← Cybersecurity Today
Cybersecurity Today
Bridging the Gap: AI and Cybersecurity in the Enterprise
00:50:50
In this episode of Cybersecurity Today, host Jim Love is joined by Krish Banerjee, the Canada Managing Director at Accenture for AI and Data. They begin the discussion with a report from Accenture that highlights the gap between the perceived and...
Loading summary
Transcript71 lines
- [00:01]
Jim Love
Welcome to cybersecurity. Today on the weekend, as I promised on the Friday show, Accenture came out with a report that I think really showed the stark difference between the perception and reality of our preparedness to address security issues we are definitely going to face as AI becomes more and more prevalent in the enterprise. I've shared my opinion that based on history, the pressures to implement a new technology that promises to have such a radical impact on productivity and costs and perhaps even competitive benefits, that pressure will push us forward even if we're not as ready as we would like to be. Call me a cynic or perhaps a realist and I'd love to be proven wrong, but I think that even as a supporter of AI in the enterprise, I'm still concerned that there are significant cybersecurity issues that we are not as prepared for as we like to think. So, as I noted, I think we'll have some interesting discussions ahead of us to prepare us for those. I think it's important that we understand AI as it's seen by the business. So I'll be adding some weekend shows that may not focus on cybersecurity, but may aim to provide some understanding of AI from a business point of view. And for our American listeners, I'll give you a heads up that this conversation is rooted in Canada, and I think that's also something to be aware of. Issues like digital sovereignty from countries outside the US and other issues are going to be things that you may encounter. But many of the things we discuss are universal in nature. And Krish, my interviewee, is a really interesting and engaging person, so I hope this interview is interesting. But once again, I'd love your feedback. We have a new website@technewsday.com and you can send us your comments there via the Contact Us form. Or if you're watching on YouTube, you can leave a comment under the video. And now, here's our show. My guest today is Krish Banerjee. He's the Canada Managing Director at Accenture for AI and data, leading a team of over 500 professionals who deliver solutions for clients across industries across the country. Krish was last on the podcast in March 2024. That was amazing for two reasons. One, there's still a podcast. I forgot to retire, but I didn't even have to remember that. I went, when was Chris here? And I went, I can ask AI that. So I asked Perplexity. And it I've outsourced my memory. March 2024 went, wow, that in AI years, that's decades so Krish, welcome.
- [02:46]
Krish Banerjee
Kim, thank you and always a pleasure to be with you and your podcast and thoroughly enjoyed our last conversation. Looking forward to this one.
- [02:55]
Jim Love
Yeah. So what's been happening recently since we last talked? Nothing much. Right. Not very exciting. There's nothing going on, have you think.
- [03:03]
Krish Banerjee
In AI times or AI years, it's probably, probably just a few blips, if you will, in terms of at micro kind of units of time that you are measuring. But the change has been significant. Obviously the market has shifted quite a bit in terms of how we have seen new opportunities exist and emerge. I'm not going to share anything new. For people who knows about Agent Dig now, which we were not talking as much when we met last time. But no, there is no conversation that can happen without talking about Agent Digg. So we will obviously talk about that. But I am also seeing an uptick in terms of how our Canadian organizations and community is actually moving towards adopt. There's quite a bit of work still needs to happen, but there is an intent is very inspiring.
- [03:56]
Jim Love
Yeah, and I was being facetious. A year is like a decade in AI right now. There's been so much happening. But the one place where I think you were prescient was you've always kept data and AI together in, in our conversation last time. I think you, you really considered that and I think that's somebody woke up, the rest of the world woke up to that at one point said wait a minute, this is important.
- [04:20]
Krish Banerjee
Exactly, exactly. And I think the most interesting quote that I heard on that topic is that we have everything ready for AI except for the data and everything cost is ready to actually launch into AI and get their hands on AI on the platforms, all the agents and all the human beings except for data. So I think we are turning towards data now, give its rightful share of attention. And I can see there's already quite a bit of work happening in the space of whether it's privacy or whether it's integration, whether we are actually building some data products which are specific to industries. I can see that the next few decades, which will be a few months in this case, will be spent on focusing on data.
- [05:09]
Jim Love
Yeah, I was watching it because we started to talk about access to data and I come out of my, I'm an old guy, I'm from the old data warehouse days when remember when somebody sent you a query or something they wanted done, I had a three month project to reorganize the data and pull it all in and do all that sort of stuff. Now we're talking about conversations with data warehouses and we may not totally be there yet, but we're. This is going to, this is going to be something that comes. That's got to be fairly exciting from your whole background.
