based on to do auto summarization of calls to do auto dispositioning of calls. So the machine just says, without any [00:15:00] prior feedback from your customer.

On what is supposed to be the behavior. So we auto disposition, we auto summarize, and the third part is act well, how we differentiate. So now you can act on it with higher quality data with automation. So if an escalation happens from Josh, and I’m the VP of contact center, at the end of the day, I want a report of all Josh’s out there.

Who reported this particular bug, you know, and were not able to get service or I did, that’s for the vp, for the agent. I wanna, if there’s a compliance issue happening, which the agent did right away, send an alert to the team lead and they are able to come in and make a connect, you know, chime in if needed.

So, detect, synthesize, act on all three areas. We’re sort of leveling up the contact center.

Josh Lupresto: I love it. Awesome explanation. So I would love to jump into a detailed example. So for this part let’s talk [00:16:00] about, you know it’s funny how sometimes that when we get into the deals and we actually talk to the customers, it’s so much different than.

What we were told might have been the problem. Right? Yeah. So I would love to hear an example from you that you got into, tell us about it. What was the expectation when you got into it, and ultimately, what did they have and why didn’t it work? And, and why did Level A solve AI solve it? You know, that’s a great point.

Ashish Nagar: Josh, so I’ll give you a couple of examples. So one is we this is for, this is for a major financial services company where we do auto qa, all their analytics and working on our agent solution agents assist solution as well. So the three specific pain points that didn’t work, one was this was nothing to do with AI.

This is like a, this is stitching together different data sources in the contact center. So they used, they have Salesforce, they have a cloud, Amazon, they have a CSAT solution, and the past pro, and then and they have a screen recording solution. The past [00:17:00] provider, which is a big which was a big, with one of the bigger players here, would not stitch all that data together into a single view of what’s happening with my agent.

The screen recording would not work on a certain channel. You know, we could not pull in certain fields. They could not pull in certain fields from SFD Salesforce. So Level AI, one of the key things which we do, in addition to the AI, one single source of truth for your entire contact center. We can pull and we pull in data from all these sources.

Easy when for us, but a big win, small step for us, big step for them. It’s like we have never seen altogether in one dashboard. I come in the morning, I see everything, number one, number two. On auto QA and analytics, they were using a, a, a legacy solution when they set up their speech analytics. I kid you not Josh, they sent us a C S V file where each tag, which they were setting in, let’s say account cancellation, customer greeting properly or not.

At hundreds of hardcoded scenarios of [00:18:00] like, what should you be listening for? What should the machine be listening for? For us, it was like, are you kidding me? This is so, and even then we set up through our AI models, within a few weeks we were getting 20 to 50 x better coverage with a machine learning model.

It took a few days to set up. Where folks are maintaining these libraries of keywords for five years

Josh Lupresto: and they, and they never have to maintain them at all. They don’t have to edit them at all going forward. Library the

Ashish Nagar: power they never have. No. That’s the machine learning, right? Like machine learning is we learn from past experiences, update it, it synthesize it, and off you go.

Yeah. That’s what’s number two, number three, voice of the customer. Now they do auto qa. They pull in all this data, but now they’re running their V O C programming Level ai. What does that mean? Which customers, to your point, like which customers are about to, could be churning? Which customers are happy with the current offering.

So they use the same conversation intelligence and tagging capability to come up with these views. And without auto [00:19:00] summarization, they’re able to see, okay, these customers are about to churn. Here is the summary of why they’re supposed to about to chart. So, and the product managers are using this data now to be able to say, okay, here’s what I need to fix the sale.

The, the churn data goes to the sales team, the service gap product gaps goes to the product manager teams. So it’s a, that’s why it’s a customer intelligence system, right? From pulling all the data, setting up better AI for trusting your data and then using it to make real actionable decisions across the enterprise.

So in our one-on-ones with them now, we just don’t talk to the. Customer service team, we talk to the product managers as well. Mm-hmm. We talk to their compliance teams as well. We talk to their sales teams as well, which is awesome for us. Yeah,

Josh Lupresto: I love that. Those are awesome, awesome examples. That’s good.

I, I think so many different scenarios there to imagine what this would solve for people. Okay. So when things like chat, g p t come out mm-hmm. And, and, and OpenAI and then obviously Bard and all these [00:20:00] others that are just gonna follow now. It seemed to me that I thought we were a little farther off from those things being that effective, that usable, that tuneable, right.

