skip to Main Content

About this episode

Unlocking Business Potential: The Rise Of Edge | Mili Iyengar – Accenture | The IoT Podcast

In this episode of The IoT Podcast we are joined by Mili Iyengar – Global GTM Leader – EDGE Computing & IoT | Managing Director at Accenture to explore how the combination of IoT, Edge computing and AI is improving lives and reinventing businesses.

We’ll about hear how Accenture empowers companies to scale, transform, and unlock value through the combined magic of these technologies. Plus, showcase some incredible use cases from major brands including Starbucks and McDonalds, revealing how they’re transforming their business models through leveraging the latest edge, IoT and AI Solutions.


Tom White (00:01.73)
Welcome back to the IoT Podcast and welcome to Millie. Hello, Millie.

Mili Iyengar (00:07.496)
Hello, hi Tom, always great to see you.

Tom White (00:10.466)
Absolutely and good to see you as well as we were laughing backstage. I think we’ve had this in the calendar for quite some time, but I am absolutely delighted to welcome a century on to the IoT podcast and learn more about you, what you’re doing in this wonderful crazy world of IoT and edge computing. So without further ado, could you just explain a little bit about who you are and the business that you represent then, Millie?

Mili Iyengar (00:35.592)
Thanks for having me, Tom. You know, super excited about our conversation today. Within Accenture, you know, we’ve been thinking for some time now when almost three years back we launched Cloud First as an organization on what the next iteration for business transformation would be. And unbeknownst to us, you know, Gen .ai was not a thing really, but we’d started thinking about how what we call cloud continuum.

would affect clients and their aspirations in leveraging IT. So there was this group that was set up called Edge Computing Practice, almost two years back now. And the goal for us was again, to start dabbling with emerging technologies. And that went all the way from, like I said, stretching cloud and cloud -like architectures all the way to the edge.

And then within the Edge remit, how then devices or smart devices, connected devices are going to play out, how IoT sensors going to fuse into that ecosystem and how it’s so much heterogeneity at the Edge, businesses are going to try and maybe adopt some standards for interoperability and how that would all pan out. So I think again, from an Accenture perspective, a little bit of our thinking into…

setting this up as a very focused area with experts in the field. And my particular role is to drive go -to -market and sales for this particular group globally. So it’s been, again, a super interesting ride, joined Accenture, about the inception of this practice. And we’ve had some great examples and great wins thus far helping customers.

Tom White (02:22.978)
Great, well thank you for the overview and well done actually for executing and taking an idea into reality. There’s lots of customers that talk about trying to get into something, but they never seem to get off the drawing board. So it’s really, really good that you’re now putting it into action and you’ve got some real life case studies and examples, which I’m sure we’ll get onto at some point today during the podcast. But if we could talk Millie just a little bit about…

your specific interest in your role today. So it’s Edge and IoT and it’s go -to -market. So for the uninitiated, what does that mean in real terms then? So you’re specifically working with new customers, is that right?

Mili Iyengar (03:09.576)
100%. This is such a great question, Tom, because the first kicker when we kick off any conversations literally Edge does not have a very defined or a standard definition in the industry right now. In the past, we would call the edge of the network as the edge and then it kind of transposed a little bit into is distributed computing or on -prem computing Edge.

I think what we are settling in within Accenture again, it’s a very loosely defined definition is anything, any workload that is not in clients data center or already on public cloud is essentially in the remit of edge computing. And we see it sort of in three different buckets. You could have smart devices which have compute inside of it. Example.

you know, cameras that come with some amount of say pre -processing or chipset designs, take your Oculus and you know, that has a silicon design embedded in it, which again does some computing. So we call that the device compute. The second bucket for us is what we call enterprise compute. And that is in a sense, if you think about, you know, distributed environments. So if it is a manufacturing client, all of the plants or sales offices.

that typically had servers in all these distributed locations and then one would have to manage them centrally is in the remit of enterprise edge. In retail, it could be, you know, stores that our retailers have in QSRs, restaurants, in oil and gas refineries, and some of these remote locations. And then we call, you know, third bucket as the network edge. We strongly believe that…

with the advent of 5G and large scale adoption across the globe, there will be some amount of computing that will sit in base towers and could be offered with the density that 5G requires at the network’s edge, the traditional way we would have operated or cloud when kind of solutions in that bucket. So that in a sense is three buckets of edge. I always keep reminding everybody edge is not this ubiquitous thing. It is literally.

