The Rise of AI with Swapnil Jain of Observe.AI

Author: Scott Kinka

The Bridge Podcast observeai LogoOn this episode of The Bridge, I’m joined by Swapnil Jain, CEO and Co-Founder of Observe.AI. We’re talking about the rise of AI and solutions that are making real-world changes right now.

Observe.AI is the fastest way to boost contact center performance with live conversation intelligence. Built on the most accurate AI engine in the industry, Observe.AI uncovers insights from 100% of customer interactions and maximizes frontline team performance through coaching and end-to-end workflow automation.

During our conversation, we discuss what lead Swapnil to build a conversation intelligence platform to improve productivity and how Observe.AI combines AI technology with human agents for improved performance and customer outcomes. We also talk about how the rise of AI had lead to more conversations about responsible AI usage and how we can balance innovation with societal impact, and so much more.

Topics covered in this episode:

  • How the pandemic and hybrid work model escalated the need for AI tools in contact centers.
  • How real-time monitoring helps supervisors identify potential sales that may not happen and intervene to convert the sale during the call.
  • The need for guardrails and protocols to ensure the right use of AI.
  • Recent advancements in AI, particularly in generative AI and large language models
  • Responsible use (and potential misuse) of AI and the importance of considering the societal impact.
  • The role of empathy and human emotions in customer experience.
  • Customer concerns regarding data privacy and security.
  • Self-policing and government involvement in regulating AI.
  • Swapnil’s predictions of large companies emerging based on generative AI and the potential threat to established companies.

The Bridge Podcast with Scott Kinka - Swapnil Jain Observe AIABOUT SWAPNIL JAIN

Swapnil is the Founder and CEO of Observe.AI. Observe.AI, a leader in Contact Center AI, transforms customer experiences and improves agent performance by helping top brands analyze 100% of calls and streamline quality assurance workflows. With Observe.AI, businesses transcribe every call with high accuracy and coach agents while gaining full visibility into their customer service operations. Observe.AI brings the power of agent assistance, automatic speech recognition, and Natural Language Processing (NLP) to modern contact centers and their frontline teams. Observe.AI is trusted by more than 150 customers and partners, including Root Insurance, Alcon Laboratories, Tripadvisor, and Pearson. Backed by Menlo Ventures, Next47, NGP Capital, Scale Ventures, Nexus Ventures, and Y-Combinator, Observe.AI’s headquarters is in San Francisco with an office in Bangalore, India. For more information, visit








Scott Kinka (00:06):

All right. Hi, and welcome to the Bridge. My guest on this episode is Swapnil Jain, he’s the c e o of of Observe, AI Swapnil, how are you today?

Swapnil Jain (00:17):

I’m very good, Scott. Thank you so much for having me. Really appreciate the opportunity to speak with you and the listeners today.

Scott Kinka (00:23):

Fantastic. Um, looking forward to it. We’re gonna give, certainly in a couple minutes, an opportunity for you to tell everybody what Observe AI does. Um, but you know, these are kind of in, in a lot of ways, informal conversations between, you know, tech people. Right. At the end of the day, executives. So I’d like to personalize at the beginning. So let’s just give us your story. Tell us a little bit about you first.

Swapnil Jain (00:46):

Yeah, no, absolutely. So, I’m, I’m Swapnil, uh, I’m, uh, the co-founder and CEO of, of Started this company about, uh, five years ago. Uh, my personal background is born and raised in India to, you know, a middle-class family. Um, did my education there, did my college there, and then from there, moved to us in 2012. Um, so came here to work at, uh, Twitter. Right. And the Twitter I’m talking about is not the Twitter that we know these days. Uh, right. It was a different time. Um, you know, we had Dick Costlow as a CEO, great time. The company was doing Fabulously well, you know, was one of the best tech companies out there, uh, in terms of Facebook, Google, Twitter, right? So, amazing run there, uh, for about, uh, amazing run there for about three years. And then in 2015, I had a lot of FOMO.

Swapnil Jain (01:35):

Uh, but you, you, you would laugh at this, but, uh, a big part of how Observe started was through FOMO. Uh, so a lot of my friends were starting companies, uh, raising a lot of money. <laugh>. I was sitting in San Francisco reading about that. And, that was the trigger for me, that, hey, I gotta do something of my own as well. Uh, so I jumped the ship, uh, did two years of research trying to, trying to find an idea through a lot of hard work in contact centers, came up with the idea to start up of ai.

Scott Kinka (02:01):

That’s an amazing story. Um, so you’re still in the Bay Area?

Swapnil Jain (02:05):

I am. So I live in San Francisco, uh, bay Area. Got married about, uh, 18 months ago. So my wife and I, we live here in San Francisco, uh, in a beautiful downtown. So that’s, that’s where we are.

Scott Kinka (02:15):

Fantastic. Did you get married in the US and then go back and get married in India and do the whole week-long thing?

