theCUBE + NYSE Wired: Practitioners Series | Vikalp Tandon, Accenture Song

theCUBE.net

Clips
More from theCUBE + NYSE Wired: Practitioners Series

Vikalp Tandon

Managing Director, Commerce Lead

Accenture Song

play_circle_outline Cloud-native, hybrid cloud, AI factories, tokens enabling new architectures and value.

play_circle_outline Importance of digital core and machine-readable data for AI enablement.

play_circle_outline Taming AI Sprawl: Governance, Compliance, and Focused Friction-Point Selection

play_circle_outline Acceleration from generative AI: faster research, prototyping, synthetic humans for testing.

Info
Transcript

Vikalp Tandon, Accenture Song

Vikalp Tandon

Managing Director, Commerce Lead Accenture Song

In this interview from the theCUBE + NYSE Wired Practitioner Series, Vikalp Tandon, managing director of commerce and commerce lead at Accenture Song, joins theCUBE's John Furrier to discuss how agentic AI is moving enterprise value creation beyond cost reduction and into net-new revenue growth. Tandon frames the shift around a clear progression — from classic pattern-recognition AI to generative AI to today's agentic systems, which build deeper memory and operate autonomously. He argues that success starts with a composable digital core and machine-readable ... Read more

explore Keep Exploring

What are the main trends and strategic questions enterprises are facing right now regarding cloud-native infrastructure, AI agents and tokenization, and whether to buy, build, or use a new type of cloud provider? add

What fundamental steps should organizations take to prepare for and implement agentic (autonomous) AI? add

How are practitioners and organizations approaching AI adoption—especially agentic AI and containerization—from a compliance and governance perspective; are they building AI-native solutions or simply bolting on AI, and what priorities and use cases are customers focusing on? add

How are agentic and generative AI being used in client work? add

bolt Powered by CUBE AI

Vikalp Tandon, Accenture Song

search

>> Welcome back everyone to theCUBE here at our NYSE studio. Of course, we have our Palo Alto studio connecting Silicon Valley and Wall Street. This is our practitioner series, part of our NYSE wired program, it's a CUBE original, a new brand for theCUBE where the intersection of capital markets and technology coming together. Technology is the market. Our next guest is from Accenture. He's going to talk a lot about agents, how work's changing. Vik Tan is the managing director of commerce and commerce lead at Accenture. Vic, great to have you on the cube.

>> Yeah, likewise. Delighted to be here.

>> Thanks for coming on. Accenture, well known for helping companies obviously transition, but we haven't seen a transition like this. This is not a transition. It's a Grand canyon of value creation.

>> And indeed it is.

>> It's going to be amazing. All these forces coming together. You got the maturization of cloud native, which is essentially distributed hybrid cloud, core, cloud, edge, AI factories. Coming online, we saw that storyline at NVIDIA again for the third year, but now you're starting to see the value of tokens. Tokens equals money, tokens equals work. Agent hype is at an all time high. You're starting to see evidence of how to execute agents if you've done gen AI right. And for people who get it right, the value creation and the value harvest is there. However, it's happening super-fast. People's heads are spinning. The world and the enterprise is like, do I buy? Do I build? Do I use a neocloud? Do I use AWS? What do I do? So this is kind of the number one topic. You guys are there figuring it out for customers. What's your take?

>> I think we start with the fundamentals. The fundamentals are, do you have the digital core ready? It is composable. Do you have your data sets organized in a way that they are machine-readable? And then how do you unlock your digital core and your data through the agentic layer? And I mean, we are talking about agentic, but the topic of AI is not new. Many, many organization, many, many use cases on classic AI where we look for patterns and use those patterns to find answers to questions. Then came agentic. So then sorry, then came generative where we are able to find meaning within those patterns and essentially translate that into language that humans could understand. And then came agentic, which is primarily autonomous, build deeper memory on you. So taking advantage of all these progressions of AI, but underlying platform and data are the core that most of the organizations needs to focus on while enabling use cases that remove some of the friction points with agentic.

>> Yeah. And people who get that right are reaping benefits. Not just cost savings, revenue.

>> Revenue.

