HFS Horizons Summit | NYC—The Evolution of the Autonomous Enterprise: Your Horizon 3 Innovation Agenda

Top three takeaways

• Ted Shelton frames a “zero-basing labor” method: redesign each end-to-end workflow as if it required zero human labor, then reintroduce people only where creativity, critical thinking, or interpersonal relationships genuinely add value.
• Armen R. Kherlopian argues that machine perception — computer vision, audition, and even olfaction — is reshaping business processes, letting enterprises triage work so a human’s time is spent on the highest-value customers and risks.
• Nitin Rakesh contends the biggest near-term prize is clearing 50 years of tech debt — for example, turning ticket-driven “break-fix” IT support into predictive, self-healing autonomous processes using pattern recognition over logs and RCA data.
• The panel debates autonomy versus augmentation and the operating-model shift it demands, agreeing that cultural and organizational change — not technology — is the hardest barrier to the autonomous enterprise.
• Angela Johnson describes driving adoption in a “herd-mentality” real estate industry through concrete wins — computer-vision project scoping and decarbonization — tied directly to NOI, IRR, and executive incentives.

Tom Reuner — HFS Research[00:03]

Over the last two days, you heard a lot about the building blocks of the autonomous enterprise. You heard even more about generative AI. But how can we aggregate that into a Horizon 3 innovation agenda, or think of other ways we can crystallize some of that into practical terms? What can organizations listening to us take away if they go back to their offices? Is it literally just analyst fluff when you’re talking about an autonomous enterprise, or what are the practical terms of that? And I think the important part of that is also cutting through all the marketing noise. As a reminder, yesterday Ralph reminded us that about 70% of organizations would describe themselves as automation beginners. We know many organizations still struggle to scale automation. Similarly with cloud transformation — we’re doing a big piece of research — there are probably two sorts of organizations, and some of them don’t fully hit the strategic objectives they initially were trying to achieve. So if you think about all these things, even the more mundane stuff like automation and cloud, let alone the new funky stuff like generative AI, how real is that? What can we take away? I think we need more context for that. We talk about the autonomous enterprise; that’s not the answer to all the burning questions, but it’s a conduit to drive those discussions. Almost the only thing to keep in mind to set the stage for today: in the end, the framework of the autonomous enterprise is all about the intersection of data and AI, but there’s also a world where technology, bots, and algorithms are the first point of interaction. It’s much deeper than front-office stuff; it’s predictive, it’s really reacting to things. If you think about things like cloud-native operations, it’s such a new complexity, and we have to be honest about that, not pretending cloud is simplification. Humans can’t deal with the complexity anymore, so we have to think differently. But perhaps the most important point is not just talking about technologies — how can we make it tangible? What are the operating models? What we’re seeing time and time again, especially on operating models — a good example is cloud — is that people just jump in and try to retrofit all the wonderful innovations, but you set up for failure because we have to rethink and reimagine many things. And that’s what we’re trying to discuss. The good thing is it’s complex, and some of my brain is mushy after two days, but I have a brilliant panel who can help me with that. So, thinking about designing the innovation agenda for the enterprise — and by all means push back. If you think it’s just typical analyst speak, it’s fluff, be frank about it. It’s all about sharing experience. Perhaps not going a full round of introductions — Ted, can we start with you? How are you trying to set the innovation agenda, and how do you try to help that? Does it resonate with you, what you heard with us? How are you driving the agenda with your clients?

Ted Shelton — Partner, Bain & Company[03:25]

Sure, Tom. Ted Shelton, I’m a partner with Bain, and I’m the global product leader for business process redesign. They asked me to take that role because of my technology background, so I also focus on automation and artificial intelligence. The advice we are giving clients today is that you need to start by doing what we call zero-basing labor. Take your processes and say, what if there was zero labor in this whole end-to-end workflow? Imagine being able to do everything with technology. And then take a step back and say, well, there may be things that we want humans to do in these processes. We think there are three categories of work that we actually still want people to do: creativity, critical thinking, and interpersonal relationships. So look at those processes and say, where does critical thinking, creativity, or interpersonal relationships play a role? Those are the places we want people. Everywhere else, take the people out. Think about it, maybe not all in one step, because maybe the technology isn’t here to do everything in that process today, but your journey will be eradicating all the rest of those tasks. And you need to put an organization in place that has that frame of mind, so they’re redesigning all of the work in your organization around that principle.