Swapnil Jain (02:24):

I actually did. So my entire, if we, if I look at the, the process, right? Yeah. Had like multiple events, two of them here in the Bay Area, then I bought four of them back in India. Uh, so Right. So it was a pretty long process, and by the end of it, we were really, really tired. Uh, so, but yeah. You know,

Scott Kinka (02:41):

I can

Swapnil Jain (02:42):

Imagine pretty, yeah, and, you know, every, every sort of celebration involved the different dress, uh, right. So you can imagine buy, like so many dresses. It’s a lot of work. A lot of work. But it was fun.

Scott Kinka (02:55):

I have a fr I asked that only. I have a friend, um, a very close whose son, um, you know, was, and he’s, you know, um, they’re Native Americans, you know, Americans I should say, from, you know, from here. And he married an Indian girl, and they went back and, um, did the whole scene, right? So he’s sharing the whole thing with me, and he is as apple pie American as you can possibly imagine, and he is sending me photos of him sequined head to toe, you know, throughout the entire process. So, uh, and I was sharing those pictures with my wife, and she’s like, this feels like it must be the most amazing wedding experience from a parental perspective. Right? But I can imagine the pride and groom are very tired at the end of it, right?

Swapnil Jain (03:37):

Yeah. They very, they are. Uh, have you been to an Indian wedding? Scott?

Scott Kinka (03:41):

I haven’t. Only, you know, vicariously through my friends’ pictures, right? <laugh>. That’s it.

Swapnil Jain (03:45):

You know, I would strongly recommend one, right? The food, the dancing, obviously, the music, everything is phenomenal. You have a lot of fun. And, of course, the clothes,

Scott Kinka (03:55):

I think I would be afraid to dance at an Indian wedding because everything looks like a Bollywood video. You know? You’re, everybody’s so well coordinated. Is that just in the movies, or is that what the wedding actually looks

Swapnil Jain (04:06):

Like? No, no, that’s, no, I think that’s a very, very, it’s a wrong perception of what it looks like. Okay. In fact, the Indian wedding, uh, dance is the loosest form of dance where you basically move, and you’re fine. You know how we go in clubs, right? You’re standing, and you’re like, sure, dude, that’s it. Yeah, you do whatever. No, no one cares, uh, uh, dance. So it’s, it’s actually very fun.

Scott Kinka (04:26):

Well, if there’s anyone out there in our listener base who is planning a wedding in India and they’re just looking for a random guest, um, my wife and I are happy to be invited, and we will certainly join you there. Um, well, that’s a great story. So you did two years of research. Yeah. And I think that’s an interesting stepping-off point before we get to the kind of the, uh, observe AI pitch. So let’s just talk about the two years of research. I mean, you saw a need; you came from Twitter, geez, that could be a whole episode in and of itself, and maybe we’ll revisit that in the future. But let’s focus on, uh, observe for the moment. What was the need you saw? I mean, and why, you know, why did you end up focusing? Were you focused?

Swapnil Jain (05:04):

Yeah, so as part of my research, um, I, I spent two years and a big part of that, which was towards the end, for about six months, I spent a lot of time in call centers. So I was in the Philippines, um, Manila, where I would go on these contact center floors, right? And, and the reason these are called floors, um, anyone who’s worked in contact center would know, you know, these are like chain task floors. We have like hundreds of people sitting in, you know, taking calls, writing emails, writing chat rights, so busy, um, uh, busy flow, right? So I would go on these floors and just sit there and watch people, right? I would sit and and look at an agent and see what the agent is doing. You know, try to look at their workflows, try to look at their systems, try to do how they’re taking calls, try to look at how they’re being compliant, try to look at like how they’re doing after call work.

Swapnil Jain (05:50):

I would look at everything. Then I would look at a QA person and see what they are doing, uh, right? How they’re evaluating calls, reviewing calls, looking at supervisors, look at an operations person, right? So I spent six months doing that, uh, right. And the interesting part about that journey was all these call centers in the Philippines and India were actually working, uh, PST and e-SDRs. So this was all at night, right? So, the actual work would start, you know, at the Philippines, let’s say, 9:00 PM and then finish at 5:00 AM. Uh, right? So just these opposites. So my entire night was just looking at these people. And what I realized, um, there, Scott was, so my background is software engineer, um, and I’m, I’m a computer science engineer. I was a tutor as a software engineer. So I’m a builder, right? And I’ve lived in the Bay area before I did that research.

Swapnil Jain (06:35):

And in Bay Area, you know, we all talk about AI and technology and SpaceX and test laws and iPhones and, and, you know, amazing things, right? But when you go and look at these people on the floors, you would realize how manual and legacy everything is for those people, right? Mm-hmm. <affirmative>, their workflows are archaic, uh, right? They’re still using pen and paper, they’re still listening to calls manually. They’re still doing coaching, you know, once a month on one or two calls. Everything felt like it, it needed to be disrupted and changed, uh, for the right reason. Yeah. Uh, right. So that’s how it started, that there’s a massive opportunity to really come in and use the latest and greatest of technology, especially artificial intelligence, to build a new class of software for, for contact centers so that they can do their job 10 times better and faster, right? So for agents, managers, and supervisors, everyone can be 10 x productive. So, that’s how that’s how Observe started.