>> I mean, we've heard that cost saving story before, but now we're seeing, yeah, knock down some cost reduction projects, no problem. People shift to another job. That argument of jobs going away is true, but not true. But the revenue piece, and I want to get your thoughts on this because you're the commerce lead. This past January at CES and then at NRF here in New York, at the retail show here, we interviewed dozens and dozens of people. And the common thread was the people who got agents right last year did the work, they grinded out the compliance, they did certain things, they were seeing revenue growth. Meanwhile, some people who didn't get it right had sprawl, too many projects, very fragmented efforts. What's your reaction to that? Because there seems to be a pattern around in the commerce side and everywhere else.

>> I mean, just think about it, right? If we just take one industry and then we can transport that into other industries as well. If you look at just retail, the basic math starts with unit economics. If you look at conversion, if you look at average or a value, retention rate, there are a lot of these KPIs. Now, there traditionally has been a lot of friction in each of those. I mean, if you look at traditional conversion rates, they're in the range of 1 to 2%. The cost of customer acquisition is very, very high. Now what agentic does is that it starts to drive the level of personalization, the level of understanding, combining a lot of diverse data sets into driving a meaningful intent signal, you will start to change the unit economics. When you change the unit economics, think about if your conversion rate goes from 1 to 2% to 5 to 6%, right? If you look at the payments, the payment rejection rate because of fraud and other errors goes from 80% payment acceptance to 95 or 97% payment acceptance. These are net new revenue areas that has never been explored or fine-tuned before. So it's not a fine-tuning of an existing approach, it's a new approach altogether. And we feel because of this unit economics, we feel that would have a massive impact on revenue generation, on customer engagement, on retention.

>> What are you seeing from practitioners that are out there doing it? Because this compliance and governance for sure, you're seeing new techniques with agents, with containers. I talked with Boomi about that, the CEO, Steve Lucas. There's been a liberation on the dev side, not just coding, having coding assistance and auto coding, generative AI coding, the velocity of the innovation, if they get it right. So is there an AI native playbook or can you bolt on AI? What's the thought on that? What are you hearing from your customers? Where are their heads at on this?

>> Yeah. I mean, I think about it. This is a top conversation. Front-runners and laggards alike. Everyone is intrigued about it just because there's so much of noise around agentic and just adding AI to your profile means that you already are much more substantiated from that perspective. So there's a lot of hype from that perspective. But I think the successful businesses are doing few things very well. They are identifying areas of friction, right? And I'll continue on the retail team. They supplier friction. As you onboard the suppliers on your retail platform, there is friction with the consumers, the way that they essentially shop with you, how you service your consumers, how you remember their preferences and deliver it. There's friction with fraud and risk. Now the successful organizations are selecting those friction areas, figuring it out, what is autonomous and if it is autonomous, can it be agentified? And solving those friction points, every friction point is a step ahead. Now the value of agentic is just because we are able to build these faster, deploy these faster, the value realization hence also become faster from that perspective.

>> You guys track the piece of the agent. I know the three Ws that you mentioned before we came on camera, but the key that I like about agents is that they're doing work.

>> They're doing work.

>> They're doing work and tokens equals more work, but there's still blind spots. Agents are only as good as the data that they can see.

>> Right.

>> So talk about your philosophy at Accenture on the three Ws with work around work.

>> Right. So let me explain that. So when we think about this reinvention through agentic, we think in three steps. What is the work that could be translated into autonomous agents? What is a workbench? Which is the platform in which these agents will run, interact, orchestrate. And the workforce that is necessary or transformed workforce rather that is necessary to make the best use of these agents. Ultimately, we are not thinking that these agents will replace humans. We are thinking that these agents will make humans more productive, more efficient. So they're able to scale. And to your earlier question, the unit economics would massively transform. So if anyone doing right could take the same number of humans and do 10 times more, that's a 10X impact. If you have your work, workforce, work bench connected in the right way.

>> Is there a playbook that's successful that you see?

>> I mean, there is a playbook, right? I mean, the playbook is for you to think about how you reach to the... I mean, if you think about four different stages, one stage is to automate the work, right? The second stage is to actually the agents do most of the work, but humans are in a pretty significant role, right? The other way around is that agents are taking more decisions and human are just supervising it. And then final is that there are certain tasks that you identify on enterprise in which the humans doesn't need to get involved. Agents pretty much do it within the guardrails as provided. If you are able to set your roadmap in this way, step one, step two, step two, step three, step four, we feel that the progression would be well-defined. You will be able to calculate the risk that you are taking. You will be able to communicate the risk to your stakeholders in that particular way, and you'll be able to drive in that particular direction.