I didn’t expect it yet, so I feel like we just woke up. And all of a sudden, the world is now in a very large potential paradigm shift, and kudos to you for being one of the first people to come to market with a product that works. And it’s, it’s productized. It seemed like people were gonna struggle for a while to say, okay, I have access to the AI, the supercomputers, the models, all the GPUs, everything but productizing it.

I, I just expected it take to take a little longer. And so curious from your perspective here. Do you think we’re in the massive paradigm shift right now? I mean, what, what’s just, this is totally in your, your own opinion. Yeah, yeah, yeah.

Ashish Nagar: So I’ll, I I, I’ll give you three or four reasons. So the answer is yes, I.

We are, but [00:21:00] the two part, yes, we are, but two, it’ll take a while for, for, for the enterprise to adopt it, you know, for me. So I, and if you talk to our friends at five nine or Genesys or Twilio or Amazon, they’ll tell you like, cloud has been around as an obvious choice for the last decade or so, you know?

Mm-hmm. But then you guys are the experts. You do do this massive transformations, right? But last I checked Cloud penetration, the contact center is around 20%. You know? So, and it’s an obvious choice, right? Like why would you not do cloud and all of that, right? So in the enterprise, even if it’s an obvious choice, you know, it takes a while, number one, number two.

But I do think it’s, why is it an obvious choice? I do think over the last I. Year or so, if not the last three, I, I think it started three years ago, four years ago. These models have become so much better that they can truly augment human productivity in a significant way up to 80 to 90%. And customer service is definitely the starting point of it.

Did [00:22:00] not believe it when I was at Alexa, but, and at Alexa, technology was not there, there back then, but now it’s, and and the third thing is, I think the customers are ready. We were in a customer conference two weeks ago or three weeks ago, and I led a panel on generative use cases in the contact center or sort of like a round table discussion, and we proposed four or five different use cases and every customer, no one asked us like, what is it and why would you do that?

They were like, yeah, of course I can do this, I can do that. How about this? Another one. And we were all reflecting on that. And I think our hypothesis was why that happened. Chad g p d made it so real for everybody to play with it. Yeah. And so anybody from a middle schooler to a VP in the contact center can make a connection on like, what is this technology can feel it, use it, and then like dream of some ideas.

So the customer mindset is there. The technology is finally there. [00:23:00] So I think it’s gonna happen. It’s gonna take time. It’s, this is like, it takes, it’s a decade long process of transformation here. But kind of like cloud. Yeah, yeah,

Josh Lupresto: yeah. Fair. No, good point. And I think it’s, it’s gonna be fun, you know, just like at the beginning of any of these great technology emergence, 9,000 companies try to emerge and say, we’re the next greatest X.

And then you see this eventual condensing and boiling down, and you’re left with. You know, a, a couple hundred or maybe even less really great companies. So love that you guys have already figured this out and you’re already in the forefront of doing this from a CX perspective. So, awesome. Yeah.

Great stuff.

Ashish Nagar: Yeah, and, and there is a lot there. So there’s a lot there to figure out. Josh, like there’s a whole layer of explainability, observability, and configurability of these systems. I always sound like big words, but this idea of. Like, you know, you, you, if you are a contact center user or your customer or even an agent, if an AI is a black box and giving me an output, I [00:24:00] don’t believe it.

So you need to explain it. Yeah. And then I need to be able to observe it. How are you doing it in the enterprise? And then configure it, you know, and how do we make those systems? How do we make products which have those three things built in? That’s a big one when we are working on it, but I think there’s a lot of work happening there over the next few years.

Yeah. Good agree.

Josh Lupresto: Final, final couple thoughts here, Uhhuh. So your. Sales tips for our partners. So obviously, yes. They, they’re gonna find you at a, at a Polaris event. They might see you on a panel, they’re gonna run into you at some point, and they’re really gonna get immersed in this technology.

But I’m curious your advice for a partner who’s maybe in the CX space but maybe hasn’t ventured into this side of it yet, or is just in an adjacent technology, like I mentioned earlier, cloud or security or something like that. Yeah. What’s your recommendation? To go approach a customer, a prospect, maybe a wedge in or maybe a bolt on, that kind of thing to start talking about this?


Ashish Nagar: so two things. [00:25:00] One is I would have a very. Strategic, a very honest conversation with your customer counterparts on. Do you feel that you have a good pulse on, on how your contact center teams are doing, number one in a, in a scalable way, and how your customers are doing a scalable way? Right.

Like we were talking to a large telecom company a few months ago, and their c e o said, you know, we just had a large holiday promotion. Like, you can guy buy this unlimited data plan instead of for 30 bucks. For 20 bucks. And I have no idea how it went. Then I’m like, where my, where are my agents pitching it?

