loading
loading
For a company that started out sending SMSes, Gupshup has spent nearly two decades learning one thing: how businesses talk to their customers. That experience is now becoming the foundation for its next bet: conversational AI. Gupshup is no longer positioning itself simply as a messaging infrastructure company. Its product stack is expanding from SMS and WhatsApp to voice AI, AI agents, campaign orchestration and tools that allow enterprises to connect multiple AI models with customer-engagement channels. The shift is already showing up in the business. At the global level, CEO and cofounder Beerud Sheth said for FY26, Gupshup is roughly $350 Mn business and profitable. But that is not the real story here given that global revenue in FY23 was around $300 Mn. More importantly, Sheth said 25% of the revenue today now comes from AI-related offerings. And this transition also resulted in some workforce reduction over FY26, but the CEO believes Gupshup is different from a conventional startup launching an AI product because it has seen how communication and messaging has evolved. Today, Gupshup is effectively trying to AI-ify an existing enterprise customer base, using the infrastructure, relationships and transaction volumes accumulated over years as its competitive moat. From Messaging APIs To AI-Powered Conversations Gupshup’s evolution has broadly tracked the evolution of customer engagement. The company started with SMS and basic messaging in 2004. WhatsApp brought richer, two-way conversations, while RCS and other channels expanded the ways businesses could interact with customers. Gupshup now sees AI-powered conversations across text and voice as the next stage, having entered the segment in February 2025. “We’ve been doing this business for 15–20 years,” Sheth said in an interaction with Inc42. The company’s focus, he said, has been helping businesses engage customers through mobile messaging and voice, with newer interfaces potentially emerging over time. The important distinction is that Gupshup does not see voice AI as a standalone channel. Its recently launched voice AI offering is a self-serve platform for deploying AI agents that handle customer calls. The company says these agents can resolve inbound support calls, qualify leads and automate operational calls such as scheduling, verification and payment reminders. The platform starts at $0.035 (₹3.50) per minute, compared with industry pricing that can range from $0.05 to $1 per minute. Sheth expects the cost of voice AI to fall further as the underlying technology becomes more efficient. Gupshup is also experimenting with smaller, specialised models that can deliver faster and cheaper inference for specific use cases. The platform is model-agnostic, with integrations including Deepgram, ElevenLabs and Cartesia. The company is also evaluating open-weight and Indian models. Gupshup’s existing voice infrastructure is another part of the equation. The company acquired voice technology business Knowlarity in 2022 for $100 Mn, giving it an established voice stack. According to Sheth, Knowlarity’s infrastructure handles around 500 Mn calls a month. Gupshup is now adding an AI layer to that infrastructure and integrating it with the broader platform. Voice AI is only one part of that strategy. In April, the company introduced Super Agent, an orchestration layer designed to bring multiple customer-engagement functions into a single interface. Businesses today have to manage messaging channels, campaigns, customer journeys, AI agents, CRM systems, payments, catalogues, human escalation and analytics, often across different platforms. Super Agent is designed to bring these workflows together through natural-language instructions. Rather than building another foundation model, Gupshup is positioning itself as the layer connecting different models to its APIs, workflows and enterprise integrations. A customer could use Meta’s Business Agent, OpenAI, Anthropic or an open-weight model while relying on Gupshup for the surrounding infrastructure. The company has also launched Superclaw, a self-hosted, on-device version of Super Agent built on the open-source OpenClaw platform. These AI products and other advanced orchestration capabilities now account for roughly a quarter of Gupshup’s revenue, according to Sheth. He added that the newer offerings are growing faster than the company’s traditional messaging business. This is not Gupshup’s first shift in response to a changing technology landscape. The company began as SMS Gupshup, allowing users to create interest-based groups and receive SMS updates, similar to early social-media platforms. After around four years of rapid growth, rising SMS costs began to weigh on the business. Gupshup subsequently moved away from the consumer-facing model and focused on enterprises in 2011. The company is now attempting another transition as business messaging changes again. “Every business has to support more than just messaging. It needs to support voice and, eventually, other channels that may emerge, such as video, smart glasses and other surfaces. Businesses will need to manage interactions across all these channels,” Sheth said. Can Incumbency Become An AI Advantage? This is where Gupshup’s strategy becomes particularly interesting. Sheth describes its advantage as a combination of “incumbency and innovation.” India’s voice AI market is expanding rapidly, with projections estimating the sector to grow from roughly $130–$150 Mn to nearly $1 Bn by 2030, while the broader conversational AI market is forecast to reach $3.7 Bn by 2033. But the opportunity comes with a crowded competitive landscape. AI-native startups are building specialised agents, voice interfaces and vertical applications from scratch. Their advantage is speed and product focus. Gupshup starts from the opposite end. It already has the customers, messaging infrastructure and