Mili Iyengar (05:19.24)
different things to different businesses and where you can harness the value across those through spectrums also differ. You can go with point solutions. So example, shopping and autonomous grab and go kind of solution as a very device edge, enterprise edge type of solution at the same time, if you are rolling out 5G in a factory somewhere and you or a company wants to tap.

large scale into using that as a compute engine as well. And that essentially could be network edge as well. So in a sense, that entire spectrum, I would say, is edge computing. Whereas IoT particularly fit into the spectrum is super simple. All of the data in the real world that gets generated is going to get into digital form.

through the use of these translator mechanisms. And we think IoT and sensors and computer vision are this translator mechanisms which move then analog signals into digital, right? So that is where the role of IoT becomes so important. It is literally the engine that generates the data, captures the data that is generated in the real world through temperature sensors and vibration sensors and pressure gauges and.

like I said, again, computer vision and the amalgamation of that is then ramification of it is to build solutions that can make sense of it and can take actions and can be connected and interposed with some kind of logic. And that is where I think the AI ML in the world of GenAI then kicks in and end to end super exciting use cases emerging.

Tom White (07:05.922)
Yeah, my god, I mean, I absolutely can agree with you, right? I think there’s so many different areas and different case studies that we see and that we talk about. And that’s one of the beauties with Accenture is that you’re touching so many different industries, you’re touching so many different, you know, areas in which people can be using Edge and IoT. So it’s great. It’s great to have you on the talk about that. So my next question was going to be,

Personally, what excites you about this whole intersection of IoT Edge and we touched on Gen. AI earlier, when it comes to business transformation, especially when it comes to your vantage point being at Accenture, because I think that’s really, really useful actually to understand where you get most excited having seen this from so many different angles.

Mili Iyengar (07:55.208)
You know, Tom, personally, I think for me, I’ve been 20 plus years in the industry, started with IBM and have seen that almost sino cycle cycle of, you know, mainframes as kind of central compute to cloud computing, which in my mind went through the other curve. And then now back to, you know, a little bit of distributed computing and edge. What is super exciting is the power of this technology to capture the data that is.

out there, which frankly, we weren’t necessarily looking at in the past, again, to drive business value, right? So I’ll give two examples. One, the field is super exciting personally, because I have, you know, family who are special needs. And I realized very early on, especially with kids, the power of smart devices. So think of a hearing aid or a cochlear implant. Think of, you know, an ADHD kid.

Tom White (08:27.522)

Mili Iyengar (08:49.096)
using smart tools, example, a smartwatch or a device that can give them prompts and reminders and notifications truly transforming lives day to day. And that in the business world interposes with these technologies almost providing superpowers to either employees or customers, right? And changing the paradigm in which we operate to make life simpler, to make life a little easier, to give us more time to…

spend in areas that really matter to us, what we value and that intersection of the possibility of this tech to be able to truly, truly impact our day -to -day life is what makes me most excited. So for example, and maybe I’ll go into an example to just make a point, right? We are working with global QSR.

Tom White (09:38.85)

Mili Iyengar (09:43.08)
where we are going into their restaurants and over the next five years, we’ll transform all of the compute that they have in their restaurants. And on the surface, it seems like, so using this, are you going to end up selling more burgers and fries? And is that the end goal and objective? But if you think about the underlying impact of some of this tech, even if we just stick with that example, right? So if, yes, this computer is going to make it faster to sell burgers, who is buying these?

It’s so surprising that in the communities that buy this, the average order is typically cash and $8 globally. So we are almost thinking about serving a strata of population that depends on these kinds of business models and ecosystem. And it’s a huge impact. One seventh of the population of the entire earth visits this QSR at least once annually, right? So.

Again, back to my point, the power of some of these technologies to transform lives at the grassroots level, make it super simple and easy at scale is unimaginable, right? If you think about autonomous driving, which is another very super simple example of edge computing and the power of sensors and IOTs coming together, it’s going to truly transform the way we operate in the world. And that is super, super, super exciting.

So I’ll pause on that note and give you another opportunity to kick back with one of them, smack the word.

Tom White (11:15.074)
Yeah, well, that really is something special. I think, you know, as you mentioned around the using IoT and edge computing for the greater good, I think that that’s really quite personal and poignant to you, you know, as you said. And I think when we talk about, you know, are we selling more burgers or are we doing things that don’t necessarily touch the heart? I think that’s really, really.

really, really poignant and it’s great what we can do now with the combination of the two technologies together. So it’s really nice to hear that, you know, really. Yeah, absolutely. So in your experience then, so moving on quickly, you know, how has, you know, IoT been instrumental in reshaping what was a traditional business model? And have you observed any kind of acceleration about that and transformation in recent years and what’s contributed to that?