Tom Reuner — HFS Research[05:07]

That’s good. Angela, you have tried a different lens on all of that — real estate. How do you try to drive the innovation agenda? What’s top of your mind from your very specific lens?

Angela Johnson — Partner, Fifth Wall[05:07]

Sure. Hi everyone. I’m Angela Johnson, a partner at Fifth Wall, and I lead our capital formation team. Fifth Wall is the largest venture capital firm focused on what I call the collision of the real estate industry and technology. What is really interesting about our model is that we are a traditional venture capital firm, but we take a very different approach to how we introduce technology to what I call probably the oldest industry colliding with the newest industry in the world. Of our $3.2 billion in assets under management, half of that actually comes from the real estate industry itself. So some of the largest owners, operators, and developers of real estate assets are actually limited partners in our funds. I spent 20 years in the traditional real estate finance world prior to coming to the venture community, and what I can say is that real estate is effectively an instantiated bond. It’s cash in, cash out; if it’s not broken, don’t fix it. It’s been really difficult to get this industry to adopt technology at all. It’s only in the last couple of years that people started paying rent online, so the adoption curve is incredibly steep in terms of what the industry needs to do to get anywhere near the autonomous enterprise. There is no ERP software solution for the real estate industry at all. You talk to real estate landlords, anyone in a real estate company, and they have no idea what enterprise SaaS is. I know that because I come from the industry. So what we do is we sit down with the real estate industry, we identify what their true pain points are, and then our venture investment team goes and finds the tech that solves real problems. That is the key to bringing this industry out of the Stone Age. Right now, the worst thing that possibly could have happened for the industry has happened with COVID. You artificially shut off demand, which is the single biggest factor outside of artificially inflating supply. So the industry is trying very hard in a hyperinflationary environment across the globe to create deflationary pressure on the OpEx line. That is where we’ve been focusing: how do you preserve NOI? How can you use data to reduce your operating expenses in a real estate building? That is how you get the attention of the industry. The other area we focus on is how do you create ancillary revenues — how do you differentiate your product offering with generative AI and other tech solutions? So that’s the approach we take: very small tidbits, go get early wins within the organizations, create your champions at the C-suite and board level, and then you have a captive audience to continue introducing tech solutions that matter to their businesses.

Tom Reuner — HFS Research[08:05]

Fascinating. Thanks for keeping us grounded — at times, especially, we’re somewhere up under the ceiling. Rajesh, as a service provider, you typically invest ahead of time; you’re ahead of many of your customers. What’s top of your mind? Is it along similar patterns and building blocks, or are you approaching it in a fundamentally different way?

Rajesh Rajappan — Senior Vice President, Hitachi Vantara[08:05]

No, thanks for that question. I’m Rajesh Rajappan, senior vice president at Hitachi Vantara, responsible for strategy and growth across all global markets. When we look at an innovation agenda at an enterprise, it is at the top of everyone’s mind. Now, what does that mean? It means different things to different enterprises in terms of where they are in the maturity curve and what they want to transform from a digital agenda standpoint. But this is a journey. This is not going to be a switch where one day I’m going to become autonomous. It’s going to be a journey, and it all starts with where I get the most return from an investment standpoint, designing the pieces accordingly and charting out the journey. And it all starts with a change in mindset from an enterprise perspective. What do I think about from an innovation standpoint, and accepting the fact that technology is going to be a co-pilot to a workforce? There is always a concern — the moment you talk about something autonomous, there is a social issue of, am I going to lose my job? So it is important to ensure that the workforce knows technology is going to be a co-pilot, an assistant, to make life easier so they can focus on much more value-added aspects of the enterprise.

Tom Reuner — HFS Research[09:42]

That’s great. And Armen, in one of your previous jobs you were chief science officer at Genpact, so I guess you had a significant part in designing innovation agendas. Of course, you have more and different jobs now. What’s top of your mind? What’s your experience?