- [05:44]
Krish Banerjee
It is very exciting. And we at Accenture have always been focused on data and making sure that the fundamentals of it are, are secure. And then everything on top of it, whether to your point, at some point it was reporting and business intelligence, which was an extremely important aspect of what we can build on data. From there we went on to analytics and using some of the models, algorithms and we can drive value out of it and then from that now to AI. But we have always been true to the fundamentals and the core of it. And that is coming back again. Like we have seen the shift of how things move towards more often shiny object at the time, which has changed over different eras of technology advancements, whether it's reporting, whether it's analytics, whether it's now AI. But the core fundamentals of what you need to build true insights out of remains the same, which is in this case is data.
- [06:42]
Jim Love
Yeah, but your practice covers most of Canada's large enterprises. So you're, you're seeing this from the enterprise version of AI, which is an exciting piece. My feeling is that we're finally grappling with that. And that is, I think people were dipping their toe in the water. They were doing things. But we're now starting to confront the, I wouldn't want to call them the issues, but the things that you have to put in place to have enterprise AI. Do you have the same sense?
- [07:13]
Krish Banerjee
Yeah, I think enterprise AI is an interesting topic. Right. And we have few examples of that. I'd say in Canada we are probably where there is an opportunity. There is a lot of pockets of AI. There's a lot of like functional AI that is happening in very specific areas, whether it's hr, whether it's finance, whether it's in areas of performance management or whether it's in areas of supply chain. But true enterprise AI has, has been the area of interest for us, which is where you can drive value, which is where you can actually drive reinvention. There's quite a bit of work happening globally on that. However, I do believe that in Canada we are a little behind when it comes to driving true productivity and value and generating value out of a true enterprise AI. So not sure if I answered your question, Jim, but that's how I'm looking at it. But double click on that. What exactly you know about it No.
- [08:11]
Jim Love
I think that's, I think that's fair. I think your perspective is that we've been building sort of point solutions and now bringing that together is the next task. I think I totally agree with you. I think the world is further ahead of us in many cases and we need to fix that. I think that's an important thing to fix. But it, but I'm starting to see the discussions take on a new maturity when you start to talk about what are we going to do about data, what are we going to do about security. And those are the points that need the foundation points, because we've been. I don't think that's unfair. I think we always do this in technology. We're led by the bright shiny object and we're led by the promise of something new. We always say, you should build the data and the security in as you're going. But we never do that. We always get far advanced, say we better go back and get control of this. And I think that's the place I see us, I see not just Canada, but across the entire world. We start to see people grappling with, saying, this is serious, it's enterprise level. We need to figure out how we're going to run it now. It can't be just something the employees bring in.
- [09:18]
Krish Banerjee
Yeah. And I think we have spent the last two plus years proving this is real. This is what I try to explain to a number of my clients and friends here is that I don't think AI needs to prove itself anymore. Like it has done its job of proving that it worked and it's real. It's really for us to make sure that we are engaging that capability and power to the right use of whether it's country or community or a specific organization. That's where our job is. Like AI has done its job in the past couple of years.
- [09:53]
Jim Love
Yeah. And I think in particular in the past three to six months, I think we've hit some levels where we can say we no longer talk about. If we're talking about when, then I think that's a, that's an important distinction we make. But also I think, and I've, and I've been happy to not have the conversation. Every conversation I had about AI started with when are we going to get to AGI? When are we going to get to. When is it going to be sentient? I'm going, you know something, guys? You have enough tools right now in the toolkit to do astonishing things like AGI. Nice conversation. Let's have a Beer and talk about that on the dock. But let's get to work with the tools we have. Do you get the same sense that we're not really using everything that we've got available for us from us?
- [10:40]
Krish Banerjee
Yeah, for sure. I think back to the conversation we were just having that there are definitely very pockets of AI and brilliance that are happening, but there's also a lot of opportunities that are like white space for us where if we stitch the things that are happening, whether you connect the dots between opportunities that exist across your enterprise, there is an opportunity to multiplex the value that you are already unlocking. So a hundred percent, there is things that are right now available that can unlock 2x3x of opportunity value savings for any organization, whether it's public, private, federal, commercial, whichever flavor of it. And then AGI is an interesting concept, but AGI to me is also relative. If you roll back 10 years, what we are doing today could have been considered as AGI. So the boundary of AGI is also what we keep shifting. Right. And I think that's like almost chasing a target that keeps changing on us. So let's just focus on the current and think about the possibilities that we will work towards but not chase something that almost like a mirage or chasing in the dessert.
