Your new favorite colleagues aren’t human
Summary
We all know AI agents can answer calls (with varying success rates depending upon their sophistication). But what if AI agents could also help CX leaders analyze their deployments, ask questions about their unique challenges and opportunities, and continuously improve?
In this episode of Deep Learning with PolyAI, Nikola Mrkšić sits down with Damian Sasso, Group PM at PolyAI, to explore how the field is moving from solely customer-facing AI agents toward agentic AI teams that can do all of the above.
We cover:
- Why MIT says 95% of AI pilots fail — and how to be in the successful 5%
- The role of QA Agents in monitoring performance
- How Analyst Agents like Smart Analyst surface business insights in real time
- The promise of Builder Agents that drive continuous optimization
- Why contact centers are shifting from call centers to command centers
If you want to understand how agentic AI is changing the landscape — one in which AI agents don’t just talk, but think, analyze, and build — this episode is for you.
👉 Subscribe today for more info about AI for CX.
Key Takeaways
- AI pilots fail without expertise — 95% of generative AI pilots don’t deliver ROI, but success comes when enterprises work with expert platforms, not by trying to build in-house.
- Agentic AI Teams are emerging — PolyAI is moving beyond customer-facing agents to add QA, analyst, and builder agents that form a full AI team for customer conversations.
- Analytics unlocks ROI — Tools like Smart Analyst give CX leaders instant insights into trends and customer behavior, turning contact centers into command centers.
- Continuous improvement at scale — With QA and builder agents, PolyAI enables enterprises to optimize and adapt quickly, far faster than human-only teams ever could.
Transcript
Nikola Mrkšić
00:05 – 00:28
Hello, everyone, and welcome to another episode of deep learning with PolyAI. My name is Nikola Mrkšić.
I’m the CEO and one of the cofounders of PolyAI. With me today, I’ve got Damian, who is one of our senior PMs, and he’s had a long career in financial services and analytics.
Damian, do you mind telling the audience a bit about you, and then we’ll kinda kick in with some news and then a deeper dive into our product?
Damian Sasso
00:28 – 00:47
Thanks, Nikola. Thanks for having me.
And, yeah, just, I mean, it’s been, it’s been a great time, being at PolyAI. I’ve had a long career in financial services, automation, communications tech, and even running a call center and support team at Bloomberg and then later on in my career at London.
Nikola Mrkšić
00:47 – 00:49
How big was that contact center?
Damian Sasso
00:49 – 01:24
I think it was about 30 people. We had both phone and chat based support, and it was like level Bloomberg was an interesting place.
It had level one, level two, and level three support, and we were a level three application technology support team with both phones and, you know, chat based at phone level like urgency because it’s financial services as you know. So, it was good.
I hope it prepared me well as I developed communications tools and then, you know, joined you at PolyAI, many. years later.
Nikola Mrkšić
01:24 – 01:28
it developed a high, you know, tolerance from trauma, which is good preparation for.
Damian Sasso
01:28 – 01:30
Exactly.
Nikola Mrkšić
01:30 – 01:45
Excellent. Excellent.
Excellent. Hopefully, we’re only a tad nicer than the finance people.
Awesome. And, you know, you’ve done a lot of work on analytics, and that’s how I think it’d be great to kinda like and that’s the focus of a lot of work.
Hopefully, we managed to touch on it today here as well.
Damian Sasso
01:45 – 02:16
Yeah. Exactly.
I mean, one of the things that really prepares you for, you know, working with data and analytics, which is really keen to our contact center customers, is spending time at a company like Bloomberg where Mike Bloomberg was quoted as saying, you know, in God we trust, everybody else brings data. So data is, like, a fundamental part of that organization, and I think it’s cool as you start to, you know, join other companies, you become very keen on how data operations need to be, to be productized and run and and and, you know, scale.
Nikola Mrkšić
02:16 – 02:53
I find it really interesting, like, the whole, like, you know, notion of product and analytics. And I feel like a lot of us, like, technical people because we’ve worked you know, we’re very numeric.