Armen R. Kherlopian — Chief Science Officer, Covenant Venture Capital[10:00]

Thank you very much, Tom. Indeed, what comes to mind here in terms of designing the innovation agenda, as a key first step, is assessing what capabilities can be brought to bear now. As Chief Science Officer of Covenant Venture Capital, we think quite deeply on these topics my colleagues on the panel have been mentioning around what is the role of human labor and how people’s time should be prioritized. When we consider artificial intelligence in particular, there is at its core biology, supporting technologies such as computer vision. We have machines — and I use the term broadly — that can see. We have machines that can hear: computer audition. By extension, there’s even our sense of smell, computer olfaction, that exists. So the opportunities can be quite profound regarding business processes in terms of what it means to have machines as part of analyst teams, operations teams, and sales teams. Considering what can be brought to bear, all of a sudden you have a massive shift in time, as Ted mentioned on this point of critical thinking and creativity. A human then will spend more time reading email as opposed to writing. That shift happens because if you can generate the content faster and highlight to the machine what needs to be tracked, then all of a sudden the time a human does spend on something becomes much more impactful. So, to underscore this idea of human senses being implemented in machines and augmenting the human workforce.

Tom Reuner — HFS Research[12:19]

Fascinating. And Nitin, I remember not too long ago we had the same discussion on automation — we worked for different companies. Again, as a provider, is this the same pattern, slightly different, or what’s changed for you?

Nitin Rakesh — CEO, Mphasis[12:19]

Just to be sure, I’m with the same company still — it’s been six years, at Mphasis. I was just worried he’s going to go to touch after smell and vision, so I’m thinking, thank God computers can’t touch and feel yet, because that’s very much a part of human interaction that I’d like to retain. I’ll shake your hand before we leave. I have a slightly different take on all of this. I agree with everything that’s been said, primarily because in the end — just one statistic that stood out to me last week when I was reading yet another article on generative AI — 60% of the jobs that exist in the global economy today did not exist 50 years ago. 60% of what we do today we couldn’t have done 50 years ago. When I first went on my first job, they were just introducing personal computers. Before that, you had to actually have files and paper and pen. I can still write with paper and pen, but it’s getting harder and harder. So then you move the needle forward and talk about Horizon 3 innovation, as you call it in the segment. I think we have a long way to go before we can think of an autonomous enterprise, because we’re still fairly wedded to the way we did business over the last 50 years. The big revolution 50 years ago was that everything was introduced through computers. We had somebody writing a general ledger at a bank, and we created a GL system and a core banking platform. Somebody was doing underwriting under a tree in the large market in London where the insurance market was born, and then we turned that into an underwriting system, a policy admin system, and a claims system. We’re still kind of far away from getting away from just automating those processes to actually thinking about autonomous enterprises. I run an IT services company — there are probably many people here from the industry. I’ll give you one interesting example. Even today, when I talk to CIOs, it stands out how this thing happens. Has anybody had to reset a password anytime in the last 30 days? What happens when you can’t fix it through the process? You have to raise something called a ticket. You’re familiar with that word. You raise a ticket, somebody will attend to it, and somebody will figure out what to do. Now, you blow that up and say, OK, an application fails and goes down. The first thing the system administrator does is raise a ticket, because then somebody will get engaged, whether it’s level 0, level 1, level 2, level 3. Every day we do this. And we are such an innovative and creative world today, with autonomy and autonomous enterprises, that we call it break-fix. We still call it break-fix, and we charge for it, by the way. Why do we charge for it? Because people pay for it. Why do they pay for it? Because that’s the only way they’ve done it for the last 50 years. Until I see a ticket, I won’t actually activate, because I’m getting paid by the ticket. My simple answer to this problem is: you see an orange light in your car and you take it to the garage because something is wrong. Most reasonable people will. Now think about it — we have the last 10 years, 5 years, 6 months of data on all the tickets that have ever happened in an enterprise, because after we do the break-fix, we do something called the RCA report, root cause analysis: what did we do to fix it? But it’s just filed somewhere in some folder. Maybe it’s on the cloud today, but that doesn’t really make any difference. You take the dump of all of those tickets — every server log, every application log gives you data on an every-second basis — and you should be able to look at the patterns. Now you use machine learning to see what pattern caused what kind of failure in the past. Can I predict a failure before it happens? Can I put a probability on the likelihood of it happening? And more interestingly, can I actually do the intervention before it fails? Now what I’ve done is I’ve taken a very simple process that we all get paid for every day and turned it into an autonomous process where I don’t really need anybody raising a ticket. I don’t even need any intervention, because the infrastructure asset — switch, server, router — will actually fix itself automatically, because you’ve now got machine learning algorithms, basically pattern recognition, going into the logs, looking at the RCA and creating an intervention automatically. Now there will still be 20% of the time that you need a human intervention because a code-level change has to happen, a patch has to be updated, some API didn’t work, some integration failed — but that’s 20% of the time; we call them L2, L3 tickets. We need a software engineer to go in and do some code. So I think we are really running ahead of ourselves when it comes to looking at the autonomous enterprise, because the size of the opportunity in just fixing what we’ve built over the last 50 years and getting that to the next era is immense. The reason I call it the next era is that there was a fundamental shift in the last 5-7 years where every enterprise is now consuming technology on demand, not on premise. That has never happened before. So I think we are still dealing with the tech debt of the last 50 years. Yes, we have some very interesting use cases coming out of Horizon 2 and Horizon 3, and large language models will change a lot of what we do and probably create a whole new wave of creative, critical-thinking jobs, but the size of the prize is so large right now in just making sure the enterprises go where they need to go in terms of removing the debt of the last 50 years. And the last thing I’ll say — the biggest benefit really will not… I mean, I wrote a whole book on it, it’s called “Transformation in Times of Crisis,” little advertisement. Thanks, Phil. This is a really tough crowd, man — nothing is sticking here. Thank God I’m not a stand-up guy. If you lose sight of the end consumer, then it’s very hard to create anything autonomous. We have so much data on every consumer — think of a bank, a retailer. How much of that data is being used? I think the application of anything autonomous for an enterprise to be able to expand revenue and automate a business process has to start in front of the consumer. To me, the biggest benefit you will see coming out of things like generative AI, large language models, and all forms of automation should really be for you to be able to really know your customer. We understand KYC as a compliance term, but actually KYC is a business term: how well do you really know your customer? If you really know your customer, you can use all the power of these new tech, and that really will pivot the whole front-to-back mindset.