Mili Iyengar (12:11.912)
Yeah. So again, two things in my head, right? One, every company across the board or any enterprise, or even for that matter, our households are going to get better value if we digitize. Simple. That is a crux of greater good, right? If we are able to digitize, if we are able to harness the basics, you know, take data feeds, it’s going to change our lives for the better. And that is true for enterprises. Now,

think about one or stream almost when it comes to IoT harnessing the power and sensors to make the products itself smart. So example, if I have a refrigerator and it’s going to automatically tell me, you know, in our busy lives, Hey, the eggs have run out or the juice is no more there or the milk is, you know, coming up. My point is these smart devices are certainly going to make life super simple and easy. So I see a lot of companies right from heavy industrials to

Even something like a tractor company trying to make that intrinsic product smarter. And again, how do you make your product smarter? By introduction of sensors so that you are able to take a lot of data feeds and are able to make sense of the environment around you. So in that particular example, if my tractor or digger is able to, while it is going and just getting the soil ready for the next crop,

also measure, you know, wetness factors, density, are there chemical imbalances in the soil? Then obviously the productivity from what’s to come downstream can be proactively corrected, right? And that is going to obviously create huge amount of economies across the globe. So I see a lot of companies across the globe trying to reinvent themselves and their product suite to become smarter and…

that’ll in turn lead to a connected ecosystem of sorts, which is going to give us huge amount of uplift, productivity and value creation is going to happen. Then on the other side of the spectrum, almost all companies want to turn into if possible tech companies, right? So our head, Karthik Narayan keeps saying like in terms of even a startup, if you think about opening a new business,

Tom White (14:25.986)

Mili Iyengar (14:34.568)
It’s not enough for you to look out for staff and building to start your operations and what’s your core product. You’ll first thing go and get a, you know, cloud account set up, whether it be on Google, Microsoft, AWS, whatever. And that is one of the basic tenants of starting a company today. So by the same remit, you know, how are you going to therefore use technology? Not as a, as a good to have, but as an intrinsic part to drive business and the paradigm shifting to it.

is I think crazy critical and more relevant now than ever with Gen .E .I. Gen .E .I. itself, I think, not just the productivity, but the value that it provides us from being able to assist us, support humans and getting more done in their day is also going to require and place huge, huge demands on compute across the globe.

And because a lot of the workloads require very low frequency or we don’t want data to be outside of our four walls, it’s going to impact almost all enterprises in thinking through how therefore edge computing and IoT comes into the picture. And then if they take a strategic approach to having an edge strategy and an IoT strategy and how they’re going to connect that to their basic reference architectures.

I think they’re going to be ahead of the curve rather than react to some of these parameters that get generated in the industry. So I think again, those two buckets, companies trying to reinvent themselves and their product suite to become smarter. And then the second bucket where companies trying to make tech almost drive their business and be a cloud first approach as against thinking of it reactively.

Tom White (16:24.994)
Yeah, I think that’s a really good explanation around the two pivotal moments that companies there face and how they can use bleeding edge technology to be able to leverage more success within that business. However, they measure that success. So I think now would be a really good time to talk about some of the case studies that we spoke about, actually, because what’s really, really useful.

here is that I touched on earlier was that there is a variety of case studies, but also some real corporate giants that are essentially working on the digital transformation to both Edge, IoT and AI solutions. So if we could talk about a couple of those, that would be excellent. I think in particular we’re going to talk about it with Donald and Starbucks. I think we’re going to talk about today as well.

Mili Iyengar (17:05.928)
100 %

Mili Iyengar (17:13.704)
Yeah, again, the way they’re using edge computing, right? If you think about it again, McDonald’s is trying to reinvent the entire IT stack that they have in the restaurants to become very reliable, available in the next generation applications like their POS to be able to really harness all of the data that can be leveraged and interperse that with other data streams. I think the key aspiration is to be real time.

Tom White (17:16.674)

Mili Iyengar (17:43.784)
So that is the underlying aspiration from a company saying, why do I need to do pricing releases in X hours? Let me do that every 15 minutes. And that’s like setting the gold standard and benchmark for how real time they want to be. And I think, again, from an Accenture perspective, you almost see this as an ability to bring the operations or the production side of the house, which is the core business.

together with the power of data and using all these basic techs and setting up the foundation first with edge computing to enable multiple, multiple use cases on top. And what I mean by that is once they have this edge compute, which is super reliable, it is a container -based architecture provided by Google.

they’re not going to be in the business of break fix anymore. So even your erstwhile SLA is what if something goes down, fix it in four hours is going to go away. The automation that is already part of the container architecture is going to almost give them a system that is up all the time. And that is the big shift I think that everybody probably needs to understand that we are talking about a highly reliable, highly scalable, highly available system.

And once you have that foundation from an edge computing standpoint, all of the use cases that are going to come on top. So think about order accuracy. Can you take computer vision models? And when a crew member is building the burger, if a particular order says no cheese, can the human or the crew member be assisted through, you got it wrong.

can order accuracies go up and that impacts that customer satisfaction and downstream effect is all positive. So that is one use case maybe that’s going to help in quality control if you think about it in that operation. The second use case could be run retail media and that can unlock almost a new revenue stream for a very traditional otherwise restaurant chain, right? So.