- [11:57]
Jim Love
Yeah, it's like we keep moving the goalposts. I don't know how I know you're not as old as me, but there was a show in the 1980s, I think, where this guy had a talking car and it would drive itself and it had a great voice and would chat with you and give you all this information. We couldn't do that now. And you know, and that's. So we are living in our fantasies from 20 years ago, but we're sitting there going, is that all you can do? Make a talking car?
- [12:28]
Krish Banerjee
Exactly. I grew up watching animation at that time called Johnny Soko and His Flying Robot. It used to have the watch where you can talk to. And that was one of the most fascinating thing at that time. I mean, that's Apple Watch or any other smartwatch at this point. I think that's an example of how AGI is in target that we set up for ourselves to achieve something. And then we say, okay, now I'm moving the boundary another 20 years further because then you can actually have more brilliance in terms of human thinking and innovation.
- [13:05]
Jim Love
Which may explain this story we did a couple days ago and we picked it up. It was, I think it was a TD bank did a survey and said that Canadians gave themselves a C at their understanding and use of AI. And I found that astonishing because even today I don't think people realize how much they're doing with AI and don't realize it.
- [13:30]
Krish Banerjee
Yeah, there's already a lot of AI even in our lives. Whether we realize it, whether we acknowledge it, there's already a lot of AI in our lives. And I can only see that we'll become increasingly more embedded and more blended into our lives. But there is an aspect of being deliberate and I think that's the part that we need more tension. Like how can I be more deliberate about unlocking things that are not just happening to me? Whether it's in my world, whether it's in how, how I'm looking for information, how am I searching information, how am I working on my next vacation planning? I think we will be working on things like Airbnb and Expedia and others that are building AI, which is somebody else is doing AI to me versus I'm going out and seeking the power and use of AI for my own productivity. I think that's, to me is the kind of the bottom up difference that we need to build as a community, as a society. If we all can be deliberate about what, everything, what we can do in our lives that can change. Even a 2%, 5% productivity improvements that will build up and snowball effect of what it could mean for the, for the country, for the community in a whole. Because if you look at our situation, like bank of Canada said, we are in a productivity emergency number 18 in OECD, which is not a great place to be.
- [15:04]
Jim Love
No. And that productivity emergency will not change unless we become better at investing in and using technology. Jim Balsilli, I've had him on the program, is passionate about this. We need to become a hub and we'll talk about Canada in a second. I do want to get there. I just want to go back to this idea of you're dealing with enterprises and with their customers. I, this is maybe just my perspective and I'm close to 70, man. I'm not supposed to be on the bleeding edge of this stuff. I am so impatient with companies that give me the old interfaces and already I want nothing to do with it. Maybe that's becoming a cranky old guy. But if I can't just type in, tell me how I do this and it comes back with the answer, I'm not, I just move on. And so I think, if, I think we may think we're not as advanced in AI, but I suspect that we're Impatiently waiting for what it can actually do.
- [16:05]
Krish Banerjee
I agree. And I was watching my 13 year old daughter the other day trying to find some information and I think the default now is to go to publicly available Copilot, Gemini or any of those platforms instead of going to the actual search. So if you go by that shift there itself, they're growing up as an AI native generation.
- [16:32]
Jim Love
Yeah. And I think that's, that's something that customers will start to demand. I'm sorry, but I'm immensely critical of government interfaces for the most part because I don't think they really understand that people who come to deal with government know nothing about government. We don't know the policies, we don't know the procedures. We just want an answer. And so many of the interfaces are set up to deal with. And we had this in corporate life a while ago. I think it's changed a lot. A lot of interfaces were set up so that the corporation could manage the transaction with, in other words, according to their rules. Consumers rebelled against that and I think most good or competitive corporations have realized that and have changed their attitude. I don't think government has quite figured that out yet. That when citizens come, they don't. They're not interested in dealing with the rules, they're interested in getting an answer. And I think that's a challenge that AI allows us to address. I hope.
- [17:30]
Krish Banerjee
I think there's a lot of positive intent and momentum and conversations happening if we collect everything around us from what we just announced at G7 to what. What we can expect out of the appointment of our first AI minister. There's a lot of good intent. I would like to take the optimism and the positivity out of that. And obviously there's a lot to be proven and a lot to be opportunity to be walking. The talk needs to happen. But there's some good signs of the first couple of steps of the talk that needs to be walked.
- [18:10]
Jim Love
I'm very optimistic. I just hope we don't because. And I think things have changed and I don't think, I don't think it's a political statement. I think if you looked at any, if you talk to any of the leaders in there, people have woken up and said we have to be better at this. And the productivity crisis may be the driver, maybe the trade issues that we have. But people are starting to realize we need to get a handle of this. And I think the appointment of an Evan Solomon, interesting appointment. Do you remember Evan Solomon from when he was a broadcaster? He's.
