Latest / Partnerships Unraveled / Harish Natarahjan - Scaling AI Solutions with Practical Execution

Transcript

Harish Natarahjan - Scaling AI Solutions with Practical Execution · Partnerships Unraveled

• 0:07
  Speaker 2: Welcome back to Partnerships Unraveled, the podcast where we unravel the mysteries about partnerships, and channel on a weekly basis.
• 0:13
  Speaker 2: My name is Alex Whitford, I'm the VP of Revenue here at Chanix and this week I'm very excited to welcome our special guest, harish.
• 0:19
  Speaker 2: How are you?
• 0:20
  Speaker 1: doing.
• 0:20
  Speaker 1: I'm doing very well, Alex.
• 0:22
  Speaker 1: How are you?
• 0:23
  Speaker 2: Yeah, I'm extremely looking forward to this.
• 0:25
  Speaker 2: You have the unique blend, I think, of being someone who is extremely intellectual and extremely funny, which I often find doesn't go hand in hand.
• 0:34
  Speaker 2: So let's see if we can keep our audience on their toes.
• 0:37
  Speaker 2: Maybe for that, uninitiated, you can give us a bit of an introduction about who you are and what you do.
• 0:42
  Speaker 1: I'm Harish.
• 0:42
  Speaker 1: I'm part of the Accenture's high-tech strategy and consulting practice and I also lead over Gen AI efforts over there.
• 0:49
  Speaker 1: In fact, I've been on a bunch of projects over here building Gen AI solutions.
• 0:53
  Speaker 1: Also, looking at Accenture's and some of our clients' current state and defining their future state and the way I think about it is that we try to help organizations scale their AI journeys in a meaningful and practical way is basically what I call it.
• 1:11
  Speaker 1: It's not all theory.
• 1:13
  Speaker 1: It's not all just build POCs after POCs, but make sure that you have a thoughtful thinking around how you do it and I'm sure we'll be talking about a little bit more.
• 1:22
  Speaker 1: But before Accenture I worked for Motorola Solutions and Panasonic.
• 1:25
  Speaker 1: I call myself the industry guy 20 years over there and I used to build public misinterpreted products for public safety and manufacturing.
• 1:33
  Speaker 1: If you ask me what my key value prop is or my through line across all of these things has been around product mindset and innovation.
• 1:41
  Speaker 1: Think about it this way I'm sitting in front of you.
• 1:44
  Speaker 1: I like thinking about thinking of myself sitting in a problem space and asking what are the jobs that people need to do and how can we solve for it in a better and easier way and reduce friction, and so I always think about it.
• 1:58
  Speaker 1: But I've also learned over my career that thinking about problems is great, but executing is even better, and taking an idea to outcome is very, very important, and that's the mindset that I work with and that's how I work with my clients, that's how I advise at Northwestern with students as part of two courses that I advise on, and so I think technology's role should be in business as well.
• 2:25
  Speaker 2: Do you find there's a running joke on this podcast?
• 2:28
  Speaker 2: I hate the word strategy because you're, hey, let's run a POC for the POC sake.
• 2:33
  Speaker 2: It's what I see so much within sort of strategic level, sort of business decisions, where it's like I've got this really clever plan and it's going to be a 74 page document and I'm always like like sure, but what about?
• 2:46
  Speaker 2: When does physics become engineering like?
• 2:48
  Speaker 2: What about in the real world?
• 2:49
  Speaker 2: Do you find that that like really like tactical and practical and execution orientated mindset is rare?
• 2:56
  Speaker 1: I?
• 2:57
  Speaker 1: I wouldn't say it's rare.
• 2:58
  Speaker 1: I think it's about being integrated.
• 2:59
  Speaker 1: And again, if I think about why accenture I?
• 3:03
  Speaker 1: I was very conscious about my decision to get into Accenture and the reason was because it was a sweet spot for me in terms of saying I don't just ideate, I also get to execute.
• 3:12
  Speaker 1: And it's not a lot of people that can talk about in consulting.
• 3:16
  Speaker 1: In fact, I had an SVP that was talking to me the other day one of my mentors from Motorola Solutions and he said consultants come in and then tell us what to do and you know what we're already doing and then you walk away, don't you?
• 3:26
  Speaker 1: I said it depends on who you're talking to.
• 3:28
  Speaker 1: That's not Accenture.
• 3:29
  Speaker 1: It's basically what I told him.
• 3:30
  Speaker 1: Right, and he's a great mentor.
