Why Paytm Is Building Applied AI, Not Just Another Chatbot

Artificial intelligence is rapidly becoming part of everyday technology. Yet much of the conversation around AI still centres on chatbots and conversational interfaces. At Paytm, we are taking a different approach.

We are building applied AI, which is artificial intelligence designed around specific business problems, merchant workflows and consumer journeys. Rather than treating AI as a separate product layer, we are embedding intelligence into how our platform operates, from merchant onboarding and servicing to fraud prevention, collections, customer acquisition, personalisation and software development.

As highlighted in our Q1 FY 2027 earnings release, AI is now part of our core operating processes. We have developed function-specific models and agents by fine-tuning open-source models, with applications across engineering, merchants and consumers. That is the foundation of our applied AI approach, and it is already showing up in how fast the business is growing and how efficiently it is being built.

Applied AI Built Around Real Business Use Cases

Our approach to AI begins with a simple question: what business problem can technology solve better?

For merchants, we started by applying AI to onboarding, servicing and business insights through the Paytm AI Soundbox. We are now building agents that help merchants market their services, engage with customers across channels including the Paytm app, and service those customers more effectively.

AI is also becoming more relevant across our merchant financial services ecosystem. Our merchant loan distribution business uses AI-led lifecycle management of device merchants, including risk insights for lending partners and collections. In Q1 FY 2027, AI-led capabilities helped drive gains in merchant engagement, retention, risk insights for partners and collection efficiency.

Across consumers, AI-led acquisition is helping us select better-quality customers more efficiently, while AI-led personalisation is helping serve more relevant use cases and improve engagement across financial services products. The common thread is that AI is being deployed where it can influence a real operating or customer outcome.

Specialised AI Models for Indian Businesses

Applied AI also requires the right model for the job. Rather than relying only on very large general-purpose models, we are building and optimising specialised models for our own use cases.

During the Q1 FY 2027 earnings call, our Founder and CEO Vijay Shekhar Sharma explained that we tune models ourselves, deploy them on our own infrastructure and optimise them for specific applications. He shared one example where a 200-billion-parameter model was optimised into a 4-billion-parameter model designed for Indian languages, and then placed on our own machines.

As Vijay Shekhar Sharma put it on the call, this means low latency, a lower cost of tokens and low running cost, operated by us. That approach makes our cost lower than it would be for a typical company, and it lets us replace expenses such as third-party call-centre costs while adding a new capability of our own.

This matters in a market like India, where businesses operate across many languages, regions and very different operating environments. For applied AI to work at scale, it needs to be fast, cost-efficient and suited to the context in which it is used. Our earlier work around the Paytm AI Soundbox follows the same philosophy, with applied AI models and smaller language models designed around small-business contexts, voice and Indian languages.

AI for Merchants and Small Businesses

India’s small-business ecosystem is one of the clearest opportunities for applied AI. A local merchant does not necessarily need another standalone technology interface. Technology needs to fit naturally into the way the business already operates.

That is why the Paytm AI Soundbox is such an important part of our AI journey. The Soundbox has evolved beyond payment confirmation into a broader merchant operating interface, bringing business information and merchant assistance to a device already present at the storefront.

In Q1 FY 2027, we had 1.57 crore merchants on subscription plans, adding 27 lakh net devices year-on-year, while merchant GMV grew 31% year-on-year to ₹7.1 lakh crore. We described the Soundbox as an indispensable operating system deployed across 1.57 crore storefronts in India. This existing distribution gives us an opportunity to bring AI-led services directly into everyday merchant workflows, rather than expecting merchants to adopt an entirely new system.

AI Across the Paytm Ecosystem

The use of AI extends well beyond merchant-facing products. Across engineering, we are using agentic assistance for coding, review, testing and deployment. As highlighted in our Q1 FY 2027 earnings release, this is helping create faster delivery cycles and lower the cost of building software, while the end-to-end development cycle is being made AI-ready across its different components.

Across consumers, AI-led acquisition helps us identify better-quality customers more efficiently, and personalisation helps serve more relevant use cases and improve engagement. Across merchants, AI is being applied to onboarding, servicing, business insights and sales workflows.

