Investing in Composio: Building the Backbone of AI Agent Intelligence

The rise of AI agents has created a fascinating paradox in enterprise software. These systems promise to revolutionize how work gets done, automating reasoning, planning, and complex execution at scale. Yet despite advances in compute and model design, they consistently hit a fundamental barrier: the inability to learn and improve through experience, a basic expectation of any effective employee.

The root of this challenge isn’t technical complexity. You can spend hundreds of hours developing LLM tools, refining prompts, and adjusting instructions, but eventually, you hit a wall. Unlike human workers, these models can’t build context, learn from mistakes, or develop the nuanced understanding that drives ongoing improvement. We believe this explains why many companies are hesitant to integrate AI agents into their workflows.

It’s the defining challenge of our current AI moment. And it’s why we’re thrilled to lead Composio’s $25 million Series A round.

The missing piece: practical, compounding knowledge across agents

The AI industry tends to focus on model capabilities: reasoning power, context windows, multimodal understanding. But, even the most sophisticated AI agent is ineffective if it can’t reliably accumulate practical knowledge the way people do.

This is precisely what we believe Composio has built: a shared learning layer that captures and distributes practical knowledge across the entire AI ecosystem. When any agent masters a Salesforce integration edge case or optimizes a GitHub workflow, those insights propagate to every other agent on the platform.

Already, Composio is engaging over 100,000 developers and supports more than 10,000+ tools across multiple categories, including GitHub, Notion, Linear, Gmail, Slack, Hubspot, and Salesforce, creating a powerful network effect. One where collective learning accelerates and enables exponential improvement in agent performance and adaptability.

Moreover, the insights layer that Composio provides—showing usage patterns, error rates, and system bottlenecks—also represents a new category of observability tooling specifically designed for agentic workflows. As enterprises deploy hundreds or thousands of AI agents across their operations, this level of visibility becomes essential for maintaining reliability and optimizing performance.

The builders fueling this breakthrough

We’ve had the privilege of getting to know Soham Ganatra and Karan Vaidya over a period of several years. Their journey spans half their lives, first meeting at a Physics Olympiad camp to then becoming roommates at IIT Bombay.

Both founders previously led engineering teams focused on integrations, where they witnessed firsthand the engineering pain points involved in connecting these complex systems. Initially, they aimed to use LLMs to automate coding agents. However, while building these agents, they realized an even greater challenge: seamlessly connecting AI agents to diverse applications. This insight became the foundation of the current iteration of Composio.

While both founders bring deep technical expertise from their previous roles at companies like Nirvana and Rubrik, we’re convinced that what sets this elite team apart isn’t just their technical credentials; it’s their obsessive focus on developer experience. With over 25,000 stars on GitHub, Composio has already demonstrated remarkable developer adoption, a clear signal that they’re solving a problem that resonates deeply with the developer community.

We believe Soham’s background as a founding engineer and technical product manager, combined with Karan’s software engineering expertise and active engagement with the developer community, creates this rare blend of product intuition and technical execution needed to have real-world impact.

The developer-first advantage

With the platform attracting over 100,000 developers, we believe this organic, developer-led adoption parallels the trajectory of other transformational infrastructure companies. When developers embrace a platform, enterprise adoption typically follows.

The shift from proof-of-concept AI projects to production-scale deployments also creates massive opportunities for infrastructure providers, who can handle enterprise requirements around security, reliability, and compliance. We believe Composio’s SOC Type II compliance and enterprise-grade authentication management equips them perfectly for this transition.

Towards AI that learns like humans

​​Looking forward, we believe Composio is positioned to become the definitive platform for AI agent integration and continuous learning. As the number of enterprise software tools continues to expand and AI agents become more sophisticated, we expect the value of a unified, reliable integration layer that enables ongoing improvement to increase.

The company’s focus on building a complete developer platform—with comprehensive documentation, extensive tooling, and strong community engagement—also creates the foundation for sustainable, long-term growth.

More broadly, Composio represents our conviction that the next wave of AI value creation will come from infrastructure companies that solve the practical challenges of AI deployment.

At Lightspeed, we’re proud to partner with Soham, Karan, and the entire Composio team as they build the connective infrastructure that will power the next generation of agentic AI. The future of work isn’t just about smarter AI, it’s about AI that can actually work and learn.

The content here should not be viewed as investment advice, nor does it constitute an offer to sell, or a solicitation of an offer to buy, any securities. The views expressed here are those of the individual Lightspeed Management Company, L.L.C. (“Lightspeed”) personnel and are not the views of Lightspeed or its affiliates; other market participants could take different views.

Unless otherwise indicated, the inclusion of any third-party firm and/or company names, brands and/or logos does not imply any affiliation with these firms or companies.