enterprise relationships. The company serves more than 50,000 businesses across 130 countries and processes over 120 Bn messages annually across WhatsApp, voice, web and mobile. That gives it a potentially powerful distribution advantage. Sheth said customers that started with SMS years ago subsequently added WhatsApp, RCS, AI, chatbots, campaign management and other tools. As customers adopt more products, their spending with Gupshup rises. That is also translating into higher revenue per customer. “For the same customer, deal sizes are increasing because they are using more products. Many customers started with SMS, then added WhatsApp, followed by RCS, and are now using AI, chatbots, campaign platforms and other products,” said Sheth. Some of its largest customers can spend as much as ₹10 Cr a month, he added. But the financial picture is less straightforward. Gupshup’s India entity saw revenue decline from ₹2,051 Cr to ₹1,943 Cr in FY25, while net profit fell from ₹54 Cr to ₹26 Cr. Sheth argues that the India entity does not represent the company’s entire global business, but the numbers nonetheless show the challenge of translating the AI transition into financial growth. The company has also faced pressure on its valuation. In January, a fund run by Fidelity Investments marked down its valuation of Gupshup to $278 Mn, more than 80% below the $1.4 Bn valuation at which Fidelity first invested in 2021 through a secondary transaction. The fund had previously marked down its holding to $697 Mn as of July 2023. The markdown came amid a broader reset in software valuations as investors reassessed the durability of traditional SaaS businesses in the age of AI. Gupshup also laid off more than 100 employees last year as part of a cost-cutting exercise. Sheth did not comment on these parts of the past year for the company. The AI pivot is more than a product expansion; it is also a test of whether an established messaging business can create a new growth engine before its legacy business becomes structurally less attractive. The Meta Problem While incumbency may give Gupshup a head start, it does not insulate the company from intensifying competition. Meta-owned WhatsApp is a case in point. Gupshup is a Meta Business Solution Provider (BSP), giving businesses access to the WhatsApp Business API. But WhatsApp itself launched Meta Business Agent this year, bringing AI-powered capabilities that can automate customer interactions, recommend products, book appointments, qualify leads and provide 24/7 support. That puts WhatsApp closer to the application layer Gupshup is increasingly targeting with its own AI offerings. Sheth argues that Gupshup’s model-agnostic approach will remain relevant even as individual platforms build their own AI capabilities. “Meta has launched Meta Business Agent, which comes with its own LLM and agent capabilities. But regardless of which model is being used, Gupshup’s role remains the same: helping enterprises put the different components together and orchestrate them,” said Sheth. Enterprises, he added, do not want to be locked into a single model as the AI landscape changes rapidly. They want the flexibility to use different models for different applications, which Gupshup believes gives its orchestration layer a role to play. The same dynamic is emerging in business messaging. Reliance Jio has introduced new pricing for RCS messaging for businesses following its partnership with Apple to extend RCS support to iOS users in India. This brings Apple into India’s business messaging market alongside Google, potentially challenging WhatsApp’s dominance in rich, two-way messaging. For Gupshup, RCS is a double-edged sword. Apple and Jio could expand the market for rich business messaging, but they also bring more competition to the infrastructure layer where Gupshup has historically made money. That makes the company’s shift towards AI orchestration more important. If messaging infrastructure becomes increasingly commoditised, Gupshup needs to capture value higher up the stack. The Road To The Public Market This mix of opportunity, competition and financial pressure makes Gupshup’s next move worth watching. Just last year, the company raised another $60 Mn from Globespan Capital Partners and EvolutionX Debt Capital to support product innovation and expansion across markets including India, the Middle East, Latin America and Africa. The company is also exploring listing venues and evaluating the legal, tax and structural implications of potentially redomiciling to India from the US. Sheth has indicated a potential 12–24 month window for a listing, although there is no firm IPO date yet. An IPO would put Gupshup’s AI narrative under a much tougher lens. But IPO talk is not new in Gupshup’s context, nor is the reverse flipping, which Sheth first talked about in February 2025. The company brought in Ravi Dugar, who took Awfis through its IPO, as CFO in March 2026, and this is a concrete signal that Gupshup is looking to overcome the reverse flipping hurdle and list in India. Investors will not simply want to know how many AI agents the company has launched. They will want to know whether AI is producing higher revenue per customer, improving margins, increasing retention and creating a durable competitive moat. That is ultimately the bigger bet behind Gupshup’s AI push. The company is betting that its incumbency is not a legacy burden but an asset, and that years of customer relationships, messaging infrastructure, voice capabilities and enterprise integrations can make it easier to move customers into an AI-first world. The question is whether that installed base can become a genuine AI moat before the companies controlling the underlying channels, models and AI interfaces capture more of the value.
document fetched 2026-09-24 02:53 · canonical https://inc42.com/features/gupshup-ai-rebuild-can-it-stay-ahead-of-the-platforms-it-runs-on