Mili Iyengar (19:51.432)
Again, the power of harnessing digital screens, connecting all of that media network and bringing that to the restaurants, huge power to counteract some of that. If we now think about all of the devices that they have within that restaurant, so take coffee machines and ice cream machines and girdles and fryers. And if there is proactive, again, amalgamation of all of that data of

hey, the fryer has been used X number of times. And so, you know, maintain it now or change the filters or whatever, back to proactiveness in operations is again, going to make life easier for the crew members and give them superpowers, right? So our thinking from an Accenture point of view is multiple customers can start thinking about laying down that core digital foundation.

And that is in effect using edge computing and using the next iteration of maybe their hardware refresh cycles to get that fundamental foundation for edge in place. And once you have that, from an IT standpoint, you are almost providing a platform to your business to unlock the value of that core foundation in multiple different manners and automate so many things on top. It could be, again,

automatic voice order taking in that particular example. And that is again, going to help you drive more efficiencies and accuracy in the way you accept orders and fulfill them. So almost every other part of business, you know, human safety. So if there is an oil spill, send a notification to the manager to get it cleaned up. If there is a security issue outside of your four walls, and you know, there’s just general unrest.

you should be notified as a manager to say, lock your doors and don’t send your people out and call probably 911. My point being it almost provides you extra eyes and ears from the amalgamation of again, computer vision, IOT, edge computing, and that gives any human superpowers, right? Or any business superpowers. So that is essentially McDonald example. And we are so, so proud to be their partners and thankful.

Mili Iyengar (22:03.688)
that we have this opportunity to really use this technology globally and help them in their transformation journey in 120 different countries, 40 ,000 restaurants, so super excited about it.

Tom White (22:18.05)
Yeah, that is quite colossal, that scale, isn’t it? 120 different countries and 40 ,000 different restaurants. I mean, that’s massive, isn’t it? And that, I suppose, really is the test to show that this is working because it’s going to be done on such a large scale, isn’t it?

Mili Iyengar (22:32.904)

100%. And I think that is the other thing we didn’t talk about, right? There are a lot of companies that are taking an approach, which is a POC driven approach. So we have another client of ours, they are Mars, you know, which makes Skittles and all of the nicer chocolates that you and I love to eat. In that particular case, in their manufacturing lines, they’ve got computer vision and data again from their OT environments.

that intercepts and says, am I filling the right amount of Skittles in every bag? Mostly for, again, quality control purposes, this data is being harnessed. And back to Tom, what you were saying, it is so exciting because it is providing the shop floor managers a little extra lease of life, extra eyes for themselves. And that is super, super powerful. So again, it’s super exciting in the way the world is going.

Our point of view is, again, a lot of firms are a little reactive to it. So if you think about industry for not zero and taking a lot of examples and even with Gen .ai within Accenture, we are driving close to 600 plus Gen .ai projects. And a lot of these Gen .ai projects, not all certainly, but a lot of them require data to be within the four walls of the firm. So I’ll give an example. If Mars wants to do chocolate mixing,

the recipe for the chocolate mixing and the steps and the processes, they are absolutely core IP to that business, right? And there’s a lot of hesitation in putting that kind of information on public cloud. So though you want to use now Gen .Eye and machine algorithms, there’s this emerging demand literally in the last six months where a lot of clients are saying, how we have been doing these POCs, maybe high performance compute and having that in our own four walls is the way we were thinking about it, but let’s pause.

Mili Iyengar (24:27.624)
because rather than do these one or two POCs, we are starting to look at this from a how will I enable scale adoption? And in that context, then getting an edge strategy or even start like people used to have a cloud strategy in the past, right? And how their adoption would be through maybe SaaS products and then data center exits. Like there was a very defined way of adoption. In the edge world, we haven’t seen that so much. And, you know, we are starting to see the pivot now in…

I need to have an edge strategy. I need to have a data strategy that interposes within the edge strategy. And, you know, we are seeing so many clients take that more strategic approach, which is, I think, going to be really game changing to move from POC to scale. That is super critical.

Tom White (25:12.226)
Yeah, yeah. Yeah, absolutely. It’s like you said at the start of the podcast, right? A lot of people talk about it, but actually executing it is the same as moving from proof of concept to scaling it out and to actually making it happen because sometimes there’s some amazing proof of concepts that happen, but it’s the business continuity and kind of togetherness to actually drive that through into scale.