Tom Reuner — HFS Research[19:15]

No, thanks for that, but also building on that — if I listen to you guys, it’s not the obvious technology jumping out. What’s top of mind is almost the other way around: the way to drive automation is much more the capability to drive organizational change than the technology. Do we need both? What’s top of your mind, Rajesh, on your side? Are you literally more building all these old-fashioned scans of technologies, trying to figure out when and at which magnitude they might come on, or is it literally built on a meeting point of changing our behavior, following customer behavior? It’s much more about change than probably the innovation itself.

Rajesh Rajappan — Senior Vice President, Hitachi Vantara[19:43]

Technology is going to be an enabler. It all starts with what you want to do and what your customer is going to expect, and how you service them well. It is about how you do things faster, how you respond to your customers faster with fewer resources and generate better outcomes. That requires, one, a change in mindset in terms of how you start looking at it, and technology is only going to be an enabler for that. There are going to be a lot of technologies that evolve over a period of time. We’re just seeing the evolution of large language models, and you’re going to see a lot more of that coming in. It’s not going to be one technology; it’s going to be bringing all of these together. But at the end of the day, it all needs to start with: what is it that you want to do for your customer?

Tom Reuner — HFS Research[20:51]

Just coming back — again, beyond the unavoidable focus on ChatGPT and generative AI — which technology should organizations really look out for? And since we have two venture capitalists, I think that’s almost natural. Angela, what’s top of your mind again? Is it technology, or are you looking at it fundamentally differently, almost literally from the outcome, from your real estate focus?

Angela Johnson — Partner, Fifth Wall[21:16]