Revolutionizing the Contact Center

Scott Kinka (07:30):

Did the pandemic and hybrid work model change the need in any way for you? I mean, you talked about these big monolithic floors, unfamiliar, as you’re aware, right? We work together on, big contact and our deals. Um, we also managed b bp, we managed BPO outsourcing to the Philippines, and, you know, and, and other offshore destinations we’re super familiar. Um, but tell us, how did that change a lot? Did the need change, I guess, during the pandemic, and then, you know, has it settled back to normal? Or are, is there a new normal even in contact centers?

Swapnil Jain (08:06):

Yeah. No, I think it actually, uh, I would say it escalated, uh, during the pandemic. Okay. Right? Because our key offerings were how AI can help humans do their job better, right? Yeah. So now, in a remote world where a supervisor is remote, where a manager is remote, where a QA person is remote, an agent is remote, right? So they’re all over the world, right? Not even all over the country, right? You know, their agents are in India or in the Philippines, or Southeast or wherever they could be, right? The manager could be here in the US or South America, right? So what that meant was there’s a, there’s a much higher need for software, because imagine, right? Previously what would happen is if you were in a contact center, a QA would just walk the floor. A supervisor would just walk the floor, right? And then kind of, you know, get a sentiment of which call is going bad, which call is going well, which agent needs help, where need to jump in, right? But now, imagine you are a large contact center with 10,000 agents; your QA is sitting in one place of the world, has no idea how their agents are doing, and they can’t even understand their world, right? So that’s where, that’s where we saw a higher need for AI tools for these people to do their job better and faster, right? So more intelligence, more understanding of conversations, a better understanding of what’s happening in these conversations. It kind of became, uh, became a much higher need, uh, in the pandemic world.

Scott Kinka (09:23):

That makes sense. So that’s probably a good stepping-off point then. So for those of our listeners who are not familiar with observe ai, uh, and I always find it interesting to compare these conversations with executives to what we’ve heard from the salespeople in the field, right? Or what’s on the website. Yeah. Um, give us the elevator pitch from your desk as the founder. What’s, tell us who is observe AI and what is it that you do?

Swapnil Jain (09:48):

Yep, absolutely. And I think, uh, it’s kind of interesting how you put, you know, a pitch a sales rep is different than a pitch on the website, because I, whenever I pitch, people say like, Hey, I’ve not heard that pitch. Why is that different than what I have been pitching? I’m like, because in my head, my, the pitch I’m telling you is the one that’s one year ahead of where the website and sales team is. You know? Sure. I have that pitch in my head and, and the team will follow through. But yeah, of course. So, you know, observe is a conversation intelligence platform, um, by, we help businesses, uh, and contact centers improve outcomes from the contact center. So if you’re a sales contact center, we help you improve revenue, um, we help you improve conversion rate. If you are a service call center, we help you improve your cost by reducing your care cost, by reducing your a h t, by improving your csad, by improving your after call work.

Swapnil Jain (10:34):

And if you’re a collections contact center or a revenue cycle management contact center, of course, we help you with more compliance and with more revenue. And the way we do that is by, by analyzing every single interaction that you’re doing. It could be chat, it could be voice, it could be email in real time, and, and analyzing that using speech and natural language and providing the right insights and next best action to every single persona in the contact center, right? So this means that the core, there’s an AI platform, and then on top of that, we have a bunch of applications, right? So starting with agent, we have an agent assist application, which listens to the call in real time and then guides the agent on what to say, what not to say based on what we have learned from a knowledge basis, what we have learned from a previous conversations, what we have learned from a top agents.

Swapnil Jain (11:16):

So imagine this, this is like an AI and a human doing the call together, right? So the, the latest word for that is a co-pilot, right? It’s a co-pilot to an agent where the agent and AI are flying the plane together. Agent is leading, and the AI is a co-pilot, right? It would take all the notes from the call, so it would summarize the entire interaction. It would automate the after call work. So the agent doesn’t need to do after call work, right? It’s kind of an assistant to the agent, right? It’s guiding how to handle an objection, how to handle a negotiation conversation, how to best pitch the product. Then we have a supervisor application where in real time supervisor’s getting a view of what’s happening with every single interaction in the, in their floors. They can see all the interactions, they can listen, they can barge in, they can chat with the agent, and more importantly, they can set up the right alerts if they need to know in real time, if a conversation is about a compliance issue, a legal issue, an upsell opportunity, essentially we are giving them the opportunity to make the correction and coaching in the moment versus after the moment.