We’ve studied maths and related fields. We all think we’re good at, like, data and that we’re data driven.
And I think, you know, analytics is quite distinct, like, product discipline where, actually, I feel like I feel for you and people working with you versus all of us where we think we know stuff. But, actually, we’re often very under equipped in terms of, like, the theory and understanding of analytics and what good looks like and what needs to go into it to be usable at scale.
Right?
Damian Sasso
02:53 – 03:27
Yeah. I mean, it’s true, it’s tremendous.
And I think we we see an evolution in the marketplace in both data and analytics and more and more tools, and we’ll we’ll get into that that try to make things more and more accessible to the average person because, you know, the the the science of data, like, data science as a profession was around just the complexity of analyzing data. And, you know, we’ve seen that com you know, the just data become a huge funnel of information and complexity around the operations of it.
And now just to focus on products, trying to make that more and more accessible.
Nikola Mrkšić
03:27 – 03:27
Yeah.
Damian Sasso
03:27 – 03:34
to the average person and the average business teams, and that’s such a keenly important thing as the market evolves.
Nikola Mrkšić
03:34 – 03:58
Totally. Totally.
I mean, and I think they’re just, like, building some of those fail safes and, like, even notions of statistical significance is so important because, like, so many people now are increasingly applying, like, AI driven products and drawing conclusions based on datasets that where they don’t really know what what a statistically significant conclusion is, and that can lead to, like, like, so much wasted work, iterations, and, like, false conclusions. But, anyways,.
Damian Sasso
03:58 – 03:58
Yeah.
Nikola Mrkšić
03:58 – 05:04
I’m starting with a really bad textbook, so I’ll move on. Like, customarily, we usually look at kinda, like, one piece of the new satellites.
The one thing that I literally think I’ve discussed in every other meeting this week is the MIT report that said that 95% of generative AI pilots at companies have failed to produce ROI. And I think it’s something that’s really kinda like especially post, kinda like, g p t five and, you know, the slight disillusionment, four o getting re enabled.
Again, I think Sean and I were both on record? where we liked it. But to our credit, we did spend half of our episode about it talking about, just, you know, the courage of Sam Altman and, you know, deprecating all the models instantly.
I was just, like, envying him because with the number of customers we work with, we never get to do things like that. And sadly, neither does Sam Altman, so even the most powerful one got Raymond.
But, I think, you know, a bit of GPT, you know, disillusionment plus, like, this report. It’s sending waves.
Right? And, how do you feel about the report and the conclusions in it?
Damian Sasso
05:04 – 07:07
I mean I mean I mean, that’s just a crazy number to put up at the headline of a report, especially in a week where, you know, NVIDIA’s reporting earnings. And as a finance guy, like, you watch the market and you look at something like that and it moves markets, but, like, you know, you saw interviews come out with the with the with the person who commissioned and worked on the report, you know, after it was published, and you really dove into it.
And you saw numbers come out that really honed in on what a successful AI deployment looks like. So I think digging Fortune did an article, did an interview with the person who had had done the study and really tried to hone in on what a successful deployment looks like.
And some things that stood out was that 65% of successful AI deployments were around the expert vendors, companies that were contracting out, teams of vendors and products and platforms that had expertise in building AI products. And that the things that really they were honing in on that would be considered like a, you know, a failed deployment were really large scale enterprises without expertise trying to build something in house.
And then the other things that had really high ROI were operations and, you know, optimization. And they talked about sales and marketing being, like, areas where, you know, at face value look like higher ROI, but then you look at areas where you’re doing operations, you know, efficiency, and you look at things like customer service, call center, business analytics, ROI, those are areas that we that that, you know, were commented as being places where we’ve seen success, where where success has happened in the market.
I think that was really interesting because you have to dig into that data whenever you see a number like 95, and then you really start to dig behind. and see that actually you know, there’s a lot of context behind it that’s quite successful, and we’ve seen this.
