Native Integrations as Core Product
Ayan Barua, CEO of Ampersand, on going upmarket with deep native product integrations
Integrations are not side work in SaaS, they are often the product work required to win enterprise customers. In Barua's prior spend management product, the hard part was pulling usage and payment data out of ERP, banks, and SaaS apps, then pushing clean records back into systems like NetSuite or Salesforce. That same pattern shows up across modern B2B software, where shipping the core workflow usually means owning data movement, schema mapping, retries, and tenant specific edge cases.
-
The shift from buyer managed integration to vendor managed integration is the key market change. Older iPaaS products let customers wire systems together themselves. Newer SaaS vendors are expected to ship native connections as part of the product, which moves a large chunk of engineering effort inside the app company.
-
Ampersand is built around the idea that unified APIs solve breadth, but enterprise deals need depth. A lightweight common model works in categories like HR, but CRM and ERP buyers often need custom objects, tenant by tenant rules, bulk syncs, and field level debugging. That is why Barua describes integrations as recurring infrastructure, not a one time feature.
-
The practical reason this consumes 40% to 50% of roadmap is maintenance. After the first Salesforce or NetSuite connector ships, every large customer asks for a different object, permission set, sync frequency, or data shape. Ampersand's product tries to turn that custom code into configuration so a RevOps or implementation team can adjust mappings without sending every change back to engineering.
This is heading toward a world where deep integrations become baseline infrastructure, like cloud hosting is today. As agents and AI apps demand fresher context and more write access across CRM, ERP, and communication tools, the winners will be the platforms that make complex customer specific data exchange fast, observable, and cheap enough that product teams can spend their time on workflows and intelligence instead.