It’s a great question, because we struggle with this quite a bit. I agree with what was said here on both fronts in terms of the customer, because I don’t think you’re necessarily going to see the organizational change in real estate, just because a lot of times these innovation teams are usually one person who sits in a silo and doesn’t really have a lot of buy-in from the rest of the organization. I think when you have these use cases that actually enable you to understand your customer, or there’s a crisis moment — look at labor shortages, for instance. That’s when we’ve been able to leverage computer vision as a source to address labor shortages. A very concrete example is in the multi-family renovation space. A lot of the standard ways things have been done for the last couple of years — when you’re scoping out a project, it typically takes six months to go measure it out and have contractors walk through the building — but we invested in a company called Taylorbird, which uses computer vision. They scrape the internet and can get within one inch of accuracy compared to using human talent to scope out that project, and then they use a two-sided marketplace to bid out to the contractors. In that process, what our real estate customers are looking for is the 500 man-hour reduction, and they’re also looking at the 100 to 200 basis point improvement in IRR on that project. So we’ve taken this example out to multiple real estate companies who are active tech-denying CEOs — that is a term that many of them have used — and we show them: if you just use this solution, it actually makes your job much easier, it drives your return profile higher, and we’re introducing you to something we’re not calling computer vision. We’re calling it an enabler to help you achieve your effective targets for the end of the year so you can get your bonuses. I think that, at least in our minds, is finding that kind of technology that has broad scalability, where we can leverage our LPs, who are the real estate industry, to showcase that the technology actually works. It is a herd-mentality industry, so everyone wants to know what their competitors are using, and that then helps us drive alpha from the venture capital perspective, because we then have an ability to roll it out to all of our real estate customers. Effectively, from a venture perspective, you need to ensure that you can get to scale very quickly and be the first mover. So I think it’s a combination of both. That’s just one example from a labor shortage perspective. The other one we’re really focused on is decarbonization. I won’t go into all the specifics, but irrespective of your politics and where you stand on climate change, the reality is physical risk is a real thing, and we’ve leveraged different types of tech solutions to help real estate asset owners address this theme across the board, whether it be from the raw materials through the construction automation process, the operating carbon of the building, the downcycling and recycling. We’ve used tech in various ways to address all of these problems, and I think that is the real area where you’re going to see a major transfer of wealth, and something that’s actually going to be good so that we all have a planet in the foreseeable future.

Tom Reuner — HFS Research[24:04]

That would be helpful. The last bit — has anything changed for you since you moved from Genpact, driving the agenda there, to being in venture capital? Is this the same approach, the same thinking, or have you literally changed your priorities and your way of thinking?

Armen R. Kherlopian — Chief Science Officer, Covenant Venture Capital[24:31]

Sure, thank you, and I’ll pick up on Angela’s comments here. Firstly, this idea of the value of human labor. As investors, we go through this exercise: what would it be if every team member, every employee, every supplier staff member — every hour of their time — was precious? All of a sudden you start seeking solutions that triage, so that at a person’s highest energy they are interacting with the most valuable customers, and at the points of their focus cycle they are dealing with the highest risks. When you have that ability to triage and to focus a human’s time, these types of technologies — in contrast, dear colleagues, to this ticket view, where each ticket has different value, let alone the concept of a ticket — it’s about optimizing around maximal human performance. As we’re discussing here, there are different technologies that can accentuate that. So companies that have this type of focus from the enterprise side are very interesting to us. By extension, also on the customer side, not in terms of productivity but in terms of experience. Right here in New York, although there was massive uptake of consumer products such as Airbnb and Uber, it was not so in the beginning. You had decades-old frames of reference — do not go into strangers’ cars, do not visit strangers’ homes — and you could quite literally go into a stranger’s car to go to a stranger’s home. The accentuation there of, OK, there can be a better experience, and it could actually be safer due to rating systems — you’re not holding your hand out for a taxi or metal hurtling by. So again, startups that are facilitating these types of experiences, which are not immediately appreciated, are also of high interest. Picking up on this point of sustainability, with some colleagues we’ve conducted some analysis, true to form, on this idea of machine perception. We asked the question: what if we could read and analyze all the earnings calls of the 58,000-plus public companies on the planet as they emerge and as they were being issued? So you have computer audition — what’s the tone? — and natural language processing under that — what’s being said. And it’s quite striking: before 2020 Q3, almost no earnings calls mentioned sustainability-related topics. Now it’s spiking, but not everywhere. There are enterprises that are not talking about it. So this has massive implications in terms of how those organizations will get to the next era, how they’re servicing their customers, interacting with suppliers, and competing. So the sentiment across the industrial base, with key points of leverage being such valuable data sources, is something also to be on the lookout for.