And we’ve seen our platform and our customers have that level of success. So it is that I feel like that in many ways validates that.
Right? It validates. what we’ve seen in the market.
Nikola Mrkšić
07:07 – 08:52
100%. You know, I think, like, people talk about, like, pilots as if it’s, like, something that should convert.
I mean, for a highly disruptive new technology, getting deployed in all sorts of new ways, it’s supposed to be failing at scale. Right? And, you know, you look at, like, Blue Sky Investments.
You look at, like, some of the best seed funds, funds in general in Metra Capital. Historically, the best ones have the highest fail rate in terms of the percentage of companies that have failed because they took the boldest bets and many of them paid off massively.
But, like, the best funds have the highest percentage of companies that went to zero. Right? So I almost think of this as, like, why is this controversial? Right? Like, it’s hard.
But that 5% should multiply and deliver tremendous value exponentially, and I think it will. I think we’ve had a lot of you know, just when I look at, like, our pipeline and the number of people that have reached out over the past week, I think the strength of successful case studies there’s kind of, like, this power law where there’s, you know, a flight to quality and, those that have proven case studies at a very large scale, you know, where you’ve got, like, 2,000 agents doing the work of well, we have AI agents doing the work of a thousand people with large enterprise, and I think we have three to four right now.
Like, that just makes it easy to claim that you’re in the 5%. And I think many companies are trying to figure that out.
But I think one way to claim that is to talk about the ROI, right, and to show the value of these things. So it can be really interesting for you to tell us a bit about how our platform is evolving to both, you know, capture the analytics, the work of, you know, the different agents, AI agents inside the platform and the whole workflow.
So, I’ll pass it over to you just maybe to kinda look at the scene there.
Damian Sasso
08:52 – 11:19
Yeah. And I think what we’ve I think what we’ve seen certainly and just to comment on that, you know, that that study and kind of, like, the scale of the operations that that that we’ve undertaken is I think there’s a tendency for large scale organizations.
People are just, you know, using things like ChatGPT and, you know, Gemini and, you know, having a successful, like, search for some consumer thing. and then thinking, I can scale this across.
a large enterprise. Like, no problem.
And, obviously, when you really start to dig into it, that’s not the case. And that’s just to kinda summarize why you see, you know, probably irrational exuberance in trying to tackle large scale enterprise, you know, deployments, like from a company without working with an expert platform.
But our platform has been we’ve seen a lot of success in the building and scaling of our AI agents in the contact center. And where we’ve seen our product go now is really not just the idea of using agents to call in and handle customer service, inquiries, but also we’re now scaling things like our, QA agent, what we call our supervisor suite of agents, to derive both business insights, you know, see operational efficiency across, like, your data as one of our customers in our platform, Agent Studio, and using that using those agents to dig in to improve the continuous quality of your of your agent, but then really derive, like, those what’s your business insights? Where are your operation ROI opportunities? We’ve really seen a lot of that, and we’ll dig in a little bit on an example of something exciting we’ve recently rolled out called smart analyst.
Yeah. I mean, we our agent breakdown has really become more than just the agents that you you reach when you call into into, into our customers, but now those things of QA agents, analyst agents, and then really getting into our agentic capabilities with our builder agents that take the the knowledge you learn from our QA and analyst agents and, you know, bring them into, into circling back into the continuous improvement and optimization.
And like the study we just talked about, optimization is really, you know, where you are seeing the ROI in the deployments of AI solutions.
Nikola Mrkšić
11:19 – 13:42
Yeah. No.
Absolutely. And I think, look, you know, there’s many ways to supply the challenge, but, like, the agents themselves, the things that we build.
Right? The core agent, I think, even had an internal rotation at one point, we called it agent agent, which was hilarious. Right? But these are the workhorses.
They do the work, and they’re kinda like, well, really level one agents, level two agents maybe. But then the question of, like, how the thing continues to evolve is really important.