- [18:44]
Krish Banerjee
I do. I do. And it's definitely interesting and I think it shows that AI is a topic that is pretty everywhere and it doesn't get limited to people who might be academic or might have a policy background or might have a very specific background. But it's an interesting kind of message that I think I'm taking away from that.
- [19:05]
Jim Love
Yeah, yeah. And don't get me wrong, you can be in. I am, I guess I am in the media. There's some very intelligent people in the media who.
- [19:13]
Krish Banerjee
It's your next grip. Maybe.
- [19:16]
Jim Love
Maybe. Yeah. Yeah. So talk to me about this. What do you have you heard heard much of. I've been digging into. I was surprised. There was already a great deal of activity going on in the federal government in terms of AI that that was kept hidden, I think. Or maybe I just didn't notice it.
- [19:36]
Krish Banerjee
I think there is a lot of interest right now in the few specific areas, obviously the areas of service. And I know that there's conversations around how citizen service could be improved using AI and using more front office capabilities of AI. There's also conversations around how could we use AI for more productive collaboration in business. How could we reduce all the red tape that we might have make business easier for us to do, whether it's for individuals or for business to business relationship. And there's also a conversation that is publicly known now about what it means from a sovereign perspective. And there's a lot of focus from the federal government which is I think is also another good step in terms of what do. How should Canada think about AI as a federal. As a sovereign, sorry, a sovereign topic. It doesn't need to be sovereign by, by physical kind of definition. Like obviously there's a physical definition of sovereign. But I think there is an. To me, there is a more broader probably in philosophical thinking about sovereign. It is a Canadian thinking, Canadian priority, Canadian investment and sovereign investment and then everything else works together in order to bring that to life.
- [21:03]
Jim Love
Yeah. And I don't think we're the only ones. I think the world's woken up to the fact that we sat back and watched the Americans develop AI and now the Chinese have it in a big way. And, and I think a lot of countries are starting to say, wait a minute, how does this fit with our operations, our sovereignty? Can we sit on the sidelines and just wait for America to develop everything for us? What should we be doing? I know Europe is grappling with this. I presume the same thing is happening in India and definitely China. Canada's another player in there. It's almost like the world woke up and said we can't be a spectator anymore.
- [21:42]
Krish Banerjee
Yeah. And you're right. Like there, there's different degrees of interest and speed and investment from different jurisdictions. So some others are probably further along. And it's also true that everyone has a slightly different angle. Some are taking more of an compliance and regulation angle versus some are taking more of open and free and innovation angle. Some would be taking this more from an. Let's use this as our world dominance angle. So we know those flavors are. Without being more specific. But I think the opportunity for me, for Canada is to think sovereign for Canadian purpose and objectives. Right. What do we need to do to help our Canadian challenges that are in front of us from innovation, from everything that is happening in the whole geopolitical situation. And how can we use that to our advantage and use that to come out at the other end of this stronger and more resilient?
- [22:47]
Jim Love
Yeah. And I think it's forced us to actually have a discussion about what digital sovereignty means. We're actually going to do a whole program on this because it's. We've thought of sovereignty as a physical thing and even with the whole digital transformation of our economies, we never actually sat down and said, wait a minute, sovereignty is a digital thing too. And we haven't, we don't, we haven't thought through what we even mean by that.
- [23:14]
Krish Banerjee
Yeah, no, I think it's a great point. That's what I was referring to that the default direction of sovereignty is physical. Let's have a Canadian data center. Let's make sure that the data stays in the country. The network is Canadian and how we can connect Canadian network on. However, to your point, like there is a whole digital angle to that and how we can use a digital and the national and the nation building side of it, like how we can use sovereign to build the nation. And to come out to my point of all of these uncertainties that are going around, come out on the other end of this stronger. And I think there's a great opportunity. We've seen that during COVID There are countries, there are organizations who came out stronger at the other end of code. And I almost feel like there's a. There's a similar kind of a challenge. May not be as prominent at this point, but definitely an opportunity ahead of us.
- [24:10]
Jim Love
Yeah, I honestly think it is prominent at this point. I think people have started to realize and part of it was people banging on the drum about productivity and productivity can be a sterile thing. And I've. People have talked about this output GDP per capita and why does that really matter? Why does it really matter? Because if you don't create wealth you've got nothing to distribute and wealth creation is going to change. I, I maybe a whole another program on this but we will not recognize our economy in 10 years and that's a short period of time for an economy to we talk about mining and oil and gas and the industries that Canada has been classically good at. These are 10 and 20 year projects. People talk about a pipeline. A pipeline, if you get it approved in two years is still a 10 year project. Heck of having a new subway line is a 10 year project but we're going to have to move at a faster speed than that because technology won't slow down for us. So it's I think we have gotten to that realization.
