From voice-powered search to on-device insurance advisors, PhonePe is quietly embedding AI into every moment of its users' financial lives, and the ambition goes far beyond convenience.
Imagine opening a payments app and asking it in your own language, "Can I book a Bharat Gas cylinder?" The app understands your query, retrieves the service, pre-fills your details, and hands you a single button to confirm.
This is not a concept from a product roadmap. At PhonePe, it is the operating direction. Rahul Chari, the company's Founder and CTO, describes more than 100 use cases within the PhonePe app being redesigned around AI-driven, voice-enabled, natural language interfaces all of which are a deliberate push to make the platform as frictionless as a conversation.
The urgency behind this shift is structural, not cosmetic. PhonePe's 300 million monthly active users span an extraordinary range of digital sophistication. The most experienced users navigate the app's full breadth with ease. But a substantial segment is only comfortable with basic transfers and bill payments and never reaches the wealth, insurance, or investment products that sit deeper in the app. AI, Chari argues, is the bridge.
Talk to your money: Voice, Language and the Search Revolution
Language has historically been one of the hardest barriers to digital financial inclusion in India. A user fluent in Tamil or Kannada but not comfortable with English or even Hindi-script interfaces has faced a narrower, less useful digital experience.
AI-powered voice search, built on large language models capable of processing vernacular speech, changes that equation. Chari's vision is for every query across PhonePe's 100+ use cases to be answerable in natural language, not just returning a text result, but navigating the user directly to the right page, pre-populated with their information, ready for one-tap action.
"To layer natural language processing, context, and semantic understanding in search seamlessly for every use case in the PhonePe app and making it almost one click, is a huge opportunity," as per Rahul Chari
The ambition extends well beyond search. Conversational transaction insights are being built to let users ask questions like "how have my household expenses trended over the past year?" or "where can I save money this month?" and receive not just an answer, but a recommended action, whether that's a fixed deposit, a mutual fund, or a revised spending category.
From transactions to journeys: Using AI in wealth, Insurance and Travel
Some of PhonePe's most interesting AI work is happening in the categories that have always required the most hand-holding. These include financial planning, insurance selection, and travel booking. These aren't transactions, they are journeys, requiring research, comparison, personalisation, and trust.
AI Guided Investing (Share.Market)
Share.Market uses AI to hand-hold first-time investors into equity markets with offerings like portfolio analyzers, goal-based investment models, back-tested recommendations, and automatic portfolio adjustment. The goal is to make users into smart traders with minimal friction, starting from as little as Rs. 5,000.
Conversational Insurance Research
Buying health insurance requires real research with claim ratios, family medical history, policy fine print. PhonePe is building a fully agentic AI flow that guides a user through this research conversationally, then completes the purchase and policy issuance without leaving the app.
Holiday Planning, Not Just Booking
PhonePe books flights, trains, and hotels, but travel decisions happen before booking. An AI-driven holiday planning interface, embedded within the app, guides users through itinerary research and date selection before handing off seamlessly to the booking flow.
Visual underwriting for merchants
For merchant lending, AI can now supplement payments data with real-time store images assessing establishment age, upkeep, and inventory to make more accurate, often more generous, credit decisions. Unstructured data becomes a financial asset.
The Privacy Bet: AI that never leaves your phone
As AI becomes more deeply embedded in personal finance, privacy becomes a genuinely hard problem. Health data, family history, sensitive policy documents - users may hesitate to share these with any platform, no matter how trusted.
PhonePe's answer is edge AI where the models are small enough to run directly on a user's device, so sensitive data never has to touch a server. In partnership with Microsoft Foundry, the company has already deployed a hybrid edge-cloud model for transaction insights. The edge component handles locally sensitive inferences; the cloud handles what requires it.
Chari's most vivid example: Imagine uploading a 15-page insurance policy document for Q&A - "what does my policy actually cover in case X?" in a way where the document, your questions, and the answers all stay on your device, and it is just a private, local conversation with your documents.
The efficiency number that tells the story
PhonePe has achieved massive operational leverage through AI-driven customer support. Around 500 agents now handle the support load of 10 billion monthly transactions, with 92% of tickets fully automated through chatbots, voice bots, and AI quality monitoring. That's one agent per 20 million transactions.
Building Digital Literacy at Scale
PhonePe's recent partnership with OpenAI is about distributing access to AI tools at population scale. By making ChatGPT accessible through the PhonePe platform, the company is accelerating the digital and AI literacy of its user base, tapping users who may be experiencing this calibre of AI interaction for the first time.
"Getting them to become AI and digital savvy faster also helps us as a digital-native platform," Chari explains. The Microsoft partnership operates differently with co-developed solutions and shared infrastructure investment. Together, the two partnerships frame PhonePe's external AI strategy with one driving infrastructure depth, the other driving user-facing reach.
"AI is a huge accelerator in our ambition. We see ourselves as a population-scale, intelligent distribution platform and AI allows us to deliver on that," as per Chari.