And then the QA agent is the thing about that kinda, like, you know, analyzes the performance. It has a notion of how well you’re doing.
Like, the basic thing that basic software in a human contact center does to, you know, analyze, like, sentiments, CSAT, MPS, predictive ones to highlight the calls that you maybe wanna listen to, you know, chat interactions you wanna look at to see what might have gone wrong so you can action it. This was historically used more to look at which human agents do well.
We are privileged in that. An AI agent is an AI agent.
It might have a different voice and might be doing different things based on an AB test, but it does exactly what we wanted to because at the end of the day, it is an application software. Right? So I think beyond these guys, there’s, like, your kinda, like, our grand architects who are really kinda like the managers of a contact center that direct the strategy, ask the questions to get people thinking and analyzing and figuring out what is the next thing to try, what is happening at scale, you know, ideally surfacing some product insights to the company around, like, its gen either its product or its customer’s preferences.
That’s a smart analyst bit. Right? And that’s really a way to have a conversation with an AI hive mind that speaks to all your customers.
And if you decide one day that, say, you’re trying to prevent churn on a group of people, you might offer them half a year off if they stay for an extra two years. But for another, you might offer a 20% discount if they stay for an extra year.
And, like, to me, I have no intuition over which one will work better. I think this is, like, you know, voodoo magic.
But an AI high mind gets to just, sir, yes, sir. Half of us will do this.
Half of us will do this. Right? And now there’s only a single person sitting behind the builder suite, right, and, in the platform, and they’re able to, like, issue that command to the AI hive mind and then collect feedback, ask these questions, and do that whole thing.
So I think that’s really, really powerful. Right?
Damian Sasso
13:42 – 13:42
Yep.
Nikola Mrkšić
13:42 – 14:00
And I think, you know, like, I remember a few calls with a client who shall not be named here where, you know, I think you spent many a month going through a manual process with one of our PSMs, Rob, where I think there was a listening session after listening session. And do you wanna talk a bit about that and kinda, like, what the queue agent.
Damian Sasso
14:00 – 14:00
Yeah.
Nikola Mrkšić
14:00 – 14:02
Did it run?
Damian Sasso
14:02 – 16:13
Yeah. And so you were starting to touch on it.
I mean, really, there’s two cool products that we recently rolled out and that really fall into that QI QA agent element. One we’re referring to as, agent analysis, and I think that’s where you were going with.
And the other way is something we’re referring to as PolyAI Score, which is these two two, key elements of our product. So agent analysis is really focused on allowing our customers to automate and bring in their own QA processes into our platform and assess calls based on any criteria they’re looking for.
It’s really about the idea of asking a question and getting an answer about a set of calls that come into your agent. And this could be things like vocal sentiment from the caller.
It could be things about, you know, trends that you’re seeing in certain calls. And that was really a powerful tool because it allowed what you just said, whereas a a process you normally had for QA ing human agents and then now bringing into the idea of how can you continuously improve the dynamic questions that come in and issues that get brought up to your now, automated, you know, PolyAI agent.
And the second thing was something we call PolyScore, which was really about assessing in a various number of factors how, your, your agent was performing in in interaction with your human callers. And so those two things are what we put in under our QA agent to, you know, use our platform.
And it’s really been something that’s focused on making our lives easier for, for that QA, QA, persona. And then the last thing that kind of falls into our our analyst agent product is something called smart analyst.
And this has really been, an agent within our platform focused on our business teams and anybody that our customers have that wanna derive insights and intelligence from the data that they get into, their PolyAI agent. So one of the things that really got me excited in my data background of joining you here at PolyAI and still gets me. I mean, you hear my voice as hoarse because I’ve been spending all week talking to our customers about this, which is exciting.
Right? And, yeah, I mean,.
Nikola Mrkšić
16:13 – 16:16
And we’re doing it. this on Friday night.
So, you know,.
Damian Sasso
16:16 – 16:16
Exactly.
