AI sprawl is already happening because different parts of an enterprise are adopting AI independently. HR may use one platform, customer service another, while developers build agents on entirely different frameworks. As that environment grows, Raj Koneru, Founder and CEO, kore.ai emphasizes that enterprises need a way to manage it without first requiring every team to standardize on the same AI platform or framework. In an exclusive conversation, Koneru tells Rajneesh De, Group Editor, APAC Media & CXO Media how the agent management platform (AMP) from kore.ai provides that visibility and control while allowing individual teams to continue using the platforms and frameworks that make sense for their needs.
As an enterprise agentic AI platform, what solutions and services does Kore.ai offer for India and what are their use cases?
We majorly serve in two areas where AI is already changing daily operations, customer experience and employee experience.
On the customer side, we cover the service operation end to end. Voice and digital self-service, real-time agent assist, automated quality management, and the contact center itself. We also offer pre-built solutions for banking, retail, and healthcare that come with the workflows, integrations, and compliance requirements those industries demand. Particularly for India, our solutions support all major Indian languages and their dialects, specific channels like voice and whatsapp. We are strong in banking, insurance and retail sectors in India.
On the employee side, the problem is simple. Too much work still starts with hunting for information and jumping between systems to get something done. Our EX offering combines enterprise search with ready-to-use agents for IT, HR, and recruiting. They answer questions, handle requests, and complete tasks, and they always respect each employee’s own access permissions.
Underneath all of this sits Artemis, our agent platform, which customers use to build agents for their own processes. Our Agent Marketplace adds pre-built agents and tools they can adapt, so nobody starts a use case from scratch.
Today about 70% of our business comes from customer experience, 25% from employee experience, and 5% from long-running process automation. I expect that last slice to grow meaningfully over the next 12 to 18 months.
How does Kore.ai’s Artemis help enterprises scale AI, and what are the challenges in scaling?
Kore.ai agent platform – Artemis is first of its kind AI-programmable platform that helps enterprises to build and scale agentic systems from a few agents to hundreds of agents working across teams and business processes. At that point you cannot have every use case connecting to systems, setting controls, and measuring performance in its own way. You need one common approach that every new agent builds on, and Artemis provides that.
A few parts of the architecture become particularly important here. Agent Blueprint Language, or ABL, gives teams a structured way to define agents, workflows, tools, and policies. The platform uses what we call a dual-brain architecture: agentic reasoning where judgment is required, deterministic flows where the path is known. That keeps behavior predictable and helps control latency and cost. And Auto Loop continuously evaluates agents against quality, cost, and business outcomes, and tunes them.
The technology, however, is only part of the scaling challenge. Enterprises still deal with fragmented data, integration with legacy systems, security, and the economics of running large numbers of agents. There is also a prioritization problem. Building agents is easy now, so companies end up with sprawl instead of value. We push customers to start with use cases that show measurable impact on cost, speed, or revenue, and scale from there.
How do the Agent Management Platform help enterprises manage AI sprawl, especially when different functions use AI instances simultaneously?
AI sprawl is already happening because different parts of an enterprise are adopting AI independently. HR may use one platform, customer service another, while developers build agents on entirely different frameworks. As that environment grows, enterprises need a way to manage it without first requiring every team to standardize on the same AI platform or framework.
What we started with the agent management platform (AMP) has now become part of the broader Artemis platform, giving enterprises a vendor-agnostic way to manage agents across the enterprise, including those built outside Kore.ai. It gives enterprises a clear view of their agent environment and a consistent way to govern how those agents operate. They can track activity and performance, enforce policies, and understand where agents are being used and how they are behaving in production.
This becomes more important as agents begin working across functions and systems. Enterprises need to know which agent took an action, what tools or data it accessed, what it cost, and whether it operated within policy. AMP provides that visibility and control while allowing individual teams to continue using the platforms and frameworks that make sense for their needs.
What are the key pillars of Kore.ai’s GTM strategy in India?
India is our largest R&D and support center, with over 700 people building, deploying, and supporting our platform from India for the world. We also have a sizeable forward-deployed engineering team here that works directly with customers and partners on deployments.
Our go-to-market in India focuses on large enterprises, through both direct sales and partners. That means direct engagement with enterprise decision-makers, deeper relationships with hyperscalers like Microsoft and AWS, where we co-build and co-sell, and a growing network of system integrators and channel partners who help us support larger AI programs.
We already have a strong presence in banking, insurance, and retail, and we see clear room to grow across other sectors in India.
How does Kore.ai uniquely differentiate itself from other agentic AI solution providers?
We have been building enterprise AI for more than a decade, and that shows in the product. The hard parts of enterprise deployments were built into our technology from the start: security, integration, reliability, and governance. They were never an afterthought. Our innovations focus on delivering scalable outcomes in production.
Today we bring three things together. A platform for enterprises to build their own agents. Ready-to-deploy applications for customer and employee experience. And a management layer for agents built anywhere, on any framework.
The differentiation is in how those pieces work together. A customer can build on Artemis, use our pre-built applications where they want to move faster, and still govern agents built elsewhere. We are not asking anyone to rip out what they have or commit to one model, one cloud, or one ecosystem. Most large enterprises will end up with a mixed AI environment, and our job is to give them the flexibility to keep adopting AI while keeping control as that environment grows.
We have been recognized as a leader in Conversational and Agentic AI technology by third party analysts like Gartner, Forrester, Everest, IDC and others and over 300 global enterprises are trusting our technology for a decade.
How will you explain the functioning of AI for Process for automating enterprise workflows and optimizing knowledge-intensive operations?
Think about processes like opening a bank account, settling a claim, or handling a product return. They run for days or weeks, involve many steps and handoffs, and depend on people to interpret information, apply business context, and manage exceptions. That is exactly why they have been so hard to automate end to end.
AI for Process extends automation into that kind of knowledge-intensive work. AI agents draw on enterprise knowledge, reason about what needs to happen next, and take actions across systems. Deterministic workflows still handle the defined rules, validations, and approvals where execution has to be predictable. And people stay involved where their judgment is genuinely required.
The result is that enterprises no longer have to choose between forcing a messy process into a rigid workflow or leaving the hard parts entirely to people. More of the routine work gets automated, the process moves faster, and domain experts spend their time on the calls only they can make.