Swapnil Jain (12:17):

And then of course, there’s a full patient performance, um, application for the supervisor where a supervisor can take down notes, do coaching sessions, all of that. Then there’s an application for QA person where a QA can come in, review the calls much, much faster, analyze the right amount of calls, look at, you know, the targeted calls to review, like the sentiment is bad, or there’s a compliance issue somewhere, and then automate that part of the process, right? And then, then there’s an application for operations to do full analytics on the contact center to find out what makes the contact center work, why is my ag low, why is my sales conversion high? All of that from that data. So that’s sort of what op what observer I does, uh, as a, got

Scott Kinka (12:51):

It. So there was a lot there. Like, let me unpack that for a minute. Um, yeah. And sort of take it one level up. So, and you did a great job of starting from the business outcomes and moving it down into the tech. I’ll start from the tech and move up to the business outcomes, and we’ll sort of do it in reverse. So bottom line, you know, you mentioned that big gigantic floor, people physically walking, you know, inefficient, where we went home, it got even more, less, less efficient, right? Yeah. Um, the reality of it is, we’ve got humans. Humans are on the phone. We need to make sure that we track them, monitor them, and prove their outcomes for the customers. And what you’re doing is a couple of things, right? One is, um, all the post-call stuff mm-hmm. <affirmative> chats, emails, call recordings, which are being transcribed, are being wrapped up, analyzed, run through nlp, and you’re getting back transcriptions and then using logic to say via sentiment, this was bad. This was good. Yeah. You were on script here, you were off script there. Just taking this voluminous amount of information and making it something that you can look at every call and see trending on to make agent management easier. Fair on that portion? Yeah.

Swapnil Jain (13:55):

Yeah. Okay. Yep. And,

Scott Kinka (13:57):

And then on the, just bear with me on the realtime part, right? And this is, I’m going to hit you with some questions around realtime for a minute. While all that’s happening, the reality of it is, is you are then also able to kind of get that heads up display on the activity that’s going on now, right? Based on your language modeling. What are the words we’re trying to avoid? What are the words we wanna ensure that they use? We wanna make sure calls are getting wrapped up effectively, et cetera. Um, and you’re providing really real-time intelligence to supervisors, Hey, this is, might be where you want to jump in, right? This might be where you might wanna pull the agent off the phone because they’re distracted. Here’s an area here that you should, you know, that’s acute. Um, I think the one that I get, how’d I do Okay. Describing you?

Swapnil Jain (14:42):

Well, you did actually pretty well, uh, we’re looking for, uh, for a new leader. So maybe we should talk one-on-one. <laugh>.


Real-Time Analytics Drive Sales Success

Scott Kinka (14:51):

I love that. Uh, so, but let me, let me ask a question because, you know, look, I’ve got a legacy. We, you know, at bridgepoint, we have all CX team, you know, we’re doing this all day. You know, my, in my history, I was at Evolve ip, we had a contact center platform that was in the magic quadrant. I was involved in working with those teams. We were always challenged with the, the gap between, you know, what realtime actually means. Let’s be clear, right? Because the reality of the timing of how the NLP engines works is, you know, they’re, you need a little bit of language to get enough context to be able to Yes. Make sure that the correct. So when you say realtime, I just wanna make sure you and I are on the same page. How realtime is realtime? How far behind are you kind of in terms of that heads up display as interactions are happening?

Swapnil Jain (15:41):

It’s actually, you would be surprised. It’s actually very realtime. Uh, okay. You know, it’s, it’s, so I would say from a, from a time lag perspective, right? Let’s say, say something, uh, right, you are looking at maybe one second before you get the response back on your screen, right? And the reason for that is, right, the data transfer is not happening from your machine itself. It’s a cloud to cloud transfer. So, okay, the way, the way ariel time works, right? So if, let’s say you’re using, you know, a good C cast provider, like a five nine or a Genesis, or a talk desk, any of these guys, right? Yep. The way we get the audio stream, let’s say you’re an agent, I’m a customer, right? The way you, the way this audio stream is being transferred is not through your system or my system, but it’s being transferred through five minutes cloud, which is in us, in a certain region in America on aws.

Swapnil Jain (16:29):

We are in a certain region in aws and AWS to says, these are extremely, extremely fast tunnels to get this data right? So it’s, it’s really fast. And, and I think if you’ve not seen the demo, we would love to show you the demo and, and have you talk to some of our customers, you would see like it’s actually really fast, right? Got, now that being said now, so there is fast, but there’s one part which is also interesting here, context, when you are doing real time, you only get the context until now, you actually don’t know the context after, right? Yeah. So there are some applications, right? If you’re doing post-call, you have the entire context, right? You can do a better job when you’re doing real time, because I actually only have context until now. And if we make decisions based on what I’ve heard until now, right?

