How Agentic AI Expands What’s Possible in Business Applications
Insights
- Business applications are evolving from data and dashboards to intelligent, outcome-focused systems.
- Agentic AI enables a progression from Copilots to assisted agents and ultimately autonomous digital workers.
- The real value of AI comes from delivering outcomes, not just automating tasks.
Jared Spataro of Microsoft and Shishank Gupta of Infosys explore how the rise of agentic AI is redefining what business applications can do. From transforming static dashboards into insight-driven systems to introducing autonomous agents that act as digital workers, the conversation highlights a shift toward new economics, new operating models, and a stronger focus on business outcomes rather than applications themselves.
Christine Calhoun:
How has the rise of Agentic AI changed what's possible within business applications and what kinds of new capabilities are emerging as a result, Shishank?
Shishank Gupta:
Yeah, I think on this one, I go back to what Jared said, right? I mean, you spoke about two patterns. There's one which are out of the box Copilot capabilities that exist for our clients and our users to leverage from. And then there are obviously the asynchronous agents. You can leverage the asynchronous agents to go off, do a bunch of things. And that may or may not be on Microsoft technologies, get something done and assist the user in the things that they do. I do believe we are at a point where we have started this journey towards assisted agents to autonomous agents. I do believe you'll start to see them. We may start small, but eventually that is the goal. And that's where agentic AI will come in, is to see how we can have a large portion of the work that we do in an autonomous fashion while there will still be some that will be assisted and with human in the loop. So I do think that those will be the three things that will happen. You will see out of the box Copilots, you will see assisted Copilots, and eventually moving to autonomous agents, which is a combination of both the patterns that we spoke about. Now here I want to give you an example of what we're doing for one of our clients. It's a large telco in US and here we are trying to leverage the out of the box capability and also build autonomous agents for them to get a single view. They call it the single glass of pain, which is essentially their dashboards. They have it today. It's a dashboard, fairly static dashboard. It comes up with information for leadership to look at it. But then it is not intelligent enough to draw insights from it. People have to manually look through it and derive insights, but we are trying to derive insights using agentic AI and agents, wherein it just does not just give you a view of information, it gives you insights as to why it is changing, what could be the reasons why the data and trends are changing. In fact, our goal is to create analytics, wherein the agents can come back and do churn analysis, because this is on CRM data, of why the customers are moving, why certain trends they are seeing in certain products. And we also be able to trigger specific campaigns to change the trajectory that they are seeing. So that's really the vision of where we think we want to go. And I think that's going to be the future. We will have many such similar examples in other domains as well.
Christine Calhoun:
Jared, What are your thoughts?
Jared Spataro:
I think it's good to back up a little bit and look at what business applications have been about. Again, it's one of those underlying assumptions that, you know, this is what powers business, but business applications as they've developed over the last couple of decades are essentially data with forms as an interface on top of them, whether it's CRM or ERP or supply chain, whatever happens to be. The most important thing to focus on is outcomes. What you're after is getting jobs done. And those jobs, as you isolate them, can increasingly be done less by humans and more by agents. So I think in that way Shishank is exactly right. Increasingly the energy I think will go towards autonomous type of agents. Sometimes people are now calling them almost digital workers that you can assign jobs and you ask them to go get that done and to bring you back not a completed job, but an outcome that you're seeking. That's a pretty powerful new business construct. The only way we could previously do that was essentially assigning a person to go get the work done and we knew what headcount costs look like. So I think there's a new economics associated with these agents. There's a new operating model associated with them. And increasingly, we're going to be less worried about the app or the application or even the interface that it provides and much more focused on the outcome that we're trying to drive. And that's pretty exciting because that's a new type of business architect. It's one that we've wanted for a very long time, but we've kind of been constrained by what's been available to us.