Mili Iyengar (25:42.408)
And I think one thing which is a little bit of still the pitfalls that one has to be aware of, right? That the typical buyer or people who actually get the best value from edge computing and IOT use cases, POC is actually the business. Whereas traditionally all of the IT firms are still used to selling to the IT in enterprises. And so to try and actually get the business side of the house and the IT together on the same page.

be able to therefore then prove out the business value for that particular use case and how it’s going to impact business outcomes. I think that is super critical. And again, if you don’t start with an edge strategy, you don’t start with the use cases, you don’t start with the business case, you don’t start with the value that is going to drive into business, downstream adoption becomes extremely hard. So, you know, tech works.

in most scenarios of POC from a technical perspective, I’ve hardly seen failures, right? There’s tweaks that are required. It takes a little time, but eventually I’ve seen all POCs get there. Do they get there in a manner which is commercially viable for the firm to adopt? Is it at the right price point? Is there a business case for it to be able to survive and thrive and expand that solution? I think those are the pitfalls. So again, from an Accenture standpoint, I think we have an extremely unique advantage.

of being almost what I call soup to nuts kind of firm where we have strategy and consulting that very closely are geared to the different needs of different industries and we are able to work backwards a little bit and have a little bit of foresight into how these use cases and impact business. And then we also have the tech to be able to bring all of the technology that is required to play.

The one unique thing that many people don’t realize is we also have song. And Accenture Song is an entire company, if you think about it, or our business unit that is primarily looking at customer experience, employee experience, and is in the business of literally bringing together ramifications of how the business outcomes will then actually impact humans and the way they operate and behaviors and adoption. And…

Mili Iyengar (28:00.968)
That is a specialized area in itself. And then you bring tech. So these trifecta effect of bringing the three pods together, if you may say so, I think is like the unique differentiator or advantage that we are able to day in and day out bring to our clients. And I see that as a very powerful combo. One without the other in this trifecta is only going to take you so far. So again.

super excited and you know we feel we are in a good position to be able to make the impact we desire.

Tom White (28:35.01)
Absolutely, yeah. I mean, it helps that Accenture has the scale as well as its customers to scale to be able to do this. And I think that’s testament to some of the reports that Accenture has also put out because you can really look into the data and really look into the industry awareness and analysis of what people are looking to do. And I can touch on one of those if I can. So.

I think one of the reports that Accenture recently put out was that there was 70 % of executives of foreseeing edge computing revolutionizing their business models in the next three years. I mean, that really is quite a colossal statistic, isn’t it? Because three years is such a relatively short period of time, as you say, to go from proof of concept to actually out onto scale. I mean, it could take three years, actually. So for 70 % of people to be saying that is really something. So my question to you, Millie, is,

What strategies do you think that businesses should start adopting and quickly if they were to look to roll out Edge? And how could they go about doing that?

Mili Iyengar (29:38.554)
Again, another very great question from you Tom. We did actually a study and the practice commission did with an external firm. So we already have our own data points and we pulled 2100 executives C -suite primarily across the globe. And the data point that you saw 70 % of executives that comes directly from that market research. And there are some other very important and interesting facts that came about, right? In that study.

So one of it was, you know, many, many clients have already started dabbling into the world of edge computing and primarily that is through POC, some pocket and part of the company and firm, but very few firms, and I think it is less than 18 % are thinking of it as a strategic mission and have a strategy around it. And I think that is the huge difference. So if you put those two numbers together, 70 % of them think that they will go to scale.

However, only 18 % have a actual edge strategy to execute. And we believe that mismatch a little bit is a great opportunity area for us to then further coach our clients, recommend them to be able to take an approach where, again, unlike the haphazard adoption sometimes from a cloud -first perspective that they took.

they pause and take a slightly more integrated approach. So to your question, how can any firm think about edge computing and what will help them set themselves up for scale and success over the next three years or a short relative horizon? I think the first and foremost thing is to be connected in their approaches. And what I mean by that is very often, even as of today, I don’t see business strategy.

an IT strategy getting amalgamated into one single strategy. There’s still two different playbooks that both sides of the firms operate under. And I’ve seen that when it comes to edge computing and IoT, because these are such customer facing or employee facing technologies and the power to transform is literally at that, you know, digital part of the ecosystem. Having very early on right from your C -suite and under, right?

Mili Iyengar (32:03.016)
take an approach where business and IT is going to come together and have strategies that are amalgamated. Not so much from, you know, tactically we want to do X, Y, Z, but strategically this is how we are going to leverage these emerging technologies is super exciting. Firms that do not have a role for chief innovation officer. I almost feel like that is a little bit of a bottleneck for them because these are roles that are orchestration roles in any firm.

And when you don’t have them, chances are that the operating model of the firm is not going to help and support that amalgamation into a single thread. So you call it chief innovation officer or transfer. It doesn’t matter. The designation doesn’t matter. My point is there needs to be somebody at a very high leadership executive level from an org chart itself who is enabling an operating model of collaboration. And we find that some of these basics, if the

if the enterprises start taking that into account, is going to set them up for future success. The second piece is almost every customer now has a hardware budget, refresh budget. This typically sits with IT. And every firm almost that I know, Fortune 500, has at least a primary cloud provider already selected. Many are multi -cloud, as we all know.

