Fordel StudiosRecruitment Screener
Semantic candidate screening that ranks on skills, not keywords.
The problem
being solved
A company hiring for 30+ open roles receives 200-500 applications per position. Recruiters spend 6-8 minutes per resume on initial screening — a single role consumes 20-50 hours before a single interview is scheduled. Most ATS systems filter on keyword matches, missing qualified candidates who describe skills differently and passing unqualified candidates who keyword-stuff.
Eightfold AI demonstrated skills-based matching that surfaces "hidden gem" candidates by analyzing trajectories and transferable skills. Paradox's Olivia handles conversational screening at scale — Workday acquired Paradox in October 2025 for this capability. HireVue processes data from over 70 million interviews using fine-tuned language models. As of 2025, 41% of recruiters used AI daily for sourcing and screening.
The challenge: most implementations are bolt-on tools that do not integrate with existing ATS workflows or provide explainable ranking decisions.
How this
agent works
This agent ingests job descriptions and builds structured role requirement profiles: required skills, preferred skills, experience level, and role-specific competencies. When applications arrive in your ATS, it parses resumes and extracts candidate skills profiles using semantic mapping — recognizing "React development" and "building UIs with React.js" as the same competency.
Each candidate is scored with a transparent breakdown: requirements met, partially met, and gaps. Scoring accounts for transferable skills and career trajectory — a candidate with 3 years of Vue.js and strong component architecture may score well for a React role despite the framework mismatch.
Top-ranked candidates receive automated phone screens via conversational AI. Role-specific questions, transcription, and summary generation. Hiring managers receive ranked shortlists with skills match scores, screening summaries, and recommended interview focus areas.
Built on Python and FastAPI, with Anthropic Claude handling semantic skills matching and structured phone screen conversations. Candidate responses are transcribed, scored against role-specific rubrics, and stored in PostgreSQL; Redis manages session state across multi-turn phone screens via Twilio. We spend the first week calibrating a skills taxonomy to your industry and role families, then wire bidirectional sync to your ATS — Greenhouse, Lever, Workday, or iCIMS. Setup runs 2–3 weeks from scoping call to first live screen.
- 01
Semantic Skills Matching
Claude maps candidate experience to role requirements at the concept level — not string matching. It recognizes that 'built CI pipelines' satisfies 'DevOps experience' and flags transferable competencies from adjacent roles. Each match is scored with a confidence level and a plain-language explanation.
- 02
Automated Phone Screening
Twilio-powered conversational screens run on a role-specific question set your team defines during setup. Responses are transcribed in real time, summarized by Claude, and attached to the candidate record. Candidates get a consistent, structured experience regardless of recruiter bandwidth.
- 03
Explainable Ranking
Every shortlist position includes a structured breakdown: requirements fully met, partial matches with gap detail, and transferable skills that were counted and why. Recruiters can audit any decision without opening the original CV — reducing bias reviews from hours to minutes.
- 04
ATS Integration
Bidirectional sync with Greenhouse, Lever, Workday, and iCIMS via their native APIs. Screened candidates, scores, and transcripts write back into your existing pipeline stages. No duplicate data entry, no parallel spreadsheets — the agent fits inside your current workflow.
Build this agent
for your workflow.
We custom-build each agent to fit your data, your rules, and your existing systems.
Free 30-min scoping call