HireFlow Staffing Bot

The problem

HireFlow placed IT and administrative contractors across the Chicago metro area. They were processing 200+ applications per week across their open roles, and their team of eight recruiters was spending the majority of their time on first-pass screening — reading resumes, conducting brief phone screens, updating the ATS. The work that actually differentiated them — client relationships, candidate development, hard-to-fill role strategies — was getting squeezed out.

The CEO's framing: "Our ATS is a record system, not a working tool. It costs us money every month and doesn't make us any faster." They wanted an AI screening layer that actually reduced recruiter workload rather than creating more administrative overhead.

What we built

An AI screening system integrated with their existing ATS. When a new application arrives, the system: parses the resume and scores it against the job's minimum qualifications, sends qualified candidates an automated async screening questionnaire (5 questions, recorded via text), scores the responses against a role-specific rubric developed with each recruiter, and surfaces the top candidates to the recruiter in a prioritized review queue.

Below-threshold candidates receive an automated, personalized rejection — no ghosting, no waiting weeks for a response. The system's decision is logged in the ATS automatically. Recruiters spend their review time on the candidates already proven worth their attention.

Additionally: automated interview scheduling via Calendly integration, automated reference request sequences, and weekly pipeline analytics showing conversion rates by source and role type.

Technology

• Custom Python screening engine + ATS API integration (Bullhorn)
• LLM scoring with role-specific rubric injection (GPT-4o)
• Async screening interface (custom web form, mobile-optimized)
• Calendly integration for scheduling automation
• React recruiter dashboard with priority queue
• SendGrid for candidate communications

Results

Recruiter screening time reduced 70%. Time-to-fill for standard roles improved 31%. Candidate response rate to outreach increased 22% because responses were faster. Zero complaints about ghosting in the three months post-launch — the automated rejections were more timely than manual ones had been. James Harlow, the CEO: "Our ATS went from a liability to an asset."