Nikola Mrkšić
16:16 – 16:17
but
Damian Sasso
16:17 – 16:17
So.
Nikola Mrkšić
16:17 – 16:18
you so much.
Damian Sasso
16:18 – 16:49
that’s what it gets. me.
excited about this. But we’ve been, you know, we’ve been going through and showing our customers how they can get more and more intelligence out of the data that they get just from deploying their agent and letting their agent respond to their customer inquiries.
And we’re going and running our customers through our new product called Smart Analyst . That’s, you know, a deployment of our analyst agent, and they’re just typing in natural language queries like they would into any AI tool. You know? What are my best calls?
Nikola Mrkšić
16:49 – 16:50
Yep.
Damian Sasso
16:50 – 17:27
What trends do you see? What automation opportunities? Where can you get more and more operational efficiency? And they get answers back, tailored that we’ve tailored and fine tuned for them right from the data that they get from our agents. So that’s all been, like, super exciting, and it’s been really exciting watching our customers see that next, like you said, level three of getting data.
out of our platform. And it makes me know that we’re on the right track because I’m often talking to our customers and being like, did you see this yet? They’re like, I’ve already played with it.
I’ve already done these inquiries. I’m already getting this, and that’s awesome.
And that we were. we.
Nikola Mrkšić
17:27 – 17:28
I mean,.
Damian Sasso
17:28 – 17:28
that one.
Nikola Mrkšić
17:28 – 19:44
have the Datadog dashboard of usage, and I was hoping to see something in it. And then I scrolled one page, second, third, fourth.
I was like, oh my god. But whoever got access and actually tried it once has come back and back and back.
And, you know, I mean, my usage of our platform has completely started revolving around Smart Analyst. And, you know, maybe just to illustrate for people, I’m sure we’ll do another one where we’ll actually, like, play with it and show people how it works.
But, you know, if you’re a contact center manager and you hear anecdotally, my favorite example is, you know, someone’s been overcharged somewhere, you go like, okay. I heard it happen once.
It takes, like, days sometimes to hear of something happening with humans. Right? Because they’re so busy.
They’re working nonstop. Right? They don’t think they can’t afford I mean, they think, but, like, they’re not going with, oh, I got two calls, and usually I don’t get any calls about that.
They got two calls. Right? Over time, if they all start getting way more calls about it, you might get, like, water cooler conversations.
Like, have you also been you know, it takes a real explosion of it to become obvious with humans that it’s happening. Right? Whereas, like, if you get an inkling of it or maybe if you’re just, like, a human manager looking at these calls and you start seeing it, and maybe if your role is the kind of, like, analyst role where you look at this stuff, previously, you would have either talked to humans or did manual search or added a call reason with people tagging it, and then you have to retrain them to tag it.
Right? We had some automatic analytics that was always really hard to do because human conversations are notoriously unstructured. Right? Now you go like, okay.
Have people been asking about billing more? Tell me tell me the incidence of billing questions and you see and there’s no more need to, like, wonder or whatever. You ask the question, you get an answer in four seconds.
Right? Versus, like, with humans, this was often impossible. And, you know, there’s been, like, generations of software for, like, call analytics, for CSAT, voice of the customer, all of that.
And I’ve not met a single, you know, kinda, like, senior stakeholder for their prospects who told me it’s incredible, and I use it every day. Right? Whereas, like, smart analyst, you know, sometimes you don’t need AB testing to know that a product is good because they just pull you in.
And I’ve found it. People even use it as, like, a search bar.
Find me a call where people ask about this. I’ve not seen that call in the last 500.
Okay. Probably doesn’t matter.
Oh, yeah. Here’s seven close.
You’ll go down. Okay.
This is, like, statistically significant. Right?
Damian Sasso
19:44 – 21:07
Yes. Yeah.
100%. I mean, I mentioned this earlier when I was introducing myself.
