New Delhi, Sept. 11 -- In his AI Tech Talk at the Global Fintech Fest 2026, Rahul Chari, Founder & CPTO, PhonePe, laid out a simple thread that guides their AI philosophy: remove friction wherever it shows up. For the consumer, it's navigating through the app to find what they're looking for. For the merchant, especially small businesses, it's trying to take their business online. For the engineer, it's last-minute threat analysis and crash debugging.
Chari's talk moved through each of these layers, showing how the core principles play out differently depending on who's on the other end.
Intent-led Product Experiences
A user who knows what he needs may not straightaway be able to land on the right menu, especially if the UI is multi-layered. It requires them to dig deep into the app to find the right icon to initiate the transaction. PhonePe's answer is to let people simply say it.
With AI Search, you can tap the search button and either type or speak - "I want to buy Gold," or in Hindi, "Mujhe sona khareedna hain". The app understands the intent, pre-fills known data, and takes you straight to the payment screen. No navigating, no guessing which tab to tap.
The same experience is applied to money management. If you need to understand your recent spending on household expenses, you can simply head over to PhonePe's AI Transaction Insights and say, "How have my household spends trended over the last quarter?". You will instantly get back a graphical breakdown of your expenses recorded on the app, across categories like electricity, gas, mobile, and broadband - giving you an insight into your spends as well as suggesting saving and investing options that will work for you.
Chari points out that insurance buying can be intimidating to many, given the jargon and choice overload. PhonePe's agentic AI simplifies the buying decision journey with its conversational interface. It helps you research, break down complexity, and land on the best possible plan. Let's say you need to get a car insurance policy, you simply enter your car's registration number and the agent will provide plans that fit, along with insurer options. Once you make your choice, it will walk you through the entire journey until policy issuance.
Creating Value for Merchants and the Ecosystem
Chari also touches upon their efforts to empower their growing merchant base by filling the digital gap. A small homegrown business selling flowers, handicrafts, homemade food, and the like can take their business online without the need for a web designer or developer. Through their AI-powered "PG SmartPages" solution, the merchant can describe their product, even in a vernacular language, and the AI generates the copy, selects the fields, and builds a mobile-optimised, payment-ready page in under 10 minutes!
For larger digital platforms, the challenge lies in integrating a payment gateway that typically takes weeks of coding effort. The PhonePe AI Integration Agent compresses that timeline from weeks to minutes, by turning a highly technical process into a simplified conversational solution that follows PhonePe PG's existing API specs, corrects payloads, and enables proper error handling. The slashed integration times enable engineers and businesses to invest in improving their products and services.
Chari also pointed to Pulse Pro and PhonePe's role in giving back to the ecosystem through data. A prospective entrepreneur planning a business in, say, Punjab can actually leverage the data available for hyperlocal, near real-time market insights. This helps them make faster, smarter business decisions. At the session, he introduced PhonePe Pulse's newest feature - Ask Pulse, an AI-powered chatbot that allows you to find the insights you're looking for simply by having a conversation.
Operational Efficiency: the Foundation Underneath it all
This consumer and merchant scale is supported by a deeply efficient, AI-augmented engineering culture. Chari speaks in detail about the AI-powered internal solutions that were built to improve developer experiences within the organization. "TraceMind", an internal debugging tool takes the chaos out of crash debugging, shrinking threat-analysis and debugging work from days to minutes while keeping the core platform stable. "Lens" gives engineering teams real-time "X-ray vision," accelerating UI generation and live mobile debugging.
The more advanced piece is "KadhAI", built for centralised code orchestration. It plans and implements code changes in isolated sandboxes, runs tests and checks, and routes everything for human approval. This cuts code scaffolding from 2-3 hours down to roughly 15 minutes, freeing developers to focus on high-level system design instead of boilerplate code.
This internal efficiency built over AI isn't confined to the tech org. Agent Hub, PhonePe's internal platform for deploying AI agents securely across engineering, business, and corporate functions, now runs more than 200 internal agents. Together they handle over 1,600 daily queries for 759 weekly active users, saving roughly five minutes per automated task, with automation covering up to 50% of daily tasks in some departments.
Outcomes, measured
PhonePe's AI story lands not in its features, but in numbers. PhonePe processes 10 billion transactions a month with just 500 support agents - about one agent for every 20 million transactions. By automating more than 92% of Customer Support tickets, the team has created 40x more value while staying lean.
Fraud investigation tells a similar success story. AI now handles roughly 80% of end-to-end fraud investigations, resolving complex, multimodal threats - that used to once take 90 minutes - almost instantly. The cost per investigation has dropped from around Rs.50 to Rs.5-Rs.15.
From a one-line voice command to a fraud case cleared in seconds, the arc Chari describes is really about one idea: intelligence, built to disappear into the background - accessible, efficient, and measured by what it actually changes.
Published by HT Digital Content Services with permission from Millennium Post.