Right? What were my customers responding to? Like, like this, like, oh, at t has something similar. Right? But I don’t know, like I got a marketing report. Hmm. But I, so my two things I would say is, do you have a good sense of how your team is performing and do you have a good sense of what your customers are looking for?

And if the answer to that is yes, [00:26:00] then. You know, how are you doing it? And there’s a better way of doing it completely right? With less setup time, much better data quality, which you can trust. And the answer to that is no. You have to start there because it’s kind of a huge point in the transformation for these businesses, for the customer service, for these businesses.

You can save costs, you can increase revenue, and from also for our channel partners like you and your, your teams and your your partners. You know, there is frankly not much changing. Fundamentally, if you move from one C C a S provider to another, it’s telephony. You know, the value add would come from the intelligence, from the automation, right?

And you can create so much value there. For every a hundred thousand dollars of C C A S, you can sell, you can sell up to $50,000 of intelligence software, right? Mm-hmm. Where there’s an opportunity to upgrade and make a ton of money on, right. You can [00:27:00] and, and you can show them the pro, you can show them value very quickly.

You know, like if a c c A tool is working, it’s working, you know, there’s not much changes, right. So that’s a huge, it’s a huge revenue uplift there. Like we estimate there’s about, I. 5 billion in installed revenue in these intelligence solutions, which needs to be upgraded, upgraded to newer solutions, which is technology change.

Right? And then a lot of greenfield, which opens up with this technology change. I.

Josh Lupresto: Awesome. All right. Final thought. Ashish as we wrap this up, would love your perspective, look out in the future with this changing so fast. We can’t look that far out, but let’s say you look out 12 months what do you see changing in the customer experience?

And really what I mean with that is any different strategies. If, if we’ve given the customer, or sorry, the partners, the strategies to go to market and talk to their prospects in the way you just mentioned. Is there anything. [00:28:00] You’d advise with that as we look out over the next 12 months?

Ashish Nagar: Yeah, I, for sure.

I think a couple of things like the I think the automation rate will increase both at the front end. Like the bots will become better, but it’ll still like five years out, 10 years out, I maybe five years or 10 years is hard to see. I don’t see the bots taking over. So some things which remain the same.

I don’t see the bots taking over in a lot of complex things, but I do see automation grade improving. Number two, in terms of what to expect from customers and how your partners talk to customers. They should expect more from these tools. As I said, sentiment and transcription is not good enough. How can you get to real intent and real meaning of why customers are reaching out in real time and not needing to maintain the tool at all or, or nearly, you know, nearly at all and thoroughly, how can you synthesize and act on it in real time?

Right? And, and configure it [00:29:00] in real time. So that should be the expectation of your customers. So, right, like, I need synthesis, I need action in real time and data, which I can trust. And if that’s not happening, they’re selling you a 10 year old enchilada, which are five year. Nobody wants that. Nobody wants that.

Right? Because, because that’s, that’s the other challenge, right? And this is something, and I, I’ll be very honest with like, we realize this. There are long-term deals in this space. People sell software for one year, two year, five years. Yeah. And if a customer in this rapidly changing environment is bought into a tool whose underlying stack doesn’t change Right.

And they bought, signed the contract for five years, that’s a problem. Yeah. You know, open AI launched G three and GPT four, four months apart, you know, and you have a five year contract AI contract, like that’s.

We love the long contract, but just pick the right partner, right? You know, which is on that [00:30:00] technology trajectory. So another thing which we would ask customers and your partner. Store technology roadmap. You know, ask the vendors to show their technology roadmap, not what this looks like next, next six months.

What did you launch in the last, last one year? Yeah, let’s start with that. Prove it. Prove it. What did you launch in the last one year? Let’s see. Are you on the path? You know, and if, and the answer is like, oh, we launched a new like color dashboard and like now this data links to that data. They’re not in the AI space.

Yeah. Oof. Yeah. That’s a good,

Josh Lupresto: that’s a good way to end it. I love it. Ashish, that’s all the questions I got. I, I really appreciate you coming on and doing this with me today.

Ashish Nagar: I. Anytime, Josh. Thanks buddy. How really nice talking to you.

Josh Lupresto: Awesome. All right. I wanna give a big thank you to our listeners. Look, the podcast keeps growing.

We want to hear your feedback, experiences, comments, other things that you want to hear. So drop us an email stands for Next Level Biz Tech. Until next time, I’ll take us out. We’ve got Ashish CEO of Level AI. I’m your host, Josh Lupresto, SVP of Sales [00:31:00] Engineering, Telarus , and this is Next Level BizTech.