Like, if I you know, running a customer service call center team at a company like Bloomberg for a period of time, you’d have to QA human, you know, human handled calls. And even there, you’d have thirty days go by before you really understood you had a trend that you needed to worry about.
And, I mean, this is in a very, like, you know, highly regulated, like, you know, financial focused market. And and when you look at now what we’ve started to do with customers leveraging our agents and then leveraging our tools like smart analysts to derive information immediately, you’re getting not just the benefit of a well tuned agent handling customer calls, but the fact that that’s there and then able to produce you data that you can query almost immediately.
And we just I was just on a call earlier with one of our customers and and our account team, and and, the customers, you know, was said, you know, you pulled me a list of calls the other day, and and our you know, someone from our team said, well, actually, I used the smart analyst tool you were just using, and that’s how I got you a list of calls you needed to look at. So you actually can do it yourself in the same way that I was able to do it.
And that’s that’s when. you really see that the power is there to get that information quicker than you would have gotten it at any point in time from looking at human, you know, human handled calls.
Nikola Mrkšić
21:07 – 21:52
Yeah. It’s unbelievable.
It’s kinda like, you know, the same way that lovable PMs, you know, do 80% of the work just to see if the whole thing feels like it makes sense. Like, now, you know, what would have taken a team of QA analysts and then people working, running, compiling reports, then looking at it, thinking hard.
Oh, I got this metric wrong. Let’s run it for another month to get a better taste.
Now it’s like, no. No.
No. No.
I don’t want that. I want this poof done.
And that’s just, like, unbelievable. It’s like an acceleration of the kind that well, I mean, it’s really fantastic to see.
What do you think the future of the contact center, its evolution into a command center looks like with this stuff? What’s the balance of AI doing work and all these agents that we discussed, you know, doing work versus, you know, just, like, human labor? Where do you think it goes?
Damian Sasso
21:52 – 22:46
I mean, I think it naturally turns into the upskilling of upskilling the teams of the teams in the contact center. And I think that’s it makes it a value add where it is now, like, the potential I mean, as to what I was excited about about this, you know, this space so much and and and sort of seeing the evolution of it.
Is it not just upscale the individuals that are working in the contact center because you now have the ability to you know, as we talked earlier, data as a science. The data scientist role is a highly technical role.
Now it can become something where the upskilling of the contact center empowers those individuals to really add more and more value to business teams and more and more value to organizations. But it also becomes something where the ROI behind the contact center and the contact center operations can now become clearer and more measured.
Nikola Mrkšić
22:46 – 22:47
Yeah.
Damian Sasso
22:47 – 23:22
in scale organizations. So it’s not just a call center now.
Right? Its data allows this to become a revenue generator and a revenue quantifiable part of your organization, and that can become information that’s just at your fingertips very, very available and very, very transparent. And that’s exciting for us working in this space and exciting for our customers.
Right? Because, you know, it isn’t just about it isn’t just about efficiency. It’s about efficiency plus, you know, value add and ROI and then the stuff we talked about earlier.
Nikola Mrkšić
23:22 – 25:49
Yeah. I mean, it’s really, you know, there was a whole period where, you know, data is a new goal, blah blah blah blah.
And, you know, I think it delivered, but not to the extent that people had hoped. Right? There was definitely a bit of a hype cycle.
And the truth is, it’s kinda like, you know, to use an unfortunate analogy, here, you know, discovering oil is kinda like Jenny I’s defracking. Right? It’s kinda like, oh, just put a bit of that.
You know? You get to a point where there’s just way more of it. It’s a lot more accessible.
And all of a sudden, you know, the equation of how much of it we have changes completely and profoundly and how much of it we can access. Right? So I think, you know, the excitement for a lot of companies that did that analytics before was there because it was always clear that the value was there.
Right? But both the mechanics of getting those thousands of people working for you to do all that, those different movements that someone had time to think of. But it was just physically impossible to maybe, you know, channel 20 different initiatives to an agent group where they did, like, a 100 different call types.