• 3:32
  Speaker 1: It's funny.
• 3:33
  Speaker 1: Funnily enough, both of his sons are in consulting and it's interesting for him to say that.
• 3:38
  Speaker 1: But it is what it is right.
• 3:43
  Speaker 1: And I feel, as we start thinking about value creation is great.
• 3:47
  Speaker 1: When you start thinking about problems, I think value capture is even more important.
• 3:50
  Speaker 1: How do you build scalable solutions that actually help solve that problem?
• 3:55
  Speaker 1: And that's what product management is, in my opinion.
• 3:58
  Speaker 1: It's not just about thinking, it's about executing that thinking.
• 4:02
  Speaker 2: And we had one of the most interesting prep calls I've had in a while and somehow I think, uh, I landed on something that sort of made sense when we talk about the world of ai one I think most people actually don't know why is, but that's a different podcast.
• 4:18
  Speaker 2: Second is, I feel they treat ai as some sort of magical hammer that breaks every single nut.
• 4:24
  Speaker 2: I know you're trying to make the magical logical.
• 4:29
  Speaker 2: Can you break down through some sort of real world client journey, how you one have that conversation, what does that conversation look like and how do you sort of drive the change that you're trying to see in a very mechanical way.
• 4:43
  Speaker 1: I'll put some data points in front of you.
• 4:45
  Speaker 1: In fact, we just had a release actions research come through that talked about frontrunners, and 8% of all organizations that we researched were frontrunners.
• 4:55
  Speaker 1: Think about it 100 companies and then eight of them are frontrunners.
• 4:59
  Speaker 1: And frontrunners are companies that have typically not just put AI solutions for the sake of it.
• 5:05
  Speaker 1: They've scaled it.
• 5:05
  Speaker 1: In fact, they've taken 34% of their ideas and then scaled it.
• 5:08
  Speaker 1: But on the flip side, we have 44% of these companies that are what we call experimenters and they don't have any scale solutions or, if anything, they're very minimal.
• 5:19
  Speaker 1: The quality of what they've scaled is not that great.
• 5:22
  Speaker 1: And when I start thinking about what is causing this problem and I've written a lot about this in my context, called AI with a product mindset, and I'm going to contrast this for you without talking about AI for a moment, I'll give you two examples.
• 5:37
  Speaker 1: One of my examples has been around a product called fee calculator that we built out when I was in my previous job at CCC Intelligent Solutions.
• 5:46
  Speaker 1: By the way, I didn't introduce it.
• 5:47
  Speaker 1: That was probably my only job that I was not in high tech.
• 5:51
  Speaker 1: Aside of that, everything has been a high tech company Now.
• 5:54
  Speaker 1: Fee Calculator was supposed to be simple.
• 5:55
  Speaker 1: Your car is in a total loss and you need to calculate fees and taxes.
• 5:59
  Speaker 1: Taxes are pretty straightforward.
• 6:01
  Speaker 1: Fees, apparently, was not that straightforward.
• 6:03
  Speaker 1: However, when you start looking at this market and you say 5 million total losses, $25 million in addressable market, you start seeing the value of what you can do.
• 6:12
  Speaker 1: But then you have to go back and ask yourself okay, I can sell this for $5 per pop, but why would someone buy it?
• 6:20
  Speaker 1: And that's where the value prop becomes important.
• 6:22
  Speaker 1: We said accurate and easy was the key message that we had to send out.
• 6:26
  Speaker 1: But if I say that accurate and easy is what I need to own, then I also need to own the data, the workflows and the change management associated with it.
• 6:35
  Speaker 1: I can't just go build a product, especially in a B2B environment, and toss it out at insurance companies and then say implement it, it just won't work.
• 6:40
  Speaker 1: And then say implement it, it just won't work.
• 6:44
  Speaker 1: The other piece of what we also did was when we started building this, this product had actually not launched a few times.
• 6:51
  Speaker 1: It was not a lack of trying or effort, it's just the strategy that you needed to take, and hence we said let's not go after these big insurance firms.
• 7:01
  Speaker 1: Let's start small, start with a few states and be iterative by design, and it's basically how we call it.
• 7:08
  Speaker 1: We built an MVP, we put it in front of our customers, we learned a lot and then we scaled it.
• 7:12
  Speaker 1: That's a non-AI product.
• 7:18
  Speaker 1: You still need to go through that same thinking when you start thinking about an AI solution.
• 7:24
  Speaker 1: Now I'll give you an example of what we recently started building.