So to be able to go back to your primary cloud provider as well and ask them, and maybe just start from there, right? How can I stretch the power of what I get in public cloud all the way into the edge? And then tie that back with what you already have as your current state or brownfield from a OEM perspective. So, you know, HPs, Dells, Lenovo, Supermicros of the world, and pull these two together.

as a coherent strategy, which means you have a little bit of budget already for hardware refresh and you are able to get into a flywheel of using that and leveraging that to move away from CapexPen to maybe OPEX, which is what the cloud providers know pretty well. Then I think one can pretty much lay down that foundation and get a super strategic approach in pretty much from the get -go.

Mili Iyengar (34:21.256)
And then you’ll not believe it, but some of our clients and I have a particular global client in life sciences right now that is absolutely pushing the boundaries of, you know, computing as a whole. And especially when it comes to edge computing, the desire is I want to be zero infra company in my own four walls. I don’t even want routers. I don’t want to have, you know, any, any server at all. And I’ve heard that from retail as well, many grocers.

What if I don’t have any servers at all in my own shop? How can I still do compute and provide this super low latency solutions back to my business? And I think these firms or the way they are thinking about being cloud first to a point where they’re thinking about getting rid of their edge in some sense is also going to push the boundaries. So in that same example, if we stay…

Maybe there is an evolving solution and they’re pushing the boundaries on it, where you will have if there are 50 plants across the globe for this particular manufacturing firm, it will be for this pharma company, maybe 30, 40 different local zones that are still going to be close enough to their factories to provide them the latency that they require.

but not necessarily the traditional cloud model that we know right now where any firm has three or four or maybe five or six landing zones across the globe. We are almost talking distributed edge local zone type of architecture that could run into a scale of 30, 40, 50 data centers. Very similar to content delivery networks, but obviously they’re very different than that, right? Because you don’t have the same kind of workload. Like if you think about a Netflix and, you know, Uber’s of the world, like,

Tom White (36:03.49)

Mili Iyengar (36:10.216)
The CDNs are exactly the same kind of files done multiple, multiple times, or YouTube, for example, right? But these new age workloads are not the same kind, they’re different kind. If it’s a grocer, it’s going to be a pause, inventory management, logistics, that different workloads that require different things. So to think of them in this, you know, connected distributed network and not being the four walls then raises new issues.

which is the telcos have to really step up and provide super awesome last mile connectivity, right? Or it’s an opportunity for whoever out there to build that bridge for network availability and reliability, which is not there in the industry right now. So again, whoever’s going to be the first to solve some of these problems will have that first mover advantage and I think huge amount of leg up to their competition, both as the firm that is…

pushing the boundaries on these new solutions and at the same time as a services provider of those solutions. So super exciting, I think, world that we live in.

Tom White (37:20.61)
Yeah, it really is. I think it’s, you know, you can’t help but be inspired when talking to Milly about your passion for it. You’re clearly very passionate, very well -versed and knowledgeable and hearing essentially so many different avenues to market and the sheer scale and size at which you do that is so impressive. My last question for you today on the podcast actually is around, well, not my very last, but one of my last questions.

is around really the future, Millie. So, you know, in terms of looking ahead, you know, no one could have predicted to share, you know, Russian and scale and size of Gen AI three, four years ago. So I wanted to ask in your opinion and that of Essentials, what emerging trends do you foresee happening in the future, shaping IoT driven business models? And how are Essentials going to try and capitalize on this?

Mili Iyengar (38:18.248)
Again, another great question from you. Tom, we believe that, you know, Gen .EI is here to stay, right? And we are just calling it Gen .EI, but essentially the world of Gen .EI, as you know, is driven through large language models right now. And all the data in text form, I don’t think is enough to give us the kind of value and opportunity and accuracy of these models that we require.

So we particularly believe very strongly that JNAI is here to stay, data feeds into large language models is not going to be enough. And what is going to evolve therefore is training our models through video. And video can teach a model, unbridled amount of training can be done. So if video feeds become the…

input to training models, we are basically going to see IoT sensors and edge computing go to a totally different ball game altogether. Now the constraint that we have a little bit is there isn’t enough computing power right now in the world to be able to take the feeds from every single video input out there and make sense of it, but we have to eventually move there. I mean, our models will not be smart enough and we’ve seen already that with Llama 3.