- [25:16]
Krish Banerjee
I hope I'd like to believe so and I'm seeing some science of it and I'm very positive and optimistic about what I'm seeing. But I also believe that there is a lot more that we could do and to the conversations we were having in the earlier part of this chat that we have the tools now, we have the opportunity now with proof of concepts and things that we have done over the past two years. We just need to now take a step to say I'm going to now try and do this for broader than just building a point solution. Cause let me reinvent the entire value chain of my industry. Let me think about how I can sell consumer packaged goods to my customers very differently than what I have done in the last 50 years. I think that is a binary point in my opinion that where we are.
- [26:12]
Jim Love
In front of and what let me ask you just let me just start with one so we could have an example. What's the project you'd most like to do? If I came in and said Krish I got a budget, don't worry about that. What's the project you'd most like to do?
- [26:28]
Krish Banerjee
It's a great question and I'd say there's probably two, two parts of the answer. One is there's a passion side of the answer which is I would like to do something which is more tangible from an community perspective. Like at Accenture we do quite a bit of pro bono work. We have done some of the work with universities and mental health organizations and others and something around that would be of interest. So if we can use the power that we have been given with AI to the cause of something better like AI for good would be an area of interest and something around that. And there could be many flavors of it, whether it's mental health, whether it's human lives, whether it's transport and traffic, which is very predominant problem for Toronto and my city. Anything we can do there would be of interest. The other answer would be like something around the whole. By Canada and for Canada from a sovereign perspective. Like if I can be part of something that will build Canada stronger and build us as the AI destination which we should be given everything that we have from our lineage and history and bring back kind of our legitimate and rightful supremacy in AI. I think that will be the other project.
- [27:46]
Jim Love
Do you know the two projects I want to do? Yeah, one, and I won't get to play in this because I don't think I'm. I don't think I'm really. I don't think I'm really a full time employee ever again. For the. I'm keep trying to retire, but the. But I'm just excited about. I'd love to take AI and reinvent health care starting with outcomes, not processes, and say what are the outcomes we want to bring about and how could we use AI for them? My example is I want to follow my doctor around in his office all day and I want to take all of the administration stuff that he's got to do and make that disappear so that all he's got to do is talk to patients and that is doable. That would excite the heck out of me.
- [28:35]
Krish Banerjee
I think that's a great one. In my first bucket of something that will touch human lives like this.
- [28:41]
Jim Love
Yeah, I think we agree on that one. The other one I'd love to do and this one I may actually participate in because we're. We started our own institute, believe it or not, to do this commercialization of AI. I just want to see our efforts go towards commercializing AI. Not necessarily even to own it, but just to say let's push these things that will make productivity so much greater for Canadian businesses. And I used to teach at Waterloo and I was having a discussion with another Prof. Friend of mine who's also retiring and look at it. I'd love to see that happen. I just love to see the mechanism of taking that wonderful enthusiasm and energy and knowledge that we have and it's. There's tons of it and driving that to a small business that can now be much more profitable. Those are my two.
- [29:36]
Krish Banerjee
No, I think that's great. And the second one you mentioned is really very interesting. It's about how we are thinking about Reinvention, like in our Accenture world and how we think about helping our clients reinvent, whether it's in the mid market or enterprise level, the G 2000s, there is an opportunity for everyone to think about the reinvention. And I think what you're saying is how do I take AI and create a path of value opportunity for organizations to reinvent and build that commercial kind of pathway for them.
- [30:13]
Jim Love
Yeah, yeah. And when that spark starts, it's exciting. I remember in the early days of it, that's how we felt. We thought we were magicians. We were building, we were building these systems, barely hold them together with bailing wire, but we were doing incredible things.
- [30:30]
Krish Banerjee
Yeah, well, what I think we, we talked about in the last one, which is about with great power comes great responsibilities, right?
- [30:38]
Jim Love
Yes, absolutely.
- [30:39]
Krish Banerjee
In the last session as well. Can't lose sight of that. Can't lose sight of how we need to think about building AI with the right responsible mindset and thinking and framework. And we call it responsible AI, Ethical AI. There's all different names to it. But essentially to make sure that we are doing it with the right intent.