• 7:27
  Speaker 1: Think of design validation as a you have.
• 7:31
  Speaker 1: Let's say, you're building a home and this is not a home, but think of it that way and you have a lot of constraints that you have to put together, electrical circuits that you have to build where do these need to be located your mechanical systems, your HVAC and everything that you need to put together.
• 7:47
  Speaker 1: So your county will come back and tell you here's what you need to build, and your designer comes back to you and saying here's what I'm going to do for you.
• 7:53
  Speaker 1: Someone's got to go and cross-check and ensure that what you're building makes sense.
• 7:58
  Speaker 1: But as we started doing this, we did a POC.
• 8:00
  Speaker 1: Six-week POC looked great and again, I'm going to stop using the word POC for a moment here, because POC is what is killing us.
• 8:10
  Speaker 1: It's proof of concept.
• 8:10
  Speaker 1: It's not a concept, it should be proof of value, and you have to start with a starting hypothesis of saying how much is this worth if I were to do everything?
• 8:16
  Speaker 1: But anyway, we did that.
• 8:19
  Speaker 1: We got good results, we moved forward into what we called a pilot solution and reality starts hitting you.
• 8:27
  Speaker 1: The standards on the data was not AI ready?
• 8:30
  Speaker 1: And there's a lot of implied knowledge.
• 8:32
  Speaker 1: And when I say vision, it was almost literally vision, in the sense that people were literally looking at a diagram and then saying it was right or wrong and they're marking them up.
• 8:41
  Speaker 1: We did not know what they were looking for, so we had to go and extract this information out of someone's mind the tribal knowledge and put it back in and you start seeing some of these problems that you start running into with ai solutions that are not native to non-ai solutions.
• 8:57
  Speaker 1: But there is lots of common threads.
• 8:59
  Speaker 1: We built it, we put it in front of a user.
• 9:00
  Speaker 1: We had to change the way they started thinking about how they use the solution, because now AI is telling you an answer.
• 9:07
  Speaker 1: You're not looking for something and they have to validate that answer.
• 9:10
  Speaker 1: And if you have people in your organization that are.
• 9:14
  Speaker 1: That's what I told you about sleuths.
• 9:16
  Speaker 1: They like the discovery and exploration of what they do.
• 9:20
  Speaker 1: They might not like the solution.
• 9:23
  Speaker 1: It's highly possible, and we actually saw this with the non-AI solution the fee calculator, as well.
• 9:28
  Speaker 1: There were people that came and told us I can calculate fees better than you can.
• 9:33
  Speaker 1: Your answer is wrong.
• 9:34
  Speaker 1: And I said okay, fine, capture every time you're wrong.
• 9:37
  Speaker 1: Give it back to us and we'll go back and check Capture every time you're wrong.
• 9:40
  Speaker 1: Give it back to us and we'll go back and check Out.
• 9:41
  Speaker 1: Of the 50 times that we got someone to log a problem, there were zero times where we were wrong.
• 9:48
  Speaker 1: It was just that people like their habits.
• 9:55
  Speaker 2: They do what they do and change becomes a very important thing.
• 9:59
  Speaker 2: So when you're, I think, one of the things that I'm constantly evangelizing, encouraging, demanding sometimes both my internal organization and also the partnerships that we build is, how do we build AI into emotion?
• 10:18
  Speaker 2: And I always think of this as partly process orientated and partly cultural, which is that you have to have a okay, we have this hypothesis, we're going to prove value I much prefer that to POC and we're going to do that and then we're going to work out how to scale, which, fundamentally, I think we understand.
• 10:30
  Speaker 2: The cultural element is, I think, the thing that particularly analytical people or strategy people will very easily overlook, because it's very tempting to forget that we're also encouraging a change of behavior of people.
• 10:44
  Speaker 2: Have you seen anything work particularly well to bring the masses on the journey with you?
• 10:51
  Speaker 1: I'll give you again.
• 10:53
  Speaker 1: I've been in manufacturing.
• 10:54
  Speaker 1: I told you I was at Panasonic.
• 10:55
  Speaker 1: I used to walk the floor all the time because that was probably the best way to understand how people use your products.
• 11:00
  Speaker 1: And you're absolutely right.
• 11:01
  Speaker 1: At a stratosphere level, where you're looking at things from a 10,000 foot view, you tend to ignore some of these real problems.
• 11:08
  Speaker 1: I've seen people literally cheat on our system.
