Data Extraction Is the Moat
Zachary Kirby, co-founder of Vessel, on building the Vercel for integrations
This draws the real moat in integration infrastructure, which is not the pretty common schema on top, but the messy work of pulling fresh, complete, reliable data out of every source system. Vessel is arguing that product integrations like Salesforce and HubSpot are easier to scale because they usually expose open APIs, so the hard engineering shifts from scraping and access problems to building fast ETL, caching, webhooks, and developer tooling around those APIs.
-
Vessel separates itself from fintech and HR data aggregators like Finch and Rutter by saying those companies win on unlocking harder source systems, while Vessel wins on how quickly it can add and operate many open API based SaaS integrations. In the interview, Vessel says it expanded from about 20 to nearly 100 integrations in roughly two weeks by reusing a common pipeline.
-
That distinction matches how adjacent companies describe the market. Rutter says Plaid style scraping was built for uncooperative banks, while Rutter works mostly with open commerce APIs and goes much deeper on orders, products, and read write workflows. Ampersand makes a similar split, arguing unified APIs work best for standardized categories like HRIS, while deeper product integrations need tenant specific handling, observability, and rate limit management.
-
Merge and Finch both present themselves as unified APIs that normalize categories like HRIS, payroll, ATS, CRM, and accounting. Their value is integrate once, connect to many systems through one model, with refresh and auth handled centrally. That is a different buyer promise from Vessel, which emphasizes developer control and managed ETL for customer facing product workflows rather than broad normalization across sensitive employment or financial data.
The category is moving toward a clearer split. One branch will keep specializing in hard to unlock systems of record like payroll, banking, and commerce. The other will become infrastructure for product teams that need many fast, deep SaaS integrations inside their app. As AI products demand fresher operational data, the winners in product integrations will be the platforms with the best extraction pipelines, sync quality, and maintenance tooling, not just the cleanest common model.