Payroll data models expanded into unified HR APIs and full payroll modules shipped, while ATS tools opened their pipelines to AI agent access via MCP.

The week in hr-recruiting

Payroll was the week's structural theme. Kombo's unified HR API added payroll data models — crossing from employment records into compensation and pay-run data. Ever Gauzy shipped a complete payroll module in a single release. These aren't the same kind of move: Kombo is adding data types to an integration layer, Ever Gauzy is adding a full product module. Both signal that the line between HRIS and payroll is contracting.

The second theme: AI access to recruiting pipelines. Teamtailor shipped MCP, letting AI assistants read candidate pipeline data and compare candidates across open roles. This is the beginning of AI-readable ATS data, which has different implications for recruiting workflow than AI-generated job descriptions or screening copilots.

Leaders

Ever Gauzy shipped a complete payroll module in a single release: PayrollRun and PayrollItem entities, a dedicated service layer, and full integration with its existing timesheet and approval infrastructure. For an open-source workforce platform, shipping the full payroll primitives in one release is architecturally significant — it completes the loop from time tracking to pay calculation.

Kombo launched payroll data models in its unified HR API alongside Skello and Zellis integrations. Payroll is a new data category for Kombo — previously its API covered employment records, not compensation. Adding Zellis (UK large enterprise payroll) specifically signals an enterprise market push rather than just API completeness.

Teamtailor shipped an MCP add-on that lets Claude, ChatGPT, and other AI assistants read pipeline data, compare candidates across roles, and surface ATS context to recruiting workflows. An ATS that's AI-readable at the data level — not just wrapped in a copilot feature — changes what's possible for recruiting automation.

Tanda launched the Reporting Agent: describe a report in plain text ("show me timesheet hours for Old Street last month broken by role") and the agent finds or builds the right report. For a workforce management tool with a large reporting surface, this is a meaningful usability improvement for non-technical HR managers.

Wildcards

Envoy brought ID scanning and verification fully in-house, replacing a third-party vendor with proprietary AI infrastructure. A workplace visitor management tool building its own ID verification AI is unusual — the category has typically relied on specialized vendors for this. The move suggests Envoy is betting that in-house control of the verification pipeline matters for enterprise security compliance more than vendor specialization.

Themes that compounded

• Payroll data became an API-accessible resource: Kombo's payroll data models and Ever Gauzy's payroll module both extend the programmatic surface for compensation data this week
• ATS data opened to AI access: Teamtailor's MCP integration makes candidate pipeline data readable by external AI assistants without custom API integrations
• Plain-language reporting arrived in workforce management: Tanda's Reporting Agent removes the query-building requirement for non-technical HR users
• Envoy's ID verification internalization signals that enterprise workplace tools are increasingly treating identity infrastructure as a core competency, not a vendor dependency

Watch this week

Watch how Kombo's payroll data models get used by integrators — adding compensation and pay-run data to a unified HRIS API is a different liability and compliance surface than adding a new HRIS connector, and how Kombo handles data sensitivity requirements will determine how quickly enterprise buyers adopt this capability. Also watch Teamtailor's MCP adoption: AI-readable ATS data could unlock recruiting automation patterns (agent-driven candidate sourcing, AI-assisted offer generation) that are currently bottlenecked on manual data access.