Razorpay, India’s omnichannel payments platform, has launched RAY, a conversational AI account manager on WhatsApp, in partnership with IndusInd Bank. Bajaj Finance is among the first businesses live on the platform, which lets founders manage payments and get support around the clock without logging into a dashboard.
What RAY can help businesses with
Businesses can send RAY a message or voice note asking to perform some routine task. This could be, for instance, asking for a summary of yesterday’s payments, or requesting an analysis of the last two months’ transactions with a sales projection for the months ahead. They’ll get an answer in plain understandable language. As of Sept. 2, RAY has already handled more than 20,000 conversations since going live.
But answering questions is not the main news; this is already available in most AI chatbots today. Additionally, RAY proactively monitors a business’s payment health. It can flag an improvement in payment success rates, or warn a merchant that an upcoming settlement falls on a day before a bank holiday and offer to expedite it. Delivering insights before the business owner even asks is one of the AI assistant’s main upgrades.
RAY acts, it doesn’t just answer
What differs RAY from reactive genAI tools is that it independently and proactively triggers actions across Razorpay’s systems. Tasks that once required logging into a dashboard, such as issuing a refund, generating a payment link, or checking a settlement status, can now be completed directly inside a WhatsApp chat. This cuts out multiple steps and saves business owners considerable time they’d otherwise spend navigating menus or waiting on support calls. It runs on Razorpay’s AI Agent Platform, that makes it a single agent that draws on shared tools, knowledge, and memory rather than separate bots for each function.
Khilan Haria, Chief Product Officer, Razorpay said, “We’ve been thinking about how AI can fundamentally change the way businesses interact with their payments infrastructure. Millions of businesses across India are building and growing every day, and we wanted to make it easier for them to access the information, intelligent insights, and support they need along the way. With RAY, an AI Account Manager on WhatsApp, a business owner can simply ask for what they need in the most natural way possible and have it acted upon.
More importantly, we are moving towards a world where they may not even need to ask us first. If we see an opportunity to help, whether it’s improving payment success, staying ahead of a settlement, or taking an action at the right moment, RAY will proactively reach out and help. Every founder, regardless of the size of their business, should feel like they have someone looking out for them.”
Why small businesses stand to gain the most
Large enterprises have long had dedicated account managers. Most small and growing businesses have not. Manual, one-merchant-at-a-time outreach simply doesn’t scale to millions of businesses. RAY is built to close that gap, reaching every business automatically through an app that most Indian MSMEs already use daily, and will probably use even more now as the new bill payment feature just launched on WhatsApp.
That gap is especially wide at the smallest end of the market. India’s Annual Survey of Unincorporated Sector Enterprises (ASUSE) 2023-24 found the unincorporated sector had grown to about 7.34 crore (73.4 million) establishments, with “own account enterprises,” i.e. one-person or family operations with no hired workers, making up roughly 88% of that total. That is somewhere around 6.4–6.5 crore (64–65 million) businesses nationally. For this segment, a WhatsApp-based assistant that both answers and acts removes a major structural disadvantage.
Amitabh Saraff, Head – Institutional and Government Coverage, IndusInd Bank commented, “At IndusInd Bank, we are committed to making banking simpler, faster, and more accessible for businesses. Our partnership with Razorpay reflects this focus by bringing account management capabilities to a familiar platform like WhatsApp, helping make business support more immediate, intuitive, and convenient.”
Product Scale
Razorpay has said it serves millions of Indian businesses. Public company materials and past disclosures have cited figures that grew from roughly 8 million (2023) to over 12 million merchants more recently. Against India’s tens of millions of unincorporated enterprises, that still leaves a large untapped base RAY could eventually reach.
A caution on accuracy
As with any generative AI system, RAY can occasionally produce incorrect or fabricated responses. That possibility doesn’t pertain to this particular product, but is a well-documented limitation of large language models known as hallucination.
Since RAY can now also take actions like issuing refunds or generating payment links, errors carry more weight than in a purely conversational tool. Businesses should carefully verify AI-suggested actions before confirming them, especially when it comes to high-value transactions. It’s also best to keep human oversight for financial approvals, cross-check settlement and payout figures against the dashboard periodically, and report inconsistent responses so the system can be corrected.
