The agent recruits. You manage talent. The Workable Agents are here.
Four AI recruiting agents are now available to every Workable customer globally. The Job Brief Agent, Sourcing Agent, Screening Agent, and Engagement Agent work together to autonomously handle the full recruiting workflow—creating job briefs, sourcing qualified candidates, evaluating applications, and nurturing talent relationships. They’re built natively into Workable, priced on a credit-based, pay-as-you-go model, and ready to transform how your team recruits.
This launch also ships with a new Workable MCP Server that connects these agents to Claude, ChatGPT, and other AI assistants through the Model Context Protocol. Teams can now manage pipelines and source candidates directly from their preferred AI chat interface.
What’s changing
Four agents, one workflow. The agents orchestrate the entire candidate evaluation pipeline without manual intervention. The Screening Agent evaluates every applicant against a structured Ideal Candidate Profile across 14 dimensions. Where information is missing, the Engagement Agent opens a personalized conversation to collect targeted follow-up questions, then re-scores instantly. The Job Brief Agent structures your hiring requirements, and the Sourcing Agent reaches passive candidates continuously. The result: a ranked shortlist of qualified, interested candidates ready for interviews.
Credit-based pricing that scales with you. Gone are the days of paying for agent access you might not use. Every Workable plan now includes 3,000 free AI credits. Screening costs 1 credit per candidate, sourcing costs 2 credits, and a completed candidate chat costs 10 credits. Buy additional credits in bundles: 5,000 credits at $0.12 per credit, 10,000 at $0.10, or 50,000 at $0.095. You only pay for the actions you actually run.
New features in GA. Recruiters can now adjust the agents’ ideal candidate profile retroactively while a job is active—changes apply instantly to recently processed candidates. You can also attach agents to jobs already in progress without creating a new posting from scratch. Recruiters receive interruption warnings before any action would disrupt the agent’s active workflow, and each candidate record shows the agent’s current processing status and planned next steps.
Headless agent operations. The Workable MCP Server connects your recruiting and HR capabilities to Claude, ChatGPT, and other Model Context Protocol-compatible AI assistants. No developer setup required. Teams can source candidates, update pipelines, manage employee records, and run reports without leaving their chat interface. The MCP Server respects all existing Workable user permissions and is included at no extra cost on every subscription plan.
How it works
Open any job posting in Workable . Enable the Job Brief Agent to structure your hiring requirements based on the role. The Sourcing Agent begins reaching out to passive candidates in your network continuously. As applications arrive, the Screening Agent evaluates each one against your ideal candidate profile. When information gaps appear, the Engagement Agent initiates a chat with that candidate to ask targeted follow-up questions—then updates their score automatically. Within hours, you have a ranked shortlist of qualified, engaged candidates ready for your team to interview.
The agents retain full context across the entire workflow. They know which candidates dropped during screening, why they re-scored after engagement, and which jobs are consuming the most credits. You can drill into candidate records at any time to see exactly what the agents learned, what they asked, and why they scored each candidate the way they did.
Why this matters for recruiting leaders
Stop the manual information sift. Before the agents, recruiters spent hours processing applications, manually scoring candidates against your requirements, and chasing down missing information. The agents eliminate that work entirely. CacheFly’s head of HR, Martin Lagler, said the Screening Agent “eliminated about 20% of unqualified applications automatically. In the past, I had to go through every answer manually. Now I don’t waste anyone’s time.”
Keep your team focused on what matters. Talent relationships, interview calibration, and hiring decisions are where your recruiter’s judgment shines. Let the agents handle the information processing. H2 Health, which scaled from 60 to over 300 locations in six years, relies on the Sourcing Agent for continuous candidate discovery across clinical roles. Director of Recruiting Carol Ribeiro noted: “Anything that’s doing the sourcing in the background saves us time. It’s less time that the recruiter actually needs to be sourcing for that specific position.”
See every credit move. A new AI Credits Report shows exactly how credits flow through your account—what’s purchased, used, expired, or transferred. Break down consumption by job, action, and time period. Filter by department, location, job status, and more. Budget decisions based on data, not guesswork.
Bring AI into your chat. The Workable MCP Server lets your team instruct Claude or ChatGPT to run sourcing searches, update pipelines, pull reports, or manage onboarding workflows directly from chat. No app switching. No integration friction. Your AI assistant now has deep access to your recruiting and HR data.
By the numbers
As of today’s launch, the agents have processed more than 120,000 candidates since their initial rollout. Early adopters include CacheFly (CDN), Belong (UK dementia care charity), H2 Health (multi-location healthcare), and HSI (Health and Safety Institute). Belong’s 3-person recruiting team filled a Head of Operations role in seven days using the agents—a timeline that would have been impossible with manual screening.
Get started
The Workable Agent and MCP Server are available now. The agents are a paid add-on to any Workable Recruiting plan. The MCP Server is included at no additional cost on all subscription plans globally.
Every account using the agents has access to the AI Credits Report in Recruiting reports—full transparency into your AI spend, by default, with no setup required.