• 11:12
  Speaker 1: They would start and stop almost immediately, and the reason they said that was interesting.
• 11:17
  Speaker 1: They said, well, I have to take three seconds to go click, do my job, take three seconds again to go click.
• 11:23
  Speaker 1: They were right about what was going on.
• 11:26
  Speaker 1: Why were we making them do things that were unnatural?
• 11:29
  Speaker 1: This is basically the question that I had to ask.
• 11:32
  Speaker 1: I had an interesting conversation with the director of engineering.
• 11:34
  Speaker 1: He was livid when he first heard that and I said, brian, this is not their problem, this is our problem.
• 11:39
  Speaker 1: We need to solve this right.
• 11:40
  Speaker 1: It's a real problem.
• 11:41
  Speaker 1: But if I take AI and step back and then think about it, we'd actually talked about this and the context was around.
• 11:49
  Speaker 1: There are two pieces to AI.
• 11:50
  Speaker 1: One is called cognitive trust, the other is called emotional trust, and the way to appeal this would be if I came and told you, and let's say you love cooking.
• 12:00
  Speaker 1: And I came and told you, and let's say you love cooking.
• 12:02
  Speaker 1: And I came and told you Alex, guess what?
• 12:03
  Speaker 1: Let's just make this up right?
• 12:05
  Speaker 1: I'm not a big fan either.
• 12:06
  Speaker 1: I just make pasta right out of it when my wife is out of town.
• 12:08
  Speaker 1: That's pretty much all I can do.
• 12:10
  Speaker 1: But so someone like me, I'd be very happy if someone came and said I'm going to cook for you, but someone that's like loves cooking, and you told them I'm going to cook for you.
• 12:22
  Speaker 1: There's two pieces to that.
• 12:23
  Speaker 1: One is a cognitive trust.
• 12:24
  Speaker 1: Do I trust you to actually cook well for me?
• 12:27
  Speaker 1: Let's say you're making a big dinner.
• 12:28
  Speaker 1: The second part is what I call emotional trust.
• 12:32
  Speaker 1: Do I want you to do that job Because I love cooking?
• 12:35
  Speaker 1: And when you take that broad stroke of cooking, people might say oh no, I love cooking, why are you taking my job away?
• 12:44
  Speaker 1: When you break that down and you say cooking has several pieces to the puzzle.
• 12:49
  Speaker 1: Cutting vegetables is one piece of the puzzle and a lot of people think of that as a mundane task.
• 12:54
  Speaker 1: Once I teach people how to cut vegetables, they know how to cut vegetables and once I trust them to do that, they would be okay with doing something like that.
• 13:01
  Speaker 1: But throwing seasoning on it and then making that final piece of the puzzle work Probably not right.
• 13:07
  Speaker 1: I think that's where I feel human in the loop when we talk about it.
• 13:11
  Speaker 1: That's where you need to make it real.
• 13:12
  Speaker 1: You have to ask people which parts of these shops do you like, and sometimes you have to automate things that people love doing.
• 13:20
  Speaker 1: Coding is one of them, where you're beginning to see white coding happening and people are going to have to learn new ways of building products moving forward.
• 13:29
  Speaker 2: In talking sort of double-clicking on products.
• 13:31
  Speaker 2: I really like your sort of narrative around AI from a product mindset because I think again, if people have a misunderstanding of what AI really is which is AI is some level of architecture that is also extremely nuanced and specific.
• 13:49
  Speaker 2: I know when we were prepping, you were talking about the importance of industry context, so sort of verticalization as a nuance right.
• 13:57
  Speaker 2: Talk to me about why that is so fundamentally important to AI work.
• 14:03
  Speaker 1: It goes back to what you were beginning to ask me at the beginning of this conversation.
• 14:07
  Speaker 1: Do we know the jobs that people are doing?
• 14:09
  Speaker 1: And when I think about a user, buyer, the jobs that they are doing are fundamentally different.
• 14:15
  Speaker 1: I was in the hospital and we had someone coming and doing my ultrasound and I was watching.
• 14:23
  Speaker 1: I was super curious about this right and I had a different problem deal with.
• 14:26
  Speaker 1: But yet I'm asking her what is going on over there and she talks to me about how AI has been solving some of the things that she would normally do, which is kind of interesting.
• 14:36
  Speaker 1: She was taking the scan and then the AI was basically running behind the scenes to do some analytics.
• 14:42
  Speaker 1: Now you would only get to know that level of detail if you get to an industry level Software engineering, for example.
