Finch builds universal employment API
Jeremy Zhang, CEO of Finch, on building a universal API for employment systems
Finch matters because it turned payroll and HR systems from closed back office software into programmable infrastructure. Before that, a benefits or fintech product either had to build dozens of custom connectors, rely on CSVs and SFTP feeds, or force employers onto one bundled suite. Finch standardized 150 plus provider connections, reached about 88% market coverage, and let software vendors plug into the system of record without owning payroll themselves.
-
The bottleneck was not just data access, it was write access. Products like Human Interest need to both read employee and paystub data and write deductions and employer contributions into each payroll run. Finch abstracts that messy workflow across Gusto, ADP, Paychex, Paylocity, and long tail systems.
-
This is different from Pinwheel and Plaid. Pinwheel helps consumer apps verify income or switch direct deposit at the employee level. Plaid made bank accounts programmable for fintechs. Finch sits at the employer level, where the buyer is a B2B software company trying to sell retirement, insurance, lending, tax, or HR tools into many payroll systems.
-
The Plaid lesson is less about branding and more about market structure. Middleware starts as a connector tax, then wins by turning raw connectivity into higher value products. Finch is following that path from organization and pay data into deductions, contributions, reimbursements, and eventually a broader employment data layer.
The next phase is a more modular employment stack, where payroll stays the source of truth but more specialized apps sit on top of it. As embedded payroll players create even more underlying systems, Finch becomes more valuable as the translation layer, then moves up the stack into workflow and distribution products built on the data and payment rails it already controls.