Right? It was really hard to tell them, hey. If you hear someone calling you about that, try this sometimes and try this other times and, like, label it as that so we can see how it works.
Like, that’s too much. Right? And the truth is if you have 10 of these kinda, like, spiritual PMs in the contact center trying out changes with the product, and their product is how they interact with their customers.
It’s a hugely important product. Right? In many ways, it’s like the leading indicator for everything from, like, NPS churn.
The equation is powerful. And, you know, the larger the enterprise we work with, the clearer their understanding of, like, what one NPS points means to their revenue.
I remember one where they were like, you know, an extra NPS point, and that says a bit more than a million dollars of profit a day. And I was like, woah.
Like but the issue was always that the equation of how it works was really hard to control. And even all these improvements were hard to roll out through humans, so they were rolled out quickly.
They were not rolled out quickly. It was very slow.
It was hard to know if you made progress. or not.
Whereas now, like, that one controller of the platform can just work on that one part, change the system for it, and then get to look at those calls very precisely to pull them out, to ask questions, reformulate their hypotheses, and just iterate. It’s kinda like they got a programming language for their contact center, whereas previously, they only really had, I don’t know, scrolls on on, you know, pyramids where it took a long time to write a single word and the script wasn’t perfect, and I’ll stop abusing the analogy.
Damian Sasso
25:49 – 25:49
No.
Nikola Mrkšić
25:49 – 25:49
But
Damian Sasso
25:49 – 25:49
No.
Nikola Mrkšić
25:49 – 25:50
yeah.
Damian Sasso
25:50 – 26:45
No. I think that’s what you’re seeing.
I mean, look, that’s, I think, where you’re seeing our product and our focus for our customers evolve because I for I mean, I remember when I when I just joined you and and and I think I had a conversation with Michael Michael Chen, our our VP of partnerships, and he had said to me, you don’t you need to understand the time it takes to roll out a change in the contact center, and human beings to be able to react to that takes weeks and weeks of time. And, you know, you touched on it earlier.
It also takes the same amount of time sometimes to derive trends from human humans. interactions that happen in the contact center.
The turnaround time on making and you know, analyzing a trend and then making a change based on that trend when human beings are involved can be weeks upon weeks upon weeks upon weeks. And now you’re looking at tools that we’re providing both from the PolyAI agent that’s interacting with, you know, our customers’ callers and our customers’,.
Nikola Mrkšić
26:45 – 26:46
Yeah.
Damian Sasso
26:46 – 27:01
You know, our customers’ customers. And then the way that you can turn our analysis tools like smart analysts into improvements, we’re now saving not just you know, we’re saving weeks and weeks of time, and then we’re also providing you with a finely tuned,.
Nikola Mrkšić
27:01 – 27:01
Yeah.
Damian Sasso
27:01 – 27:28
you know, a tool that’s really specifically for the data. It’s not a cookie cutter thing.
It’s something that’s been tuned specifically for. the interactions that you have.
And, that’s unique. And that’s you know, as a product manager for someone like let someone like me, that’s why you see my me getting excited about this because it’s a unique challenge that we’re tackling in a way that, you know, we see the excitement from our customers and see us moving the ball on in customer service to, like, the next the next level.
So.
Nikola Mrkšić
27:28 – 29:44
No. No.
And listen. I can tell you that, you know, Yun Zhang, our former CEO, had this saying that I always find incredibly strong, like, information in companies flows upstream in sentences, not in, like, charts and graphs.
Charts and graphs are, like, middle management, the people who really need to, like, work on it. And then there’s, like, the one key piece of insight.
Right? And I can tell you that previously, you know, I spoke to the CEO of one of the largest European banks this week. Right? And, you know, two topics.
One was, like, smart analyst and just, like, the point where, you know, in a language other than English, we were, like, varying it. And it was just, like, talking to an extremely insightful person who knows everything in their contact center where, you know, the person was looking and going, like, okay.
