When we announced our Series A, we stressed that interoperability is the defining problem of the workforce ecosystem. Kombo exists to solve that. Today, our APIs move people data for 400+ companies, including Indeed, Experian, Greenhouse, Paychex, Thatch, Juicebox, Benifex, and Gem.

But API integrations cannot solve this problem alone. And this is exactly where our new browser agent integrations category fits in.

APIs cover a lot — but not every use case

There is a growing share of use cases that companies cannot solve through APIs. They might not expose the required fields or operations. Even when they do, access is often blocked by lengthy security reviews, partnership requirements, or onboarding costs exceeding the revenue they would bring. And there's a long tail of legacy and custom-built systems without (production-grade) APIs.

Translating UIs into deterministic APIs

For the first time, an AI agent can do what an API does — read and write back structured data reliably at scale — without that system having to provide an API. It interacts with the system through the public browser interface, just like a human would.

We call this a browser agent integration, and we're proud to have built the first one for people data. It translates UIs into deterministic workflows with API-grade reliability.

The progress of the big AI labs made this possible. Their models can reason about UIs with a fluency that didn't exist a year ago. But their general-purpose agents are built to figure out a task once, with a human in the loop.

That's why we spent the last twelve months building a layer that adapts them to the realities of enterprise UIs in workforce tech — strict security and compliance requirements, multi-step and conditional flows, messy data, and the many edge cases you hit at scale.

The result is a browser agent integration layer that is fully deterministic, debuggable, and compliant. Every run produces the same structured output, all actions are traceable, and integrations meet the reliability bar enterprises expect from critical infrastructure.

Introducing AI Apply, the first use case

The first use case we've productized on this layer is AI Apply. It lets you read jobs and submit applications across any ATS without an integration. We started here as many of our existing customers asked for it — so if you have unsolved integration use cases, don't hesitate to reach out.

It's used by leading recruitment players, including Recruitics, Bandana, Placed, CV-Library, and College Recruiter. Having collectively submitted hundreds of thousands of applications, they report:

• 3–5× more completed applications
• Near-100% attribution
• 99.9% delivery success rate
• Under 2 hours to onboard a new ATS

AI Apply is the first use case we shipped for ATS integration needs, and we're already exploring the next ones across HRIS and payroll.

One infrastructure layer for every people-data workflow

Our API and browser agent integrations are not competing integration approaches — they complement each other. Where partnership and APIs make sense, we'll keep building the deepest APIs in our market. Where they don't, the browser agent layer closes the gap.

Our customers often don't come to us with a clear integration question. They come with a use case, and our job is to solve it end-to-end. This will be through APIs where they exist, through the browser agent layer where they don't, and always with the reliability and technical support quality you expect from critical infrastructure.

People data is the most sensitive, most fragmented, and most important data a company handles. It can't be solved as a side feature of a generalist platform. It needs an infrastructure layer of its own. Every system, every workflow, every AI agent that depends on people data runs on one trusted data layer.

This is what Kombo is building towards.

— Alex Kübel, CEO and Co-Founder