The reasons why these models are better sometimes are not just how fast they can be or the computing power, but the sheer amount of training or input data that goes into training it and therefore the inference is better. So my point being again, we are not going to now unlock and rewind from where we began with Gen .AI and Advent. More and more video.

inputs will be required to train our models and make more sense of it. And in order to do that, we will require sensors and IoT and compute. We cannot take a video file and send it all the way to public cloud to train there. So I do see a scenario where in order to be most effective with compute, we will have to interperse that compute in multiple sections and paths. So as I was explaining right at the beginning of our top track, right? That…

Mili Iyengar (40:36.232)
there will be some compute within these cameras itself. So, you know, you discard the rest of the frame sets or only if there is a movement, you capture that entity and parse that to the next stage, which is the enterprise. And there you are taking, you know, 40 different camera feeds, say if you are a factory or a store and making sense of it. And then you are interposing and sending only that amount of relevant data upstream. I think that…

distributed way of computing at every stage is going to become super, super critical. Also from a perspective of just security and cyber security and getting the right firewalls in place for access of that data and XYZ, right? So I think that in itself is going to be a very, very big driver for IoT and edge computing. And again, in almost all businesses, latency,

security, data sovereignty, examples of where a firm wants to have its own LLM models or be very contextualized in how I have my operating or manufacturing lines and what videos can teach my particular scenario and inform them, I think is going to be crazy contextualized. And that will come from harnessing more and more the power of IoT and Edge, not in a generic sense, but…

in a very custom sense for the businesses that we talk about. So that is, I think, my personal point of view in what will fuel this industry over the next three, four years. Of course, the use cases and the power of that is going to be remarkable. But I do think the intrinsic piece of it will be the more we want to use Gen .ai and be contextual and bring out productivity, then the more we will require edge computing in the firm’s own four walls.

contextualize that. So, you know, those are kind of my two cents of how we will see the industry moving.

Tom White (42:37.762)
Fantastic insights. Yeah, really, really quite on point, as I would say, actually, because that’s what a lot of people are saying at the moment and a lot of people that I have on my podcast are saying similar things. So that’s how I know that you’re very much on the money with that. I think that’s really, really important. But it’d be great to see it. It’d be great to kind of see in the future of Centres, growth in these areas for what you’re talking about.

I think that’s one of the beauties about having a podcast is because we’ll say it today and then we can look back in a year or two years time and actually see how you know things have emerged and hopefully we don’t have another J -A -I curveball that comes in and really just disrupts everything for the good because you know that would really set us off our course but thank you so much Millie for all your insights today and for learning more about you, your role at Accenture. It’s been really really great having you on to the show.

And as I come to near the end of the podcast today, we always ask a series of questions a bit more lighthearted to our guests. And I’ve got a couple for you, Millie. So my main question is, what challenges do you face personally that there isn’t a tech solution for at the moment? And you really would like to see that change. Do you have anything in mind?

Mili Iyengar (43:55.396)
So many, right? But just super simple, I think in many solutions, storage is becoming extremely important. Like where do you store the data? In many simple examples, I think battery power for these devices and longevity. So I have clients who want to use drones, but the battery life for those drones is just 20 minutes. And so how do you expand that? The same with EV charging and unlocking so many other.

Tom White (43:58.818)

Mili Iyengar (44:24.84)
models and also from a sustainability angle. So I do feel the, not so much the personal challenge, but professional challenges. Can we solve for some of these technologies? And I’m always, always counting for solutions out there from startups or even mature firms in this area of battery power and storage. And how can we…

keep data at rest and how we are going to leverage that. So that to me personally is a super exciting area to watch out for. The other area that I’m again personally watching out for is what is called TinyML. So neuromorphic computing, you know, if you think about it, again, my understanding is even speaking to Stanford professors and MIT professors that we are at the peak of what density we can pack with transistors on a chip.

Tom White (45:01.826)

Mili Iyengar (45:18.504)
Like that is at the peak and Zenith and, you know, Moore’s law is already kicked in. So the next iteration of what we require will be custom silicon. So specific purpose, you know, chipsets and silicon. And in that regard Accenture has already made an acquisition. So we have a, you know, edge computing practice that also has engineers who can actually design chipsets because we see that as an emerging area. But back to my point, you know, TinyML, how can we do?

machine learning with less than one watt power within these devices and chips. I think that’s super exciting as well. Neuromorphic computing is super important therefore. And again, the more sensors are going to be across the world, the sustainability impact of it as well. So it’s great to say, yeah, we’ll do all this data harnessing and run it through models, but how can we do it in the most sustainable manner and not…

basically have more than, you know, 1 % of electricity use across the globe right now going to data centers and flip that over and do it in a responsible manner. Right. I think to me, those are extremely interesting. And because I’m a mom at the end of it, I always think about data privacy and how some of these feeds are going to play out. So if you think about it, you know, data is going to live forever. It’s not going to really get quashed.