And, like, it all started with just a conversation from the MIT report where it’s like he’s like, we know this is working. I see a lot of other things failing.
Like, what else can I get out of it? And I couldn’t wait. And I showed him the smart analyst.
Right? And I think at that point, like, he’s like, well, how’s this? How’s that? How’s that? Oh, I knew this was right. And I’m like, you know, honestly, there was, like, a boyish, playful, like, excitement about this thing, and it’s like, that’s something.
Right? Because it makes the whole and, also, like, for that ROI, right, like, that one person with that idea, I think, you know, you don’t try out a lot of things if the barrier to entry and ROI is so high. And if you have to retrain humans, derive a trend, and maybe you do it once, twice, and you see, like, 1% difference, something you can’t substantiate.
The next time you show up at a staff meeting and, you know, in the morning when they kinda, like, have a roll call and tell people what’s new, they go like, hey, Damian, like, stop, man. Like, your last two things were down.
They’ll be like, unless you’re sure or someone you got, like, a c level approval to do this. Like, come on.
Don’t bother us, man. Just let us, like, pick up the calls we have to pick up.
Right? And that discourages you from experimenting. Whereas now you’re completely not interfering.
There’s no human kind of weight in terms of having to take away from their time doing something else to roll your thing out. And that, I think, is just like a multiplicative positive effect on how contact centers can provide value for the business.
Damian Sasso
29:44 – 30:11
Yeah. Hugely.
Hugely. And I think that’s where we’re gonna continue. I mean, that’s where we’re gonna continue to see this, see the contact center operations evolve.
And we’re, you know, on the front lines of building those solutions, and I think that’s what yeah. I mean, it just makes it exciting.
You talk to our, you know, our customers and our leaders and our customers all the time. You know, talking to customers, you see the excitement around getting that level of value that we’re able to provide and that we’re able to see, you know, coming from these solutions.
Nikola Mrkšić
30:11 – 30:11
Yeah.
Damian Sasso
30:11 – 30:45
So it’s really it’s really, yeah, it’s really interesting, and I think we’re going to see more and more of the of the analysis and, like, data analysis tools start to cycle back into that kind of, you know, overly buzzword and used agentic, but it really becomes that agentic flow of I’m now taking this information. I’m gaining quickly with those tools and quickly making changes, and now I’m getting closer and closer to having that level of agentic automation.
And, you know, it’s something that’s hugely powerful for our customers.
Nikola Mrkšić
30:45 – 31:24
Absolutely. Well, look, Damian.
I’m in New York next week. I think we’ve hyped up smart analysts to the point where I think we owe another episode where we demo that next week.
So, maybe we can even do it in person. And, thank you for joining me.
I am super excited about all this and, you know, kinda like how Shane and I did, like, I think in the end of three episodes, on, on telephony, which caught cumulatively, I think, almost a million views. I think, I’m excited to kind of let’s commit maybe to kinda, like, showing people just what smart analyst is and how powerful it is, and maybe we can, hopefully, provide some interesting commentary on, like, how that gets rolled out, interchange.
Damian Sasso
31:24 – 31:34
I’m excited. I’m excited.
Maybe we can go for those big, big, big views numbers because I think we’ll have a lot of our customers and our, you know, prospective customers excited to see it. So.
Nikola Mrkšić
31:34 – 31:50
Absolutely. Well, Damian, thank you for joining me tonight, and we’ll speak again, and we’ll continue working on all this.
Thank you for all that you do to make this happen. It’s and it’s really the whole way of excitement with all the polymers, with our clients, and, we’ll we’ll show more.
Damian Sasso
31:50 – 31:55
Yeah. Sounds great.
Thanks again for having me, Nikola, and let, you know, have a good one.
Nikola Mrkšić
31:55 – 32:04
Yeah. And, of course, guys, share, like, subscribe, and, yeah, we’ll get in the habit of saying that.
And everyone, thanks for listening, and see you all in the next one.