Swapnil Jain (17:10):

So it’s limit certain applications, um, that you can do it. So the big one I always use is notes, right? Mm-hmm. <affirmative>, when agents takes notes, right? And you’re doing an after call work, observe won’t do the after call work once the call is over, because now I have the entire story. If I’m drafting the notes while the call is happening, it can go in multiple places. And the notes I might have taken might not even be relevant until now. Um, yeah. Right? So this is an example of an application where, you know, you gotta wait till the call is over and then in one second the summary pops up. Uh, right? And it’s all real time and good before, and the agent can take that and put in the CRM in other places, but you can’t do that as the call is happening because you don’t have the context on the other side, uh, of the point.

Scott Kinka (17:50):

Got it. Makes sense. So with the idea of context and sort of the timing on the, you know, if you were to add context, the reality of it is the natural language engine would start giving you words, and then as it adds context, it would sort of improve the beginning of the sentence, right? Exactly. Kind of as you go. Um, just thinking about that modeling, can you share with me some of the, can you situationalize realtime for me? And what I mean by that is, like in the a in the supervisor heads up display, let’s forget about post call, during call. What are the kinds of things that, how do customers apply that? I mean, what are the kinds of things that they’re looking for that throw the red flag up? You know, when does the flag get thrown on the field? Gi give, give, give of the listener some examples?

Swapnil Jain (18:35):

Yeah, absolutely. So there are two most common use cases we have seen, right? One is a sales opportunity use case, right? Um, I have a team that is running sales, right? You could be doing an outbound sales, it could be an inbound order taking sales, right? And every sales matter, every call matters, right? So now your team is, sure, you know, these calls, running the sales process, running the sales script. You have your supervisors who are, who have set up the right alerts on when a sale might not happen, right? So now I’m sitting as a supervisor, right? I have this dashboard, I’m seeing like my 20 agents are doing these 20 calls, right? And on 19 of them, I’m getting this green indicator that the sale is gonna happen. And, and they’re following the sales process and it’s good on one, I see red, right?

Swapnil Jain (19:17):

And I’m like, oh, seems like the sale is not gonna happen. And then I have to act now if I have to convert the sale, yeah. If I act tomorrow, the sale is, is over. So now, now is the opportunity for me to make additional buck, right? So I go in there, I click and observe will precisely tell you that. Why are we saying it’s red? And, and in this case, why we think the sale is not happening. And that’s the opportunity for you to either barge in, right, take over the call, or go ahead and talk to your agent and say, Hey, propose this offer, or make this, uh, uh, make this discount and get the sale done, right? So very strong revenue use case, right? Yeah. You don’t wanna lose your sales.

Scott Kinka (19:56):

And the things that it’s hearing that are, that make it red, give me those specifically. So it’s, yeah, maybe there are words that are non buying words. There are chains of texts that we’re expecting. Are people to say that on that side of the recording are not being said, is that those are the kinds of things that make it turn red.

Swapnil Jain (20:16):

So let’s use an example, right? So let’s say I’m a telco company, uh, right? I am, uh, um, this is a plan upgrade call, right? So all my agents, you know, when you call in and say, I have a plan, I wanna change my plan. So they route you to a certain set of agents, right? In telco, you would see agents ranging from service agents, retention agents, upsell agents, uh, right? Your service agents are regular. Like, Hey, I have a problem with my bill. How do I pay? Then you have retention agents, which are like, yeah, I’m canceling, I’m gonna move to a different provider. That’s retention agents, right? And then you have the upsell agent. So now you’re talking to an upsell agent, right? The upsell agent’s job is to, you know, upgrade you to a higher plan, so get more money from you.

Swapnil Jain (20:53):

Yeah. And then also maybe lock in a watch or an iPhone or something else along with that plan, like, Hey, you are running this great discount, all of that, right? So e each of us has a kpi, right? Each of us has like to sell two iPhones a day, or like three watches a day. So I, I’m running this, uh, contact center now. I have these agents who are doing this now, and I have a sales process, right? Uh, I, as an agent, I’ve been given the sales process that, okay, this is my pitch. These are the available offers to you. These are the available discounts that you can offer. And this is the value pitch. And this is how to handle objections, because you’re gonna get a lot of objections in the call, right? So now from using this and using the past interactions, observe is DESI designed what a great sales call looks like?

Swapnil Jain (21:36):

Yeah. Right? So now we say, if the customer is showing you, you know, intention to buy, they’re talking about, you know, contract. They’re talking about, tell me more about the pricing. Tell me more about when it can start, when it’ll be delivered. These are buying signs. But if you’re hearing from the customer things like, yeah, you know, let me talk to my wife or husband, Hey, you know, hey, let me call you two days later, right? Or let you know, let me think more about that, right? These are signs where they’re not buying. So we use some of these signals and then alert the supervisor that, hey, this agent is not gonna make this sale because they’re not able to handle this objection. And they can see in the dashboard what is the objection. And they can either barge in, you know, tell the agent and how to best handle, got it. Put in an offer, put in an offer right there, that agent. Make this offer to the agent because the supervisor knows that this can convert the agent, convert the call.