- [31:00]
Jim Love
Yeah, yeah. And the right tools. Speaking of responsibility, the. We're starting to have the conversation about the economy and where and the changes in the economy. I was interviewing someone today and we were talking about the almost. It's not a total disappearance, but entry level jobs are disappearing faster than, than kids can there. There are far fewer enter entry level jobs than there ever were. And that's not just economics. That's something that we're starting to see the impact. I don't think we fully, I don't think we fully imagine the impact of AI. We know, I know ultimately at one point or another, most of what we can do can be automated. That's. It's just a question of how many years, decades, weeks, months, whatever it is. But the change in the economy, how do you see that? What do you, what are you looking forward at? And in terms of what.
- [31:56]
Krish Banerjee
I think there will be few things if I have to hazard prediction or a guess in terms of next maybe two to three years, because anything beyond that is, is almost unknown at this point. So I wouldn't go beyond three years. The first one is we will certainly see a point where we will have to make a decision as an individual, as an organization, which I was referring to as that binary decision point where we have choices in front of us, which is am I going to be an AI leader or am I going to be still following? And we have seen from, from the past. Similar kind of. It pivots or other moments that there is a 2 to 3x difference between those who will make the choice of leading versus following. So I think we are there like that binary kind of big bang point is already there. So the second one will be about differentiation. At some point, AI will become commoditized. Like we all will have some agents, we will all have some kind of an AI capability. It might be six months, it might be 12 months, 18 months for some. But there will be a point where you will draw the line and say it's like digital. Having an app on the cell phone is no longer differentiation. Used to be in 2013, 2014. And somebody will say this pizza company has an app. The other one doesn't have. It's only on website. And I will go and use it, the one that is an app, but it's no longer a differentiation. So what is going to differentiate? What is going to be your face in the future of your organization? How are you going to look differently than others is going to be important. And the third things I'll wrap up with what I believe is we will see more of an interlock between the physical world and the soft world or the AI world. Like whether it's in robotics, whether it's in how we are looking at manufacturing. We are already seeing humanoids. We in Accenture, we made investment in humanoid research and firms that are building those. I think that's a little bit on the out yet. But if you look two to three years, it's not going to be too far. When we actually see robots in our houses doing stuff like we already have Roomba and other stuff doing things. It has quite a bit of AI built in it. Having a humanoid do other stuff is probably not too far away.
- [34:28]
Jim Love
Yeah. I'm more optimistic than you are. I think we're going to see commercial robots in 18 months that will be at the high end of purchase. And most of them are estimating coming in at 30 grand or so us. And I have to say, not. Not immediately. When you look at the future world I'm looking, I live in a place in the country difficult to maintain. I'm 70. How many years more can I actually look at doing this? I'm thinking long enough to get a robot man.
- [34:58]
Krish Banerjee
Yeah. And then find it be like when does the price comes from 30 to 20 to 10 to 5, so that it becomes more of a commodity. And the necessity that everyone will feel like they will have and are restricted to only those who can Afford. And then at point, at that point it'll be how do you differentiate? Right. Everyone has some kind of a mechanism to clean their house whether it's an AI or not. But the differentiation will be how intelligent it will be going forward.
- [35:27]
Jim Love
Yeah. And robots being the outside of our two to three year prediction, the I was remiss in not allowing you to get give more of an explanation of agentic AI because I think that is what's going to power the next two to three years is AI that can take action. Have I got that correct in your definition?
- [35:47]
Krish Banerjee
For sure. And it's great that we talked about AI and this topic for 45 minutes without coming to agentic, which is good in my opinion because to some extent I am thinking that we are over indexing on agentic and I'll explain why. And I think there's a ton of value in thinking agent. We have the tools, we can build agents that can make decisions. You provide them with a purpose and a goal and it solves towards that goal. Instead of LLMs which is prom like you ask a question, it gives you an answer in an agent you give them a goal. It goes and does it's all kind of things. Even if it talks to agent does, orchestration comes back. However, if it's only as successful and as relevant as the process that you are trying to reinvent, whether it's simple in your life, whether it's what you do in your day to day, whether your vacation plan or planning for your weekend to and supply chain company who's trying to think about their source to pay process and how are you using agent tick to actually get the value. And that's what I am. I am more interested in having the conversation because we have an natural kind of opportunity to use some of these shiny hammers if you will and look for the nails. And there's a bit of that happens whenever there is in something new. And I think there's something like that happening with agent tech which is great. It's getting all the attention and it should get all the attention. But let's use the opportunity to think in my insurance organization, what is that we can drive in claims? What can I do in underwriting if I'm a bank, what can I do with KYC and fraud and aml or if I'm in retail organization, what does it mean for my inventory optimization? What is my full process flow of that that I can then start picking and dissecting and say where does an agent play? And I'm going to get 20, 30% of an improvement in terms of productivity. That to me is more interesting and to me is the topic of the next six months, 12 months when it comes to agentic. Agent tick for sure, but agent tick for what and to do what and how is where I would like to lead us to.