• 14:49
  Speaker 1: If you're talking about companies like HP, dell, motorola Solutions, they build a lot of embedded systems.
• 14:55
  Speaker 1: You go to a company like Salesforce, the type of products that they build are lot of embedded systems.
• 14:58
  Speaker 1: You go to a company like Salesforce, the type of products that they build are fundamentally different and the type of software that they write is also different.
• 15:04
  Speaker 1: You can't take what you build for a company like Salesforce and then try to adapt it as is into another firm.
• 15:12
  Speaker 1: That's building hardware.
• 15:13
  Speaker 1: You can take patterns, but you can't take the product as is.
• 15:17
  Speaker 1: That's where I believe industry nuance is very important, especially coming from high tech, and I can speak to this far border and I've worked in the insurance for a very small sliver of time CCC and I've seen the differences in terms of what happens over there.
• 15:29
  Speaker 1: I think that's the fundamental difference, and now you basically start taking every piece of this job and everything that people do creatives, for example.
• 15:36
  Speaker 1: What they do for high-tech firms in B2B is different than what Apple would be doing in a consumer market or Bose would be doing in a consumer market.
• 15:45
  Speaker 1: I believe that's what is important and I'm saying even within high-tech, if you think about it, apple and Bose are a different type of company than HPE HPE and Dell, for that matter.
• 15:58
  Speaker 2: Yeah, I think AI gives us the ability to do sort of extremely specific answers, but that means you need extremely specific context.
• 16:07
  Speaker 2: My bet we'll see if it comes out is we are going to see such specialism in verticalization and so businesses that previously sold to or served and serviced manufacturing and agriculture and pharma that will soon die out, because what will happen is there'll be an AI service model that is just specialised in finance.
• 16:30
  Speaker 2: That's all they do, because they can lift and shift and serve and learn at a far better rate, which gives them the more specific answer and therefore, I think, the tighter value.
• 16:41
  Speaker 1: I would add to that there are always going to be general-purpose products.
• 16:45
  Speaker 1: If you look at Microsoft Office 365, it's a general-purpose product and you can build a lot of things on top of that.
• 16:53
  Speaker 1: There's always going to be that space.
• 16:55
  Speaker 1: In marketing, for example, we see a lot of companies like jasper writer come along that can do a lot of things well, but context becomes important.
• 17:04
  Speaker 1: So you build a context on top of the product and there are going to be products or ai solutions that you're going to have to build that are very specific to an industry.
• 17:13
  Speaker 2: Potentially they're both going to happen and um, we are speaking because of a friend, of a mutual friend of ours in terms of david, who works at hp.
• 17:23
  Speaker 2: I've sort of been aware for a long period, essentially sort of just thrown into a bucket of gsis that I'm assuming everyone sort of wears black robes and sort of drifts between the trees, I'm not really sure.
• 17:38
  Speaker 2: Talk to me in terms of where Accenture, large enterprise customers and HP, what that Venn diagram looks like.
• 17:46
  Speaker 1: Yeah, I'll put this in a slightly different context.
• 17:50
  Speaker 1: I say and again, I come from the industry and I've been there for several years and I came into consulting.
• 17:56
  Speaker 1: My learning about Accenture specifically was that we play three roles translator, integrator and a catalyst.
• 18:03
  Speaker 1: Translator, because we talked about industry expertise, bringing that jobs to be done notion.
• 18:08
  Speaker 1: And integrator because you're beginning to see that AI solutions can't just be a dropship anymore.
• 18:16
  Speaker 1: You have to implement it, you have to drive change management, so integration of the software becomes important.
• 18:20
  Speaker 1: And a catalyst because there are go-to-market motions, for example, that we work with several clients, and the fact that we have access to some of our clients in very different ways and nuanced ways, especially within IT, within their business units, gives us the ability to go have a conversation about how to bring a product into that, into their environment.
• 18:44
  Speaker 1: Is that a help answer or Nothing?
• 18:49
  Speaker 2: like what you're saying.
• 18:50
  Speaker 2: I'm maybe then to sort of focus in on the energy environment, because I think that's the bit that's interesting.
• 18:56
  Speaker 2: I really like macroeconomics.
• 19:00
  Speaker 2: I think history has a lot to teach us in terms of where the future is headed.
• 19:05
  Speaker 2: Ai is a bit different.
• 19:08
  Speaker 2: So we have seen every sort of bleeding edge technology.
