01
It reads
The file opens. The agent reads the CV against the job you opened, not a bag of keywords pulled from a title. The document is the input, not a filename standing in for it.
Teams searching for CV screening or AI resume screening usually get a keyword ATS, or a chatbot that summarises a paste. Kumo is the world's first Agentic HR Employee: a People OS with an agent colleague. It reads the CV against the role you opened, sees each tool's real result, and keeps going through the hiring mission. Sensitive steps wait. This is not legal advice.
Kumo HR Agent 2.0 reads a CV against the job, not a keyword cloud. Screening stays grounded in the role you actually opened. It reads the document, not a filename summary. Then it looks at what the last tool returned, and continues the mission.
The file opens. The agent reads the CV against the job you opened, not a bag of keywords pulled from a title. The document is the input, not a filename standing in for it.
After every tool call it looks at the real result. If the last step came back unexpected, the plan changes. This is a loop on live results, not a one-shot summary.
Screening is one step in a hiring mission. The same colleague can open a role, add the candidate, move the pipeline, schedule an interview, extend an offer, and hire, with you on the confirm.
Confirm gates, an honesty kernel, and a full audit trail apply. Sensitive actions wait. It cannot claim unfinished work. Every step, tool call, and result is in the log.
Most “AI CV screening” is a keyword parser with a new label, or a chatbot that talks about a paste. A traditional ATS stores the application and waits. An agentic colleague reads the document, sees the result, and keeps going through the hiring mission.
Give it a hiring goal in plain language. The same agent colleague that reads the CV can continue the work already in the live capabilities catalogue. Each of these waits for your confirm. This is not a chatbot bolted onto an ATS, and not a traditional HCM that stores the record and waits.
Create a job opening with an AI-drafted description, so screening has a real role to read against.
Add someone to the pipeline with a chat-attached CV stored on their record. The document travels with them.
Move candidates in the pipeline, schedule interviews, extend an offer, and hire. You stay the decision-maker on every sensitive step.
Autonomy is earned on trusted, repeated kinds of work. Until then, and always on anything sensitive, the agent pauses. Offers and rejects wait for a person. Screening is a first read against the job, not a guarantee of hire quality, and not legal advice.
Sensitive steps wait. Nothing that should not ship, ships. You stay the decision-maker on offers, rejects, and hire.
It cannot claim unfinished work. If a step did not complete, the log says so. No invented finish line.
Every prompt, retrieval, tool call, and approval is logged: searchable, explainable, and built on your private data, never trained on it.
Kumo is an agentic HR operating system: hiring, onboarding, payroll, leave, performance, policies, and compliance. The colleague that reads the CV is the same agentic HR employee that runs the rest of the mission.
Not a chatbot. Not a copilot. Not a traditional HCM. The category page for Kumo HR Agent 2.0.
Open /agentic-hr-employee →Missions, artifacts, skills, the 2.0 loop, and the honesty kernel. The full product explanation.
Open /intelligence →People, hiring, pay, performance, leave, learning, culture: one source of truth the agent colleague uses.
Open /platform →The live list of work the colleague can finish, including open a role, add a candidate, move, interview, offer, hire.
Open /platform/capabilities →How to buy agentic HR without the fog, and where Kumo sits in the field that actually shows up.
Open /blog/best-ai-hr-agents-2026 →Live product. Your real hiring questions. A human on the other side of the form.
Open /contact →30-minute demo · live product · your real questions.