Tom Reuner — HFS Research[28:13]

Great. Just building on your point — you have to work differently. For me, a key point, and it’s a tough one to get your head around, is how we should think about the operating model. It’s probably not one operating model, but I think yesterday Chetan from Amelia, previously IPsoft, was ganged up on by service provider clients over his prediction about digital workers — you’re slightly behind, that’s one way of looking at it. But if I look at it the other way around, an organization transforming to cloud needs new operating models, unless you’re just looking for efficiency gains. If you want to be product-centric, if you want the velocity and all the other good stuff that cloud should bring, you have to change your operating model. Ted, I know you’re quite skeptical of the term autonomous anyway — here’s your chance to let rip, that’s one thing — but also, how do you guide your clients on those issues around operating models? Or are we too early? Do you think that has to come much later? How’s the sequencing on that?

Ted Shelton — Partner, Bain & Company[29:12]

Yeah, so let me first respond to the skepticism about the word autonomous. It’s not because I disagree that many of the activities in the enterprise are going to be sort of lights-out, straight-through processing, but rather that if we start saying autonomous all the time, there isn’t room in that word for employees. I think employees are actually still going to be really important for a while — until Skynet and the Terminator show up and then we’re all done. But leaving that dark possibility aside, we’re still going to have companies that have employees, and I think we need to talk about the augmented workforce. We do need to think differently about our operating model, though. Why is it that we have the M-form organization that we have today? It’s because, in order to scale a large organization, we had to make communications very efficient between employees. To do that, we created pyramids, hierarchies, and we separated out functions. So a finance function has all the finance people working in it; wherever you work in the world, you have a local manager, a regional manager, and it goes all the way up to the CFO. Well, what does technology do? It reduces the cost of communication. It reduces the cost of coordination. It also reduces the number of people that have to be in that communication loop to effectively get the work done. And we can now start thinking differently. We can think around process, and we can think about how all the people within a process are matrixed into a process owner. So that operating model for how we structure work within organizations is going to fundamentally change because of these technologies. I think the interesting challenge is that the people change is going to be much harder than the technology change, and we talked a little bit about that throughout the last couple of days. But what we tell clients is that you will be on a journey for the next decade to arrive at a very different enterprise model, to run your business fundamentally differently than you do today — one where technology is much more central to the way you think about and run everything. So it’s not discontinuous with your autonomous message. I just think that by saying augmented, we say: oh, by the way, the people still are here. They’re being helped by the machines, not replaced or displaced by them.

Tom Reuner — HFS Research[31:33]

Nitin, it’s almost a segue — as a service provider, there’s a lot of internal change, and I think you have to be honest, even on automation on the supply side there is displacement. The jury is open on whatever happens with gen AI and all those technologies, but how do you drive that change, how do you take your teams on that journey?

Nitin Rakesh — CEO, Mphasis[31:44]

So, Tom, one very big value prop from our industry has always been creating net new talent in whatever is the next best thing since sliced bread — all the way from the days of the mainframes to cloud and now to AI. So for us, we always operate on the edge of extinction, because if we don’t change, we’re not going to have a business to run. I’ll give you an example. 10 years ago in this industry, a third of the revenue was all related to some form of testing — simple testing, unit testing, black box testing, regression testing, UAT, BUAT. We used to joke about it and call it XAT, any form of testing. Today that number is probably less than 5% which is standalone testing. So we had to reinvent ourselves, because we had to reskill those people to become full-stack engineers. Obviously, each company had a different business model of how to create those scaled pyramids of talent. You can’t really take a pure tester and turn him or her into a full-stack developer, but if you have the core engineering skills, then you can. So the reason we always think ahead — five or six years ago at Mphasis, we bet the whole company on cloud and cognitive, because that seemed to be the wave. Cloud’s been playing out quite nicely; cognitive is starting to play out quite nicely. We set up a quantum lab three years ago in Calgary and in India, primarily because we think it’s probably another two or three years away, but, as always, it’s been around for 10, 15 years and not really making an impact — but it’s starting to make an impact in areas like optimization. So that’s one example of a Horizon 3 bet we made, where we think having a little bit of a forward-leaning stance will help us create that capacity at scale as it’s needed. So today we’re able to create talent on demand primarily by reskilling rather than by hiring, because this talent doesn’t really exist. To me, the biggest change — every enterprise client we talk to, the change isn’t to do with skill set or talent or technology or the application of it or the use case of it; it’s almost always cultural. I think that’s the big elephant in the room, always, when we’re trying to sell a large deal. And the cultural change is multifold. It’s about using a third party; it’s about using technology in an area where somebody was doing it and saying, I was doing it better, if you automate it, it won’t be as good. I understand the pulse of the customer on the phone — but that doesn’t mean you don’t automate the contact center and the incoming call. I asked a simple question to many of our clients: define what your digital drop ratio is to me. And they say, what is that? So, how many times does a transaction — let’s take a bank, an insurance company, or a healthcare company — start on the app or on the website and end up in a call center? Those of you who use Amazon have probably never called Amazon. I probably called them once because I couldn’t fit a mattress back into the box I opened it from. In almost every other enterprise, the digital drop ratio can be as high as 80% which means 8 out of 10 times the transaction actually ends up in the call center — and I’m not talking about a balance inquiry, I’m talking about a real transaction. That is partly cultural, partly technological, but I think the issue of implementing the change very much starts with the mental model shift, and that can only happen top-down.