>> What's it like at Accenture? Because Accenture is a classic service business. We saw with cloud, you got a lot more operating leverage with the cloud. I'm imagining AI is a force multiplier-

>> It is a force multiplier....

>> for the leverage of service that you guys provide. And in fact, I would say that generally speaking, some of the investments I see in the capital market side is in service industries. Which you go back 15 years ago, no VC would invest in the service business unless you were really scaling up linear scale. Accenture is a great case study of that actually. But now that you have operating leverage with the domain expertise, both in the system and in the humans, talk about that dynamic and what that means for customers, or for Accenture and for your customers.

>> Right. I mean, think about some of the traditional problems that have not been solved before, right? Let's take many, many banks have a huge amount of legacy platforms that are still running in Cobalt, right? And up until now, there wasn't a capital investment that allowed you to modernize those platforms. They were working, but they were also limiting your ability to diversify, launch new products faster, create new experiences. But what agentic is doing, which is a playbook here, what agentic is doing is to allow you to transform those platforms and experiences in a manageable cost, which unlocks value for us because we have been trying to pitch to many of the banks, "Hey, let's get out of this platform." But there wasn't any ROI on that one. Now, because of the cost of transforming it, there is a much better ROI and creates a humongous amount of opportunities in the services business. That is unlocking for us.

>> Talk about what you're working on. Take us through a day in the life. I'm your Copilot. I'm an AI agent. What am I working on with you? Take us through some of your engagements with customers.

>> That's a very intriguing question. So think about this one, right? I mean, the first is deep research, right? So as we are trying to solve problems for our clients, it used to take a lot of time for us to do research. That research has been completely fast tracked. You can assemble a lot of very profound signals from vast areas of data that the clients can give it to you, and then you can find issues that you essentially can solve for. So that is a number one area where we are applying agentic and essentially working through a Copilot like scenario. The second one is to reimagine experiences, right? And in order for us to do more, show less tell, the idea is for us to rapidly build prototypes, showcase and test those prototypes, and in most of the cases, test those prototypes with synthetic humans, right? So that you can rapidly score and experience and therefore you will deploy it. That's another big use case of agentic and Gen AI where we're using the power of it so our conversations are moving much faster. Now the third one is, as you are convincing the clients that, hey, this is the right place for you to invest, you are building the business case for it, you're also able to deploy systems and production much faster and help them scale much faster too, right? So from concept to production, that life cycle, we are able to significantly shrink it, and that continues as an everyday task for us that we meet the clients, we talk to them, we do deep research and essentially activate those experiences in a very short order.

>> What are you focused on now? What's your goals for the year? What's your focus?

>> Well, the focus is primarily think about building capabilities that we could rapidly deploy in the market, right? So if you look at the big shift that we are doing as a services business is that we often only led with consulting or our SI and operational capabilities. What we want to lead with is agents essentially that we can monetize and all our combined knowledge that we have in the enterprise sits with those agents. So that way, think about a merchandiser. It's a very complex job. If we can codify that steps of merchandising, steps of content production, steps of marketing and performance optimization into an agent, we can easily monetize that much more rapidly. So we feel that will be the biggest stream of revenue that we could generate out of this experience and also accelerate solutions for our clients.

>> Well, Vik, great to have you on theCUBE. Again, thanks for contributing to our practitioner series. We'll have to check in with you. Again, we love the commerce angle because retail is again, healthcare and retail, two big verticals, commerce and certainly shifted. Like I said, the metrics are there and they like analytics. So they're-

>> Yeah, ....

>> perfectly poised to lead the transformation.

>> Yeah. No, the parting thought that I have is it's a phenomenal time for us in the services industry. What we see is a continued momentum in this particular space. What we also see that certain objectives that were difficult to accomplish because of lack of capital will get the investment for us to move much faster. And we also see that we will be able to create unparalleled value in terms of unit economics for our clients that they could realize much sooner than they could before.

>> And you can bring AI intelligence to their systems. Again, this is the focus of the practitioner series, I'm John Furrier, your host of theCUBE, doing our part here to bring as much data to you. Thanks for watching.

>> Thank you.