Responsible AI Usage: Balancing Innovation and Societal Impact

Scott Kinka (22:23):

Totally. So that’s it. You brought up a really interesting point that I wanna explore a little bit on, right? Um, you’ve gotta observe ai, you’ve got AI in the name of the company. Yeah. Right? So I’m going to broaden the topic a little bit cuz we, I, I honed in very specifically on practical in world applications for a reason, right? I mean, there’s a lot of madness going on right now, at least in the markets related to what’s going on around artificial intelligence in our world. You know, our customers generally have a, you know, they think it’s magic, right? Yep. I’m just being transparent, you know, uh, I’m gonna reduce agents by enabling bots, you know? And then, you know, my very next question is, will, okay, you know, what’s gonna in that’s great. What it’s gonna inform the bots? Is your process written? Now what do you mean?

Scott Kinka (23:09):

Right? And so the whole idea of like providing AI context is really what this boils down to. I mean, it’s only smart insofar as you can tell it things that it can be smart about. Um, so I totally get that. So here we’re in this, you know, what I perceive to be semi irrational, you know, largely art imitate or life imitating art kind of thing around, you know, my gosh, AI’s gonna change the world because we’ve, we’ve got some now available modeling and chat GT, H P G P T and things like that that are, that are going on. You know, obviously you work with AI technologies every day. Um, you know, what are your thoughts on where, where we are as an industry around, you know, good guy, bad guy, government involvement or non-government involvement. Like, you know, it’s kind of getting the, the narrative’s getting crazy. Um, and as you’re answering that, I’d be curious to know if those are some objections that you guys are hearing out there in the field, you know, where people are like, this sounds great, but I’m scared. You know, it was a very broad question, but I think I’m just trying to ask you to open it up a little bit. Like, where are we, uh, in society around this AI topic right now? What’s real, what’s not?

Swapnil Jain (24:23):

Oh, no, no, absolutely. And I think, uh, this, this is the burning topic everywhere, uh, right? I’m glad you brought this up. So I think according to, you know, observe, uh, and me and then my computer science background, I absolutely believe, um, that we are turning a corner with AI right now. Um, right? In the last six months, we have made some of the biggest advancements in ai, um, that, that, that we have made in a while, right? So, you know, in 2017, you know, deep learning was a thing, right? That was the biggest thing that everyone spoke about. Um, right? Mm-hmm. <affirmative>. And in the last, I would say since then, five years, uh, or six years since then, what we are seeing right now with chat, G P D G P T, generative AI are some of the biggest advancements, right? So I, so from a, purely from a technology advancements perspective, you know, this is a big change right now in terms of, you know, how scary it’s gonna take over the world.

Swapnil Jain (25:18):

All of that. I think there is, uh, there is some value in that conversation, uh, right? Because with every technology, right, you gotta use a responsibility, right? So there is, there is, there is value in like where you could take AI and really use it to harm the society. I’ll give you some examples, right? So we are talking about generative ai, right? Generative AI can generate content for in, in a way it looks very, very human, right? So today, for example, I can take your picture, Scott, I can take my picture and I can, I can actually generate a podcast. Oh, actually I do need picture because it’s a podcast. I’ll take your voice and I’ll take my voice and I can generate a podcast and I can, you know, make it about anything I want and I put it out there mm-hmm. <affirmative>, right?

Swapnil Jain (26:04):

And it’ll be a hundred percent like you and I are speaking now, that’s not the right use of, uh, of technology here, right? And technology has reached a point where you actually cannot differentiate, right? So I think there, there is, there is the part of AI where we all have to be really responsible about it, right? Where we all have to be really careful about how we use it to, to the advancements of human society and not to pull us back, right? Yeah. Especially for, especially divide the class into, there’s people who understand these technologies and know how to use this technology to, to sort of dominate the people who do not understand these technologies and don’t know what’s happening in the world, right? So it, it’s, I think it’s on us to make sure, you know, we don’t let that happen and we use AI in the right form.

Swapnil Jain (26:48):

Now, talking about sort of like in context center, right? You spoke about, you know, I’m gonna automate my agents and do bots. I think doing bots and chat bots and conversational AI is not new. Started about five years ago, or even seven. We’ve been talking about chat bots, voice bot agents going away for a very long time, right? We have made progress on that front. I feel like, you know, we obviously see a lot of chat bots, voice bots, but actually believe generative AI and, you know, large language models and what’s happening with chat GPTs is actually gonna advance that journey for sure, right? But it’s gonna do a hundred percent, absolutely not, right? But it will definitely take you one step higher. So if maybe you were automate, you were able to automate 20, 30%, now you can automate 50%, 60%, right? It’s never gonna be a hundred percent.