On the earnings call, Vijay Shekhar Sharma described how AI now supports our field sales workflow, with an agent helping identify what our field sales executives should do in small-business merchant acquisition, built in-house. Together, these applications reflect a broader shift, where AI is becoming part of how the business operates rather than remaining a standalone technology initiative.

AI-Led Operating Leverage

The impact of this approach is becoming visible in our operating metrics. In Q1 FY 2027, our revenue increased 28% year-on-year to ₹2,448 crore, while total indirect expenses grew by only 6%. EBITDA reached a record ₹203 crore, up 182% year-on-year, our highest ever quarterly EBITDA.

The cost of building our platform declined 3% year-on-year, from ₹749 crore to ₹729 crore. Software, cloud and data-centre expenses declined 5% to ₹159 crore, even as we continued to invest in AI. We attributed the reduction in platform-building costs in part to significant AI-led optimisation and productivity gains that helped absorb annual appraisal increments.

At the same time, we continued to invest in growth. For us, AI-led efficiency does not simply mean cutting costs. It means building operating leverage so the business can grow faster than its underlying cost base. On the earnings call, our President and Group CFO Madhur Deora noted that our EBITDA margin, adjusted for the PIDF incentive, improved from 1% to 8% year-on-year, even as we kept investing where it made sense.

From AI Efficiency to AI Monetisation

The next phase of our applied AI approach goes beyond internal productivity. During the Q1 FY 2027 earnings call, Vijay Shekhar Sharma said we are beginning to move from an AI optimisation journey towards an AI monetisation journey. He also said some of our AI products have already started generating a small amount of revenue, and that these non-payment, non-financial-services lines are an area of personal focus.

The focus is particularly on merchants and businesses. We are building AI solutions that we first deploy within our own operations and can then take to other businesses, having relied on those solutions ourselves. Management has indicated that smaller businesses and larger enterprises are likely to need different types of AI products and services.

That creates a natural progression: build for a real internal problem, improve the technology through use at scale, and then identify where the same capability can solve problems for other businesses.

Why Applied AI Matters

The value of AI will ultimately be determined by what it can do, not simply by how convincingly it can talk. For us, that means AI assisting a merchant onboarding process, helping improve collection performance, supporting better risk insights, making software development faster, delivering relevant consumer experiences, or bringing business intelligence to a merchant through the Paytm AI Soundbox. These are practical applications tied to existing workflows, products and distribution.

At Paytm, we are building AI with that purpose in mind. Not another chatbot, but intelligence applied where it can make technology more useful for merchants, consumers and businesses across India.

FAQs

Q. What is applied AI at Paytm?

Applied AI at Paytm refers to the use of AI models and agents for specific business workflows such as merchant onboarding, servicing, fraud prevention, collections, customer acquisition, personalisation, risk insights and software engineering, rather than as a standalone chatbot product.

Q. Is Paytm building its own AI models?

Yes. We fine-tune and optimize models for specific use cases and deploy them on our own infrastructure. 

Q. How is Paytm using AI for merchants?

Paytm is applying AI to merchant onboarding, servicing, business insights, collections and lifecycle management of device merchants. We are also building agents that help merchants market their services, engage customers and service them across channels, including the Paytm app, with the Paytm AI Soundbox acting as an operating interface at 1.57 crore storefronts.

Q. How is AI helping Paytm improve efficiency?

In Q1 FY 2027, Paytm reported AI-led productivity and optimisation across platform-building and engineering workflows. The cost of building the platform declined 3% year-on-year, while software, cloud and data-centre expenses declined 5%, even as we continued investing in AI. Revenue grew 28% year-on-year while total indirect expenses grew only 6%.

Q. Does Paytm plan to monetise its AI products?

Yes. Management has said we are moving beyond AI-led optimisation towards AI monetisation, including AI-led services for merchants and businesses. On the Q1 FY 2027 earnings call, Vijay Shekhar Sharma said some AI products have already started generating revenue at a small scale.

Q. Is Paytm’s applied AI focused on consumers or merchants?

While AI-led personalisation and acquisition benefit the consumer business, our AI monetisation focus is particularly on merchants and businesses. On the earnings call, Vijay Shekhar Sharma said Paytm has decided to take AI solutions it uses internally to merchants and businesses, with smaller businesses and larger enterprises likely to need different types of AI products and services.