Automation · Workflow Design

My Dream Talent
Automation Stack

✦ Drew Costanza March 2026 5 min read

Screening 200+ applicants a week isn't scalable — not if you actually want to give each candidate a real experience. Here's the end-to-end automation workflow I'd build if I had full control of the stack.

End-to-end automation workflow
01
Top of Funnel
Source, screen & filter
t
tofu
Resume screening & fraud detection on inbound applicants
Screening
Gem
Outbound cadences, sourcing & pipeline management
Outreach
Qualified candidates advance
02
AI Screening
Async video interview
ez
Ezra Recruiting
Candidates answer 3–4 structured questions on their own time. AI evaluates tone, clarity & fit — cuts 80% of screening load before I review a single name.
AI Interviewer
Shortlisted candidates move to human interview
03
Interview & Scoring
Zoom → auto-scored scorecard
Zoom
Interview transcript recorded automatically
Interview
Z
Zapier
Picks up transcript and routes it to ChatGPT on completion
Automation
AI
ChatGPT
Fills scorecard from transcript — key quotes, flags, follow-ups
Scoring
Offer decision made
04
Offer & Close
Automated offer + personal touch
🌱
Greenhouse
Offer approval triggers the entire downstream flow automatically
ATS
DocuSign
Offer letter sent immediately on approval — no manual steps
eSign
Personal video
Short video from hiring manager attached — the human moment that converts maybes
Closing
Candidate signs & accepts
Human touch preserved at every stage that matters

← Back to blog

The problem

Let me be honest: when you're managing 15–20 open reqs at a time, the math doesn't work. You simply cannot give every applicant a thoughtful experience, personally review every resume, and still spend real time with the candidates who matter — not without some help.

So here's the workflow I'd build. It's not about replacing the human touch. It's about protecting it — by offloading the repetitive stuff to tools that are genuinely good at it.

Stage 1 — Tofu + Gem

The moment a req opens, Tofu kicks off targeted outbound sequences to sourced candidates while simultaneously flagging inbound applications for fraud signals — duplicate submissions, mismatched experience, ghosted-offer patterns. Gem sits alongside it, keeping the pipeline organized and doing first-pass resume scoring so I'm only eyes-on the people who actually fit.

Stage 2 — Ezra Recruiting

Qualified candidates get invited to an async video screen through Ezra. They answer 3–4 structured questions on their own time, no scheduling back-and-forth. Ezra's AI evaluates tone, clarity, and fit signals. By the time I review, I'm watching a curated shortlist — not 40 cold calls. Candidates appreciate the flexibility. I appreciate getting an hour of my week back.

Stage 3 — Zapier + ChatGPT

When a Zoom interview wraps, Zapier picks up the transcript automatically and sends it to ChatGPT with a pre-built scoring prompt mapped to our scorecard criteria. Within minutes, the hiring manager has a structured summary — key quotes, green flags, questions worth following up on. No one has to take notes. Everyone has more context.

Stage 4 — Greenhouse + DocuSign + video

Offer approved in Greenhouse? DocuSign fires automatically. But here's the part I care about most: attached to that offer is a short, personalized video message from the hiring manager — recorded in advance, triggered at the right moment. It says: we picked you, and we mean it. That human moment at the end of an automated funnel is what converts a "maybe" into a "yes."

The bottom line

None of this replaces recruiting instinct or relationship-building — it just gets the noise out of the way so you can use those skills where they count. The best candidate experience isn't fully human or fully automated. It's the right blend of both.