Swapnil Jain (27:30):

There’s gonna always gonna be an element of human emotions, human empathy, that will play a big, big role in customer experience. Uh, right? It’s, it’s not about, uh, getting a right answer, but it’s also about bringing empathy into, into the mix which, which technology is not gonna do. So all in all, I actually think this is a big moment for ai. Um, you know, and then talking about customers and objections, I actually, I see that when we talk, talked to our customers about the latest and the greatest of technology. So for example, two days ago we hosted a g PT Innovation day where we invited 12 of our largest customers in the Bay Area, spoke about ai, spoke about G P D, and yes, I could see that people are worried, uh, to an extent, which, like, what happens to my data is, is Chad g p D gonna leak my data?

Swapnil Jain (28:12):

Is, is my data gonna be misused? Right? So we as a company have to make sure we protect our customer’s data, just privacy, the security, uh, right? So all in all, to summarize, yes, I think this is a big moment in ai, right? Uh, I encourage, you know, enterprises and CIOs and head of IT to actually look into some of these technologies to advance their contact center. Uh, it’s a natural evolution. It’s gonna happen mm-hmm. <affirmative>, uh, but do in the right way, right? Think about security, think about privacy, think about ethics, all of that into account as you bring some of these technologies into your contact center, into into enterprise.

Scott Kinka (28:46):

This is a hard one. Um, and maybe it’s just a yes or no answer, but I mean, as an industry, have we earned the right to self-police or is, you know, this, or, you know, are the government overreach? Okay? You know, where my opinion is just by me saying that is the government involvement, uh, uh, conversation warranted?

Swapnil Jain (29:11):

You know, I think I’m, I’m kind of split on that, right? The, the software engineer me, you know, the builder in me is like, no, let me build all the cool stuff in the world. Um, right? Yeah. Let me do amazing things with technology, right? But at the same time, I feel like there, there are some things that I could do which are not good for the society, right? There are use cases that I can solve, which are not, not the, the best outcomes, right? So, so I think either as a government or us as, you know, the, the builders, right? We put our guardrails and we put our protocols, right? But there, there is some value in having some guardrails around this that will help us advance this for the right cause versus the wrong causes. Yeah. Be government, right? Or, or maybe government is not the right solution because they might not understand the technology.

Swapnil Jain (29:54):

Maybe it’s us, right? It’s the builders, right? You and me, right? We define the guard rails and we see guys, these are the biggest problems we should solve and we should advance AI to do this, right? Versus, versus going into use cases which are not right. Uh, right. So I, I think of it that way. I think of less as like government and not government, but I think of more of like, Hey, there, there, there, there’s, we can, we can put our right structures around this as builders, uh, to solve for the right use. Interesting.

The Future of AI

Scott Kinka (30:16):

So we’ll be, we’ll be following that closely. Um, you know, let me add one more question on it. I mean, I think the object, I think the hysteria falls into two categories, right? The first one is the very clear and appropriate one you just mentioned, right? The ability to create content that, that is fake. The ability to deep fake images, the ability to, you know, ask a generative model to create, to write you malware, but also to ask it to solve that malware. Like, yes, a hundred percent. The other area of concern is the very kind of sci-fi, you know, approach to the end of humanity <laugh>, right? Thank you to the Matrix and iRobot, and, you know, going all the way back to Isaac Asimov, right? I mean, we’ve been talking about this and it does it, that part of the argument to me does feel very, you know, life imitating art. Um, but it is amazing how that self-fulfilling prophecy does potentially, you know, come to bear. So I’m just gonna ask it. I mean, we want this, we don’t have to dwell too long here, but I mean, you know, if somebody looks at you in the eye and says, yeah, the, all of this is going on, so the singularity is near, what is your answer to that? Like, how would, how, how do you think about that?

Swapnil Jain (31:30):

I would say it’s nearer than it was, uh, two years ago, right? I don’t think it’s, it’s near, we are not there yet, but we, we have made a joint progress.

Scott Kinka (31:39):

Yeah. Got it. And is that something to be, is that something to strive for? Right? I mean, I think the hysteria is based around the singularity being accidental, you know what I mean? That Yeah, yeah, yeah. That the computers wake up, you know what I mean? <laugh>, as opposed to that we inform them the of, of the level of you kind of consciousness that we’re expecting them to have or build towards that. I mean, are we str, are we actually, is it our goal to get the singularity, I guess is the question?

Swapnil Jain (32:06):

I don’t, I I don’t think so. And that’s what I feel like we as technologists, we as we as builders, uh right. Should put our systems in place on what do we want AI to be doing for us? Uh, right. Yeah. It, it should be like AI is there to help us, right? Ai, we, we are the ones creating AI to really help us do our jobs, you know, much better, much faster, be more happy, be, you know, have longer lives, right? It’s for us to do, right? Versus, hey, we mistakenly created the system, which is now not the right thing for us, right? So I think that’s where I feel we, as we as a society, and we as the builders, not the government should come in and put together some structures in place on how, how do we wanna leverage AI going forward?