- [38:10]
Jim Love
Yeah, and I think the interesting piece of that I think, I hope we'll get to is to not look at it as something that has to automate everything. It. I think we need to collaborate with our AI and realize that many of the things that we think of as barriers to agentic are just failures to us, of us thinking we need to be involved in the process, we need to collaborate. And I think that's a more mature way to look at it because AI always gets judged way worse than humans or by bigger standards. And you look at this, you say, I'm going to look through all these lists and banking's a great example. We look through all these lists and I'm going to find fraud or indicated fraud. You're going to get false positives, I'm going to get false positives if I put a human on it. And as a matter of fact, if I put a human on it, they're not going to be as fast and they're probably going to fall asleep during this. So we, we accept a certain amount of false positives or errors. We, we can't seem to get to that with AI in like we all, we always raise the bar so high that you could never do it. I booked a flight the other day. The system double booked me. It wasn't AI. I dealt with it. And I think we, we have to get to that some somehow to that idea that we're not, we're collaborating with something that is, that is new.
- [39:35]
Krish Banerjee
I think it's a great point. And the way I think about it is that simple example. Let's say I'm going from point A to point B and that's my goal. I want to travel from point A to point B. And the only way I can go there is by using different modes of transportation. I need to take a car, I need to fly some distance, maybe walk to a ferry and take a ferry and cross the river. Now if at the beginning of the travel I give you the condition that you can only use one mode of transportation, which is fly, then how would your travel look like? You will try to use an airborne mechanism of travel every single step, which you could have either walked or you could have taken a car, or you could have taken a Train or a ferry, that's AI. Like you are trying to use AI for everything rather than to your example, there's pure automation, there's probably human needed in some parts of it. That's why I always say take the value angle, lead with the value, look at the process you are trying to reinvent. Maybe 40% of this is AI and agent and everything else is probably just pure human efficiency.
- [40:50]
Jim Love
Yeah, it's interesting when you think from the outcomes. Google has now released its new Google Maps in Europe and it will now do exactly what you said. And are you going to take a car? Would you like to bike? Would you like to walk? And offer you those choices? And I thought that's why I always keep saying there's more AI in our lives than we think. And it, it's already there. Can we talk about a little bit about just the democratization of AI and open source AI has become a big thing? Certainly. One of the studies I saw we did a story on it was from the Linux foundation said 89% of organizations are using open source AI in one form or another. What's your position? What are your thoughts on that?
- [41:37]
Krish Banerjee
I think there, there are a couple of things there. There's value in how we use OpenAI or rather AI which is open is. Sometimes it becomes, yeah, we have to.
- [41:50]
Jim Love
Be careful on that.
- [41:51]
Krish Banerjee
I am careful about that. So let's say AI which is more open for use. There's definite advantage to using that. If you know what you want to do. If you want to use that to actually go under the hood and make adjustments that will serve your purpose, then sure, go for it. But do the same conversation we were just having that not every single nail needs the same hammer or the same problem needs the same solution. You don't need AI which is open for everything. You are completely okay with things that has been built for you and you can use now. If you want to go under the hood and adjust some of the parameters for a specific use case, then obviously you need like you need that. But try to understand where you need and why you need it. I often find that people just without understanding the real need of it says oh, this AI is not open. But the question is why? Like why does it need to be open? It has given you the APIs, it has given you the opportunity to go and make those adjustments. But certainly for the right type of use cases it is a value that you can have, but you need to be able to use that power. Because if you are given an AI which is open but all you can, all you are using that for is asking questions and it gives you an answer. You might as well do that with a black box in front of you.
- [43:22]
Jim Love
Yeah, yeah. Just want to turn it around and because we're coming to the end of our time here and goes by so quickly when you're talking about this stuff, two things I want to ask you. One is, is a lot of people who watch the show or listen to the show are executives and they're trying to figure out where they go with AI And I think one of the things I always want to know is what are the questions that executives should be asking?