Tom White (46:17.538)

Tom White (46:21.442)

Mili Iyengar (46:40.808)
what we say today may literally come back in 20, 30, 40 years later to, I would say, come back in haunt. I’m an optimist and I believe in the good use and power of technology, but I think we have to be extremely intentional. And this applies to every single human being, right? That the digital footprint that you create is going to talk a lot about who you are and it’s going to shape us and, you know.

that is going to be the humans that we evolve into. So for the younger generation, certainly, like I said, as a mom for my child and his friends and others in Gen Alpha and Z, I always think about their core understanding of the digital footprint that they leave behind or they’ve started creating, almost like your credit history, like you have to start thinking about some of these as well. And how can we educate and sensitize across the globe, you know, not just in pockets.

about some of these basics, so you’re not taken advantage of, or you know how to react in certain situations, I think is going to be super critical as well. So, you know, I’m always kind of watching out for those things.

Tom White (47:53.57)
Yeah, it seems, yeah, crikey. I think if you’re interested in TinyML, we had Evgeny Gusev on from the TinyML Foundation on our podcast last year, and we’ve got an upcoming one as well. So we’re big fans of TinyML and what it’s gonna do. Last couple of questions. So name a book, movie or song that sparks your inspiration, Millie.

Mili Iyengar (48:10.024)

Mili Iyengar (48:17.)
it’s neither of those, but it is an app. So I love, love, love the masterclass app. And frankly, I like it because, you know, thankfully they don’t have an IT stream yet there, but everything else. I do think that I personally make a lot of sense of the world when I engage myself with just refreshing into areas that are…

Tom White (48:21.954)

okay, I know it.

Mili Iyengar (48:41.544)
totally outside of my core day -to -day job, which is an IT and tech and business outcomes, to hear artists train themselves and sports personalities train themselves. And I love the episode from Coach K on how he actually built a team and what his leadership fundamentals were. So I absolutely adore and I have a lot of time for masterclass. And then I’m also a lot into podcast.

So some of my favorite ones that I love to listen to is acquired. It just talks about multiple companies and their evolution and their stories. So it’s very insightful. I also like a podcast that is masters of scale. I think especially that is very relevant to core job right now as well. So example as edge computing and IOT goes through scaling, what are going to be the challenges and.

They talk about a little bit, I mean, just they invite so many people who’ve done it before me. So I feel like I’m learning from the masters and stalwarts and some very basic insights of edge computing is still very ephemeral. So when you do a demo that people can actually see versus putting PPT slides are some of the secret sources that I’m learning from these podcasts. And I love anything with Adam Grant.

Tom White (49:49.218)

Tom White (49:58.146)
Yeah, yeah.

Mili Iyengar (50:01.96)
Scott, Simon, Seminick. So I do feel some of those constant reminders in being better human beings and, you know, putting a lot of love out there is very critical and keeps you grounded. So those are my other favorites.

Tom White (50:02.114)

Tom White (50:15.906)
Yeah. Yeah, I listened to the Masters of Scale podcast. It’s Reid Hoffman’s, I think, isn’t it? And if I give you a recommendation, it would be Damian Hughes podcast. He does it with another chap. His name escapes me, even though I’m going to kick myself when I finish. But it’s called the High Performance Podcast. And the High Performance Podcast, I think you’d really, really enjoy because it’s…

Mili Iyengar (50:23.652)
Yep, that is correct.

Mili Iyengar (50:40.264)
Tom White (50:44.738)
It’s similar to Masterclass in the sense that it has a lot of sports professionals and business coaches and people about how to get better at what you do, right? And it’s a UK -based podcast, but very, very popular and has a lot of international guests on it.

Mili Iyengar (50:49.032)

Mili Iyengar (50:57.64)
Wonderful. I’ll definitely check that out.

Tom White (51:02.21)
great. Well Millie, thank you so much for coming on to the IMT podcast today. It’s been great having you. I’ve really, really enjoyed it. I’m sure we could talk for hours, but I think we’ve got enough great content there for the show. So thank you so much.

Mili Iyengar (51:17.032)
Thanks so much again, Tom, for having me. Always a pleasure. Cheers. Bye -bye. Have a great day.

Tom White (51:21.954)
Thank you.


About our guest

Mili Iyengar is Global GTM Leader – EDGE Computing & IoT | Managing Director at Accenture. She is a seasoned technology leader with over 19 years of experience in IT Outsourcing & Cloud Transformation Sales. Currently focused on Intelligent Cloud EDGE Computing & IoT solutions, she advises C-Suite and client leaders on leveraging emerging technologies for competitive advantage. Mili excels in consultative sales leadership, crafting Go-To-Market strategies, and building strong client relationships. Her deep tech insights drive tangible business value.

Connect with our guest