• 19:12
  Speaker 2: I think if we look at how networking and MPLS changed Communication and go to markets for large enterprise businesses, it was enterprise first because it was so cost prohibitive.
• 19:23
  Speaker 2: Smb and mid-market couldn't adopt the technology we seem to now with AI and cloud and SAX as a sort of Venn diagram.
• 19:32
  Speaker 2: That cost prohibition is not the same.
• 19:35
  Speaker 2: I think there's a fair argument to say we aren't going to see enterprise lead, then mid-market, then SMB, then consumer, which sort of every other life cycle of sort of technological change has happened.
• 19:48
  Speaker 2: What's your thoughts?
• 19:52
  Speaker 1: Let's scale this back to five years even for that matter, and maybe let's just say I kind of like show up, give out my age over here, way back in 1999, when I was building my first neural network.
• 20:05
  Speaker 1: That was my graduate project.
• 20:07
  Speaker 1: In my undergrad I had to do everything.
• 20:10
  Speaker 1: I had to write every piece of code to make these weights work and whatnot right.
• 20:15
  Speaker 1: But if you go back, like 2015, 2017, ai has become a lot more accessible, and what I mean by that is if you think about TensorFlow or all these PyTorch and all of these other deep neural net models that have come through, the way you use them has shifted.
• 20:33
  Speaker 1: It's not changed the thinking.
• 20:34
  Speaker 1: You still need to have a good data scientist that knows how to use these solutions, but at the same time, it's there Now.
• 20:40
  Speaker 1: We talked about cloud and I'm going to.
• 20:43
  Speaker 1: Since we talked about David over here, that's the one that made the introduction I'm also going to talk about the other piece of this puzzle, which is an AI PC, and as I start looking at workloads and where cloud was playing and asset service models were happening, some of the workloads that could potentially happen on an AI PC.
• 21:03
  Speaker 1: Put it this way you're going to buy a PC anyway.
• 21:04
  Speaker 1: It could be an AI PC and the price points are going to come down, become more favorable.
• 21:07
  Speaker 1: But when you do, you could do some of your work on an AI PC directly.
• 21:13
  Speaker 1: But AI PC might not be the right workload for everything.
• 21:15
  Speaker 1: It might have to go into the edge, might have to go into the cloud.
• 21:19
  Speaker 1: And that's the point of view in which I'm kind of beginning to think about how are you going to build an orchestrated set of solutions?
• 21:26
  Speaker 1: And I think the solution mindset is basically what I feel you're going after over here, where people are going to say, well, I got a solution, but I can't just basically run it on the cloud anymore.
• 21:39
  Speaker 1: I'm going to have to run it everywhere.
• 21:40
  Speaker 1: And what is this model going to look like?
• 21:42
  Speaker 1: And I don't know that as a service and all of those things people keep talking about.
• 21:45
  Speaker 1: As a service instead, I don't believe that it's still going to be there.
• 21:48
  Speaker 1: It's just going to evolve into something else.
• 21:49
  Speaker 1: But that's where I feel things are going to land.
• 21:56
  Speaker 2: And it's all to me.
• 21:58
  Speaker 2: I'm always sort of there's the strategic element in terms of where it's going to finish.
• 22:03
  Speaker 2: How can brands like HP encourage that change of behavior?
• 22:09
  Speaker 2: How can they accelerate that mission?
• 22:12
  Speaker 1: If you're talking about HP, and I'll talk about all organizations, all OEMs in general, because I think it's probably only fair if you do it that way, right?
• 22:21
  Speaker 1: One of the things that I've been saying is take a and we've done a lot of research around this.
• 22:29
  Speaker 1: In fact, we had a position paper, we put out a point of view just recently March or April.
• 22:34
  Speaker 1: I know March or three months is a long time in the world of AI, but nevertheless, I think some of these things were interesting.
• 22:40
  Speaker 1: They said take a solution.
• 22:41
  Speaker 1: Mindset is basically what we started talking about.
• 22:45
  Speaker 1: If you think about what Android and iOS did when we first started with smartphones and it's not like smartphones did not exist before that Chrome had a smartphone, Windows had a smartphone, but the ecosystem did not exist.
• 22:59
  Speaker 1: Apis did not exist.
• 23:00
  Speaker 1: There were no smartphone.
• 23:01
  Speaker 1: Windows had a smartphone, but the ecosystem did not exist.
• 23:03
  Speaker 1: Apis did not exist.
• 23:03
  Speaker 1: There were no out-of-box capabilities.