- [43:52]
Krish Banerjee
Three things I think a lot of that we discussed on this call already. The first and foremost to me is what value am I trying to drive? What is the problem I'm trying to solve, which I think is. Hasn't changed in my 25 years in, in this space and consulting, right? What problem am I trying to solve and what value am I going to drive? So that's question number one. The number two would be do I have the right tools in my toolkit? Whether it's the right digital core, whether it's the right data platform within the right AI tools, you need to understand, do I have the right things? Because if you are going to do a renovation in your backyard, you need to know whether you have the right tools or your kitchen. Do I have the right tools that I need for the next few years that will get me to the end point. And the third, and probably one of the most important ones is about talent. Where are my people? Are they ready? How are they going to accept what training, what kind of skill, upliftment do they need? Because the first two can be true. And I've seen so many that has not gone to the extent of delivering what they have to or should be because of the third point, which is about talent and change and how my organization is going to react. How am I going to change the ways of working? Because I can put in great agentic solution for my supply chain. How am I going to make sure that my organization is going to adopt this new ways of working? So those are the three things problem I'm trying to solve. Make sure that I know exactly what my North Star is. Where am I driving, leading with value, understanding the toolkit that you have and I have the right tools or not, and then making sure that you have the people angle covered and you have the people who can lead with this change.
- [45:48]
Jim Love
And as a coach, I'm. I don't know how you work, but I suspect you're a pretty good one. I'm feeling behind, feeling a little overwhelmed. What can I do?
- [45:58]
Krish Banerjee
I always say, and I told this to my, my daughter who's 13, and also to my father who's old enough to be worried about you getting his hands on AI and say, just get started. Get your hands on something like play around with within simple interface. Make some small incremental changes. Like the way Jim, you mentioned that you're starting to use some aspects of AI. Move from search to synthesizing. Use AI to build some of the things that we do in our lives, whether it's a vacation planning, whether it's planning your weekend getaway or how you're going to plan your daughter's birthday. These are simple things that are available. And then there are further advanced things that we can always get people exposed to, depending on their interest and skill level and all. Taking that first step of adoption is important. So get started. There's not much you can break by getting started for yourselves and for AI.
- [47:02]
Jim Love
It's big enough to take the abuse you're going to give it. Yeah. And as a. And I get this a lot. This is, I think we all talk about it as a family person and you've got your children growing up. I've heard all kinds of advice from teach them to code to what do what's your. How are you trying to help your kids move forward into this new world?
- [47:24]
Krish Banerjee
Yeah. So I have two daughters. One just graduated from university, other one going to high school. They both have degrees of interest in this space and trying to get their hands on some kind of an AI education. My, my older one took cognitive science as an discipline, so she's already exposed to that. And the younger one is getting Python lessons in summer. So that's one way to get them interested. I couldn't have imagined learning Python when I was 13 or going to grade nine. But that's what she's going to do for grade good part of her summer and see how she likes it and build some app or build some kind of a decision engine. I think it's important to build stuff like how we used to learn by playing with Legos, like our kind of thinking about structures and building and all the mechanics of how things connect with each other. That grew when I was young with playing with Legos. I think for this generation it's about playing with AI, as you rightly said, like, AI is bold and big enough to be abused by anyone. Go and play with it. And I think it will have the patience to be with you.
- [48:36]
Jim Love
Yeah. If anybody was going to break it, it would have been me and I would have done it by now. But I love your answer because I think that's as a parent and my kids are now are grown and have their careers and I hope that even pre AI that I left them with the idea that be passionate and learn. It's no matter what people say, don't learn to don't learn this. Don't learn and be passionate and enjoy learning and because nothing I've ever learned has been wasted. And I think that's a. Yeah. So here's your chance to evaluate me as your interviewer here. What didn't I ask you about that you're excited about talking about.
- [49:17]
Krish Banerjee
I think we had a great conversation across varying different topics. You covered things that I would have wanted to talk about this, like more than a Q and A. I think these are the topics that I'm passionate about. And as you probably heard me talking about the value and the reinvention for Canada, these are topics that I'm very passionate and very close to. So happy to come back to your show anytime to double click on any of these topics if interested. But I'm hoping that the next time when we speak we would assess ourselves against some of these dimensions and say we have moved the needle on whichever way we are trying to do it in our lives.
- [49:58]
Jim Love
I think that's excellent and I think that's. I'm looking forward to that. I always look forward to our conversations just. But it's been great talking. We won't make it so long next time. We won't make it decades in AI time between.
- [50:11]
Krish Banerjee
I. I would be happy to be back, sir. Thank you. Great.
- [50:14]
Jim Love
Thank you very much.
- [50:15]
Krish Banerjee
Thanks, everyone.
- [50:16]
Jim Love
My guest today is Krish Banerjee. He's the Canada managing director at Accenture for AI and Data. And thank you very much, Krish. Great to have you on the show.
- [50:26]
Krish Banerjee
Thank you, Jim.
- [50:27]
Jim Love
And all of you out there listening to this. I hope you're on the dock having a nice time and enjoying the podcast. But you had lots of places you could have been on your weekend and you spent it with us. So thank you very much. I'm your host, Jim Love. Thanks for listening.