‘These are like AI employees… they will impact job market, but not how we are thinking’: Indian AI agents startup founder

KOGO, the deep-tech AI company founded by Gopalakrishnan and Praveer Kochhar in 2020, has created an operating system that enables companies to develop AI agents, AI plugins, and AI tools to work seamlessly within business environments and enterprises. “AI agents now have cognitive ability, meaning they can reason and autonomously make decisions. So, it’s no longer just automation; it’s autonomy—the ability to make decisions and complete workflows independently,” Gopalakrishnan explains in an interview with indianexpress.com, distinguishing AI agents from traditional chatbots.

KOGO is meant for both small businesses and large enterprises. (Express Photo)

“A chatbot operates on an ‘if-then-else’ sequence and cannot make decisions if something falls outside its predefined responses. For example, if you ask, “Can you tell me the cancellation charges on my ticket with this PNR number for a flight to Bombay tomorrow morning?” a typical chatbot will likely be unable to answer that question. However, an AI agent can access the API, retrieve the exact details, and provide a complete answer, reasoning through the information just like a human. It might explain that if you cancel the booking at four o’clock tomorrow, a certain cancellation fee will apply, specific taxes will be deducted, and this is the amount you will receive. This is something a chatbot cannot do. It’s a very complex reasoning process,” he explains.

An AI agent has certain skills and abilities. It can gather data—whether structured, unstructured, from the internet, or visual sources. It can process all of this information, comprehend it, and generate insights. “It can understand human intent in multiple languages and interact across various interfaces, taking real-world actions. For example, it can post, pay, record, edit, and update—all automatically. A chatbot cannot do this.”

Gopalakrishnan’s company has developed a proprietary AI operating system—a framework that allows AI agents to be deployed, managed, and built directly on the OS. Think of it as an “AI Pass” for AI agents, plugins, and tools. Enterprises and SMEs can select an AI agent based on their specific needs and workflow, deploying it in minutes (in fact, under 10 minutes) without requiring technical expertise. “We are a horizontal stack, not a vertical stack, with a low-code, no-code platform,” Gopalakrishnan notes.

Different AI agents and their pricing on KOGO. (Express Photo)

KOGO is aiming to simplify how AI is built and implemented. Comparing Large Language Models (LLMs) to electricity and AI agents to appliances, he adds: “An enterprise cannot simply take intelligence from an LLM and plug it into their workflow or system. It needs AI agents… We are the ones who created the operating system for this. These agents also need to communicate with various platforms, access data, and interact with the world. We have built over 600 integrations into the platform.”

Gopalakrishnan explained that autonomous AI agents can be deployed on a website, an app, within a chatbot, or through voice-over IP. This means they can be used over telephony, in WhatsApp, on social media, and more. These agents support multiple languages and can operate across various interfaces. KOGO powers its AI agents with several AI models and has also trained its own small language models for on-premises deployment, integrated with Microsoft Azure, Google Cloud Platform (GCP), and Amazon Bedrock.

KOGO gives customers the ability to build their own virtual agents that can perform tasks such as handling client queries and identifying sales leads. Think of them as “AI employees” who perform tasks on their own, autonomously, completing multi-step decisions on users’ behalf without the need for human intervention.

“Choosing an AI agent is simple,” says Gopalakrishnan. Customers can visit the marketplace operated by the company, which went live earlier this month, choose a pre-defined AI agent from hundreds of options available, select it, customise it to their use case, and start using it immediately. He noted that this approach works well for mid-market clients who run chains of dermatology, skin, and hair clinics. For example, a client might want to implement a customer service agent to automate the process and expedite responses to customer queries, which currently rely on call centers where agents answer questions. Once they deploy the AI agent, it can not only take bookings on behalf of users but also handle other tasks.

However, the deployment of AI agents varies from customer to customer. For instance, one option is cloud-based, allowing anyone to visit the marketplace, select an AI agent, and customise it using a no-code approach. Conversely, if a large enterprise wants to deploy the AI agent, it will be entirely on-premises or on a private cloud. This setup includes locally hosted models, whether they are LLM, vision, or speech models, as well as the enterprise’s own data and systems that reside within the organisation. The operating system is then installed within the enterprise’s premises, and the OS serves as the framework that supports everything and runs their own AI store. Their AI store contains all the custom-created AI agents used within their environment. For example, an insurance company might want to use ready-made agents from KOGO’s AI store, along with about ten custom agents tailored to their processes. All of these agents reside within their AI store.

AI agents are built for productivity and designed for specific use in businesses, such as booking flights, filling out insurance forms, or filing expense reports. “We are not in the B2C space; we are hardcore in the enterprise space. We are B2B, and that is where we have built the AI,” Gopalakrishnan explained. He emphasised that these AI agents make more sense in businesses and enterprises where there is a need to increase productivity by freeing up time for more valuable tasks and reducing human intervention.

“I am not here to sell a service to end customers. I’m here to help an insurance company sell its services in a much better and more fulfilling manner with the power of AI, providing a faster customer experience for their clients. I’m here to help a bank offer better customer service and fulfillment, leading to faster, more efficient, and cheaper operations for their customers. I aim to create a more cost-efficient and effective AI implementation for a manufacturing company so they can reduce their cost of goods and offer better prices to their customers.”

The scope of artificial intelligence, Gopalakrishnan said, is vast because many organisations still face challenges in integrating their AI infrastructure with legacy systems. When these systems cannot communicate with modern-day systems, they become inefficient due to the sheer number of processes involved.

AI agents are the next big frontier for major tech companies, as tech heavyweights including Microsoft, OpenAI, Amazon-backed Anthropic, and Google are all introducing autonomous AI agents to the market. These agents can perform complex tasks and are capable of making decisions on users’ behalf, significantly reducing the time required for various processes from hours or days to mere minutes.

“AI agents have already hit the mainstream. It’s crazy how much business has been done in just six months. It’s only increasing, and we’re having to expand our capacity,” Gopalakrishnan said, adding that he expects the adoption of AI agents by small businesses to pick up in another six months.

Gopalakrishnan sees these autonomous artificial intelligence agents making the most impact in domains such as sales, customer service, operations, data management, HR, and especially manufacturing. “It [AI] will impact the job market, but not in the way that people are thinking. Those who know how to leverage and use AI and AI agents within their systems and organizations will keep their jobs, while those who do not will be at a disadvantage. Certain jobs will be lost, but a whole plethora of new jobs will also be created because of AI. This is very similar to the digitisation of banks that occurred in the early ’90s.”

Gopalakrishnan didn’t reveal the clients his company is working with in deploying these AI agents due to confidentiality, but he did mention how closely his company has worked with MapMyIndia, which has an app called Mappls that lists about 500,000 businesses.

“AI agents serve as the reasoning layer. It doesn’t matter what the model is; it can become larger and better, but the reasoning capability of the AI agent is already built-in. As the agentic framework improves, it will improve the ability to achieve higher levels of autonomy. This is what will lead to significant changes. The models are already quite advanced, and they will continue to improve as we move toward AGI. However, the path to AGI—often referred to as the path to the singularity—will largely depend on the agentic framework. When AI is capable of reasoning and making decisions, that will be the key to achieving AGI, and this is a fundamental part of the agentic framework.”