Scott Kinka (32:43):

Yeah. What was it, what was it in iRobot, it was like the three laws, right? Like yeah, we’re gonna have to have those at some point, which I totally get. Yeah, exactly. Um, which makes complete sense. Well, that was that one in a, in a direction. I wasn’t entirely ready for it to go in, but I loved every second of it, and I’m, we could just keep going on this. I’ll definitely, we’ll look for a time to have you on again, to get a little bit deeper on this model as the industry continues to move. Yeah. Um, let me just end with a couple quick questions while I have you, and we may have already answered the first one, but I always love asking this kind of give us the, give us a shameless prediction, 18 to 24 months, um, doesn’t necessarily need to be in tech, but I feel like given the conversation we’ve had so far, that maybe that’s where it’s gonna land. Um, what do you, you know, look in your crystal ball, 18 to 24 months,

Swapnil Jain (33:34):

I think there’s gonna be extremely large companies that will be created, uh, based on generative ai, and there will be a lot of companies that will be under fear and threat using this technology. Right? Uh, that’s my big prediction, I think, and, and the biggest example that we are seeing in this direction is Google and Chat g pt, right? Yeah. No one, no one could have ever imagined in their lifetime that anything can threaten Google’s business, right? But it’s a very common topic that we all talk about now. Yeah. Which is, right. So that’s, that’s my biggest prediction that we are gonna see, uh, a completely new set of companies which are powered by these technologies have taken, you know, are, are very large now, and then some of the largest companies out there under threat, uh, who have not advanced themselves to use these technologies.

Scott Kinka (34:29):

That’s amazing. Um, two more. And I agree, by the way, I think that’s, that, that’s awesome. Um, let’s just assume that eventual robot driven, you know, extinction level event happens <laugh>, right? And, and there’s, there’s one app left on your phone, you’re a have to have it one app remaining functioning. What’s it gonna be?

Swapnil Jain (34:51):

You know, I, I, I would, it’s, it’s not even an app, it’s just the basic use of phone. I would like an ability to talk to my family, <laugh>, right? Yeah. My wife, you know, when I have kids, my parents, you know, just gimme that, right? I think that is what we live for, right? At the end of the day, right? Scott, you and I, what do we live for our families, our kids, those moments, right? Our parents. So I think forget technology, forget business. I, I, I want access to that in the end.

Scott Kinka (35:16):

Okay. Got it. I love it. Is there something in, we should be reading? That’s the last one. What’s on your end table right now is that you’ve got some kind of recommended reading for the listeners?

Swapnil Jain (35:29):

I think one of the things that, that has been on my mind, uh, with, with so much advancements in technology is like, just continue to build my thinking ability. Uh, where like, how I think as a leader, how I think as a ceo, how I think as a founder. So I’m, I’m reading this book called Art, art of Thinking Clearly, um, right? Uh, so it’s all about, it’s less about ai, it’s less about software, it’s less about business, it’s more about like, how do I continue to evolve my thinking muscle and process so I can process everything that’s happening in the world. Uh, it could be business, it could be ai, it could be people, it could be anything, right? So just continue to evolve me as a human, is what I’m focusing on right now.

Scott Kinka (36:07):

That’s a, that’s a whole other podcast episode in and of itself. So while I’ll, I’ll wait, I’ll wait for you to get through that book and we’ll have another conversation. Um, but words of Wisdom, maybe get a link to that in the show notes. I think lastly, you know, if anybody wants to find out a little bit more about what you guys are up to, uh, how best to do that.

Swapnil Jain (36:26):

Absolutely. No, thank you so much. So I would love to chat if anyone wants to learn more about Observe and what we do and in general, wanna talk about ai. Uh, you can reach out to me at Swapnil, that’s my first name, S W A P N I L at observe obse r, or just go to our website, www dot observe ai. Uh, and you can leave your phone number and name and, and we’ll reach out to you.

Scott Kinka (36:48):

Yeah, fantastic. And of course, you know, observe AI is a, is a Bridgepoint Technologies partner. Um, they work very closely with our CX teams, so we build strategies around all of this in terms of implementing, not just observe ai, but the CX platforms and the, the c a partners that you guys have, um, in businesses. So we’d be happy to, to help in that process and, and move it along. This has been a great conversation, this webinar. Thank you so much for your time today and for our listeners, thank you for your time. Um, super excited to have spent this time together and I hope you guys found that as insightful as I did. We will catch you on an upcoming episode of the Bridge.

Swapnil Jain (37:24):

Thank you so much, Carl. It was great, uh, speaking with you. I’m glad we could cover a bunch of topics and you from contact center to ai, uh, and, and as you said, hopefully was, uh, insightful for, for our listeners.

Scott Kinka (37:34):

Thank you. Fantastic. Thank you so much.



Experience the Bridgepointe Way

Start today with a no-obligation consultation with one of our experts.

Table of Contents

Related Blog Posts