• 23:04
  Speaker 1: There were no capabilities for developer SDKs to go build solutions and then put it in a marketplace.
• 23:07
  Speaker 1: A marketplace did not exist, even if developers could build solutions.
• 23:12
  Speaker 1: So what can companies like HP, Dell, Lenovo, all do and there are lots of other players over here to reduce this friction in deploying an AI PC, for example, in the market quickly, and I believe that's part the hardware itself, which they're really good at, and part the orchestration layer that they're going to have to build in order to make sure you're not just solving for the user journey, but also the enterprise journey.
• 23:41
  Speaker 1: So let me explain that.
• 23:42
  Speaker 1: You might have an AI PC, I might have an AI PC, but my AI PC might be running at 40 tops.
• 23:49
  Speaker 1: Your AI PC might be running at 20.
• 23:51
  Speaker 1: Things that I can do on my AI PC might not be things that you might be able to do on yours.
• 23:56
  Speaker 1: Now, if you start thinking about what, I can't, tell you that you can't get any AI goodness for the next two years until you get to a new laptop.
• 24:03
  Speaker 1: Right, I'm still going to have to give you the benefits of AI.
• 24:06
  Speaker 1: I'm just going to have to find a way to orchestrate that better.
• 24:09
  Speaker 1: Maybe it goes to a near edge or it basically goes to the cloud.
• 24:13
  Speaker 1: Whatever path it needs to take, that orchestration needs to happen, and so that's something that I believe these OEMs can do.
• 24:20
  Speaker 1: The other piece of it is with great powers come great responsibility, I guess.
• 24:25
  Speaker 1: So how do you make sure that people don't misuse AIPC on because, oh, I'm just going to go get a new large language, small language model put it in and you don't know if it's been vetted properly by the organization, you don't know if it's allowed to be used.
• 24:40
  Speaker 1: These checks and balances still need to happen and I think organizations have to have a way to solve for it.
• 24:47
  Speaker 1: So I think that's where I believe the integrated solution that these OEMs can build is going to make it easier for channel partners to implement or SIs like Accenture to go and implement these solutions into it.
• 25:02
  Speaker 2: That's awesome.
• 25:03
  Speaker 2: So maybe pivoting fairly significantly.
• 25:06
  Speaker 2: You have a complicated and very interesting job.
• 25:10
  Speaker 2: I know you've worked very hard to build a team around you to sort of support that motion.
• 25:16
  Speaker 2: When we were talking about hiring practices or characteristics that you look for which is one of my favorite things to find in organizations where they have not a standard go-to-market position, where it's like, hey, I do marketing and this is marketing you raised intellectual curiosity and intellectual relentlessness as two characteristics that you sort of hone in on.
• 25:39
  Speaker 2: Why?
• 25:40
  Speaker 1: and how.
• 25:43
  Speaker 1: One of my mentors started talking to me about this, and this was during an interview.
• 25:47
  Speaker 1: I was getting into this job, let's say, one foot on what I know, one foot on an area that I had no background in, and we were talking about it, and he said, listen, I'm going to hire you.
• 26:00
  Speaker 1: And I'm going to hire you because there are two things that I can never teach people intellectual curiosity and intellectual relentlessness.
• 26:08
  Speaker 1: You seem to have both.
• 26:09
  Speaker 1: I can teach you everything else.
• 26:11
  Speaker 1: And I think six months later, he and I had a conversation because I was pitching a different idea to him, and he asked me what did you do to get to the answer that you did?
• 26:22
  Speaker 1: I said, well, there was a towing truck, and this was my previous company called CCC, and they had a towing truck and I needed to see what happened when a car got totaled.
• 26:30
  Speaker 1: So I just followed the towing truck and there's nothing wrong with it.
• 26:33
  Speaker 1: I was not stalking them, I didn't get into the yard, but I just needed to see what was going on and I understood these towing companies have their own incentives to do what they do.
• 26:43
  Speaker 1: And as I was talking to him about it, I said, mark, this is what goes on and he's like that is really cool, and I think that's what I look for in people Are you intellectually curious?
• 26:54
  Speaker 1: Curious is basically are you interested in understanding a problem space, exploring it?
• 27:01
  Speaker 1: Relentlessness is about executing, and that's not lack of strategy.
• 27:06
  Speaker 1: It's basically taking that idea into practice, right, and I, when I look for people, I generally look for two types of people.
• 27:16
  Speaker 1: One person are people that can solve complex problems.