Tom Reuner — HFS Research[35:25]

Just staring at the time — it’s running away. I want to open it up to questions. Audience, any questions for the outstanding panel?

Armen R. Kherlopian — Chief Science Officer, Covenant Venture Capital[36:25]

Sure, I’ll immediately reply as an optimist, but as a scientist, hold on. Yes, thank you for the Matrix-style plug-in there, Ted. So to unpack a few of the examples: if we consider computer vision, there’s processing in our primary visual cortex that is layered. For example, if you see a loved one or an automobile, there are circles there, whether it’s the wheel or the face of someone you recognize, that’s processed in layers. Hearing is on frequency, almost like you unroll your cochlea — you have a piano in terms of frequency. Olfaction is even higher-dimensional; you can sense different compounds, which is very interesting at scale at different real estate sites for safety, as an example, or industrial security. What you’re alluding to is that this neural network structure around observation has been played the other way to also generate content, and indeed there is research that can decode, as an intercept, some of the connections around what we are thinking. It’s basically an extension of pattern matching — very interesting research in terms of functional MRI. So a lot of these biological insights are being implemented in the machine, increasing the processing speed and being distributed in networks, and you get all sorts of exciting examples for different business models. For example, Apple bought Shazam for half a billion for the simple audio recognition around what music is being played in a social setting. I believe in the enterprise and business processes there are many Shazam-style apps to be discovered and deployed, and when you do that, you start to blow apart things like what is a ticket or what’s needed. So that sensory capability, and imagining where it could be deployed across the industrial base, I find tremendously exciting. Neuralink is a good example of what you’re referring to, by the way.

Tom Reuner — HFS Research[38:28]

So next time he’s going to get Musk to come and talk to us. As I said at the beginning, I need an outstanding panel with the brain power. Patrick, do you want to run with a question?

Rajesh Rajappan — Senior Vice President, Hitachi Vantara[39:30]

Yeah, I think this is going to go back to what are the outcomes we’re going to be helping the enterprises with. We’ve talked about testing and how that has evolved. Pretty much all aspects of this are going to evolve. How does coding happen? Today, coding can be improved in terms of productivity using generative AI — you could generate the initial code with generative AI and then refine on it. What that means is more productivity, which means you could do more for less. I think similar aspects are going to be applied across different parts of services. Contact center is one where there are a lot more autonomous aspects that could be brought in, where human intervention would be needed for more high-end responses compared to what is happening today. I was having an interesting conversation with a CIO last week on something similar, and what I heard — and I think this has been said in public as well by the same person, but I won’t give their name — he says it’s not that AI will replace jobs or people; AI will replace people who don’t use AI. So I think you have to upskill yourself, you have to be able to use tools, and you have to be able to get more productive. Just because you can write code faster doesn’t mean we don’t need a programmer or a developer, because the number of lines of code being written is, by the way, growing exponentially. We need more developers; we can’t find enough. So it all depends on your current job. Sales can never be automated, let me tell you that. So if you’re in sales, you’re golden, as long as you meet your quota.

Tom Reuner — HFS Research[40:50]

Yeah, of course, great. Unfortunately, time is linear, not autonomous, so I’m staring at us. But a big thank you to my great panel, and give them a big round of applause.