Key Takeaways
- • Ampersand raised a $15 million Series A led by Bessemer Venture Partners, with participation from existing investors Matrix and Flex Capital and new investors including Yelp, Tenacity Capital, CTO Fund, and Mana Ventures, according to the company's PR Newswire release.
- • Ampersand is building integration infrastructure that lets agentic AI applications read and write inside legacy systems of record such as CRMs and ERPs, addressing a bottleneck the company says slows enterprise AI adoption.
- • Lauri Moore, Partner at Bessemer Venture Partners, will join Ampersand's board, and the company announced the beta release of Andi, an AI integration agent that helps developers implement customer-specific integrations.
- • Co-founder and CEO Ayan Barua argues that the so-called "saasapocalypse" is overstated and that systems of record like Salesforce and NetSuite will remain central, positioning Ampersand as a bridge between the two software generations.
Summary
Ampersand announced a $15 million Series A led by Bessemer Venture Partners, according to a PR Newswire release issued October 6, 2026. The round included participation from existing investors Matrix and Flex Capital, and new investors Yelp, Tenacity Capital, CTO Fund, Mana Ventures, and several prominent angel investors. As part of the financing, Bessemer Partner Lauri Moore will join Ampersand's board.
The company is positioning its product against a practical obstacle to enterprise AI deployment. According to the release, CRMs and ERPs in many verticals have been in place for decades and are highly customized, so AI-native software vendors selling to enterprises must acquire deep domain expertise on how these systems function before they can deliver value. The release argues this integration burden slows enterprise AI adoption and limits the addressable market for AI vendors.
Ampersand's approach is to give each end customer's system a custom data model that accommodates custom fields, objects, permissions, and workflows, so agents have the context they need. Agents can then take bi-directional actions within systems of record. The company describes its offering as a complete integration solution for AI-native companies, combining integration infrastructure, domain expertise, forward-deployed support, and agentic support. It also announced the beta release of Andi, an AI integration agent designed to help developers with implementation work in each customer's environment.
John Pena, CTO at Hatch, an AI communications platform for service businesses acquired by Yelp in 2026, said Ampersand allowed Hatch to serve customers previously unavailable to it because of complex integration requirements, expanding its addressable market without months of maintenance work. Ayan Barua, co-founder and CEO, said that CTOs still face the legacy systems problem and that Ampersand is designed to make systems of record and newer agentic software legible to each other.
The founders bring integration experience to the company. Barua was CTO and co-founder of Siftery, which was later acquired by G2, and then served as VP of Engineering at G2, where CRM and ERP integration work consumed much of the engineering team's time. Co-founder Lauren Long built Firebase Extensions at Google, connecting Firebase apps to third-party APIs at a scale reaching billions of executions per day. The founders said they spoke with more than 100 companies building integrations before writing Ampersand's first line of code. Ampersand is based in San Francisco.
Moore said in the release that while the Model Context Protocol (MCP) lets agents describe tools, it does not address runtime challenges such as expiring tokens, schema drift, and differing customer configurations. Connecting an agent to a system of record is a demonstration, she said, while sustaining that connection across hundreds of customers with reliable writes is much harder.
The company is positioning its product against a practical obstacle to enterprise AI deployment. According to the release, CRMs and ERPs in many verticals have been in place for decades and are highly customized, so AI-native software vendors selling to enterprises must acquire deep domain expertise on how these systems function before they can deliver value. The release argues this integration burden slows enterprise AI adoption and limits the addressable market for AI vendors.
Ampersand's approach is to give each end customer's system a custom data model that accommodates custom fields, objects, permissions, and workflows, so agents have the context they need. Agents can then take bi-directional actions within systems of record. The company describes its offering as a complete integration solution for AI-native companies, combining integration infrastructure, domain expertise, forward-deployed support, and agentic support. It also announced the beta release of Andi, an AI integration agent designed to help developers with implementation work in each customer's environment.
John Pena, CTO at Hatch, an AI communications platform for service businesses acquired by Yelp in 2026, said Ampersand allowed Hatch to serve customers previously unavailable to it because of complex integration requirements, expanding its addressable market without months of maintenance work. Ayan Barua, co-founder and CEO, said that CTOs still face the legacy systems problem and that Ampersand is designed to make systems of record and newer agentic software legible to each other.
The founders bring integration experience to the company. Barua was CTO and co-founder of Siftery, which was later acquired by G2, and then served as VP of Engineering at G2, where CRM and ERP integration work consumed much of the engineering team's time. Co-founder Lauren Long built Firebase Extensions at Google, connecting Firebase apps to third-party APIs at a scale reaching billions of executions per day. The founders said they spoke with more than 100 companies building integrations before writing Ampersand's first line of code. Ampersand is based in San Francisco.
Moore said in the release that while the Model Context Protocol (MCP) lets agents describe tools, it does not address runtime challenges such as expiring tokens, schema drift, and differing customer configurations. Connecting an agent to a system of record is a demonstration, she said, while sustaining that connection across hundreds of customers with reliable writes is much harder.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.