• 27:21
  Speaker 1: You can throw it at them, you can give them a 10-minute word of mouth of here's what I'm looking for, and you would see some dramatic results come back from them, because they would have thought very deeply about the problem statement.
• 27:33
  Speaker 1: The other set of people that I generally look for are people that can bring in deep research and then synthesize that answer into two or three sites.
• 27:41
  Speaker 1: We've done client work where we had to put a product market together, and electronic manufacturing is not easy, especially contract electronics manufacturing.
• 27:50
  Speaker 1: All we had to do was to basically get that data out.
• 27:54
  Speaker 1: It was about 20 or 25 pages of information that we had collected.
• 27:58
  Speaker 1: But how do you synthesize this into a go-to-market strategy?
• 28:01
  Speaker 1: That's important, right?
• 28:02
  Speaker 1: So there are some people that are able to do it with guidance.
• 28:05
  Speaker 1: That's what I look for?
• 28:06
  Speaker 1: Are they interested in solving the problem and can they actually solve the problem?
• 28:11
  Speaker 2: And do you find that one is rarer than the other?
• 28:15
  Speaker 1: It's just different and I basically say I pick my people based on the needs that I have.
• 28:22
  Speaker 1: I'll give you an example If I'm doing a deep assessment and then I'm trying to build what would look like an operating model for a client and operating models, we have standard templates of how operating models should work.
• 28:36
  Speaker 1: Believe me, every client is different and the operating models need to be customized and tuned to their specific environment.
• 28:43
  Speaker 1: And when you do that, you need to have people that have a deep understanding of the client and then bring that answer back.
• 28:49
  Speaker 1: So it's not a lot of information that you're consuming, it's just the answers that you're putting out can be very complex and you have to be able to say it in an advanced way.
• 28:59
  Speaker 1: The other is more around.
• 29:00
  Speaker 1: Can I throw a lot of information at you and help you synthesize it?
• 29:04
  Speaker 1: That typically is like a product market fit, product strategy, road mapping, prioritization type of work.
• 29:11
  Speaker 1: That's a different.
• 29:12
  Speaker 1: I think there are two different skills.
• 29:13
  Speaker 1: You will see some overlap, but I'm okay with either of them.
• 29:17
  Speaker 1: It depends on what kind of job I'm looking to solve for at that point in time.
• 29:22
  Speaker 2: Awesome, parish, that's been amazing.
• 29:25
  Speaker 2: It's really interesting to get into the world of consultancy.
• 29:28
  Speaker 2: I need to ask you for a bit of consultancy on this podcast.
• 29:31
  Speaker 2: We always ask our current guest to recommend our next guest.
• 29:34
  Speaker 2: Who did you have in mind?
• 29:45
  Speaker 1: you have in mind.
• 29:45
  Speaker 1: I'm going to propose EJ Tavella.
• 29:46
  Speaker 1: He is the SVP of product supply chain, product set Anaplan.
• 29:47
  Speaker 1: I knew him from when he was at Accenture a few years back.
• 29:49
  Speaker 1: In fact, one of the POCs that I was talking to you about was Andres Diotlich.
• 29:55
  Speaker 1: He's a phenomenal leader.
• 29:57
  Speaker 1: In fact, I still keep in touch with him.
• 29:58
  Speaker 1: I talk to him every so often.
• 29:59
  Speaker 1: He's a phenomenal leader.
• 30:01
  Speaker 1: In fact, I still keep in touch with him.
• 30:02
  Speaker 1: I talk to him every so often, more from an intellectual perspective.
• 30:03
  Speaker 1: I think one of the best learnings that I got from him was you can't be good at everything, so focus on your strengths and leave the rest to other people that can do it better than you can, and I think product management is one of those places that teaches you a lot of bad things.
• 30:20
  Speaker 1: You tend to be a lone wolf and do a lot of things yourself.
• 30:22
  Speaker 1: Consulting is a different place.
• 30:25
  Speaker 1: You've got to bring people on board and get them to do what you want them to do.
• 30:30
  Speaker 1: So that was my learning from him.
• 30:33
  Speaker 1: I still keep in touch, so I think he'd be a fantastic next guest for you.
• 30:37
  Speaker 2: Awesome, Harish.
• 30:37
  Speaker 2: Thank you so much for the recommendations.
• 30:39
  Speaker 2: Thank you so much for sharing your wisdom.
• 30:41
  Speaker 2: It's been awesome.
• 30:42
  Speaker 1: Wonderful.