The 5-Filter Hiring System

5 Filters. The A-Players Survive All 5.

Run them every time. Save yourself the rest of your career. Each filter teaches itself in the next 60 seconds.

Filter F1

Speed

Did the candidate respond inside your 48-hour window?

What it scores on

Time-to-response (hours)
Time-to-completion (hours)

How it looks in practice

Example

Applicant Maria submits her residential HVAC application at 10:47 a.m. Tuesday. She finishes the whole loop by 6:12 p.m. Tuesday - 7.5 hours end to end. F1 = 95.

Applicant Doug submits at 11:02 a.m. the same Tuesday. He completes the video by Saturday afternoon - 4 days, 5 hours. F1 = 32.

Why it matters

A-players have options. The minute they hit 'submit' on your application, they're applying to four other shops. If they wait three days to finish your loop, they're going to wait three days when a customer calls from a flooded basement.

Filter F2

Pattern

Do they fit the blueprint of who has won this role at your shop?

What it scores on

Tenure-fit (years in trade / years per shop)
Revenue-fit (avg ticket size / close rate, if reported)
Callback-rate fit (when available from prior employer)
Behavioral notes match (blueprint phrases vs. answers)

How it looks in practice

Example

Your A-player blueprint for a multifamily roofing lead: 3+ years on the truck, 1+ year as a crew lead, at least one prior role > 24 months.

Carlos has 6 years on the truck, 18 months as a lead at his last shop, and one previous job he stayed at for 4 years. Pattern fit: 4 of 4. F2 = 88.

Why it matters

The pattern beats the resume. Top performers in your shop look like each other in ways that are obvious in hindsight and invisible in a stack of resumes. We make the blueprint explicit, then we grade against it.

Filter F3

Synthesis

Live trade scenario. AI-graded on thinking, not on which path they picked.

What it scores on

Clarity - do they write a coherent plan, or list every possibility?
Prioritization - do they handle the customer relationship before diving into diagnostics?
Trade-off articulation - do they name the decision they're making and why?

How it looks in practice

Sample synthesis scenario - Residential HVAC service tech

A customer calls on a 96°F July afternoon. Their 12-year-old Trane XL16i is blowing warm air. Last tech (one of yours, three weeks ago) replaced the capacitor and the system ran fine for two weeks. Diagnostic on arrival shows low suction pressure, high superheat, normal subcooling. Customer mentions two small kids in the house including a 6-month-old. They have a home-warranty policy and are annoyed it's the second visit in a month.

What do you do, in what order, and why?

Why it matters

This is the literal scenario that runs on the residential HVAC role template. It's ambiguous on purpose - three of four candidate paths are defensible. Top candidates handle the customer relationship BEFORE they get into refrigerant theory.

Filter F4

Taste

Real customer email. Reply to it. Taste shows up in 60 seconds or it doesn't.

What it scores on

Tone - calm, direct, no defensive language
Empathy - name the customer's actual frustration
Professionalism - no slang, no over-apologizing
Brand alignment - does it sound like your shop?

How it looks in practice

Sample taste prompt

The candidate sees a real (anonymized) customer complaint email - the kind your CSR opens every Monday morning. The prompt: "Reply to this customer. The real reply. The one you'd actually send."

We don't print the full prompt on the public site - candidates would game it. The full version ships inside the role-template library you get when you sign up.

Why it matters

AI can write you twenty follow-up emails. Taste is the human judgment of which one your customer will read and which four will get them to unsubscribe. You either have it or you don't, and it shows up fast.

Filter F5

Reverse

The questions they ask YOU at the end of the interview. Scored against the A-tier rubric.

What it scores on

Decision-making depth (how decisions get made)
Success-at-90-days framing
Coaching / feedback orientation
Vs. C-tier signals: PTO, benefits, snacks

How it looks in practice

Candidate asked about...	Tier
PTO, benefits, free snacks	C
Pay structure, raise cadence	B
How decisions get made; how you handle an underperforming tech; what success looks like at day 90	A

Why it matters

If they ask about PTO and benefits, they're a C. If they ask about how decisions get made, how you handle a tech who's underperforming, what success looks like in 90 days - they're an A. The reverse interview has been the single most reliable filter at CCR.

Scoring

The Total Score

F1 through F5 roll up to a single 0-100 composite score. Default weights are calibrated against CCR's own hire history: F3 and F4 carry the most weight (they predict performance), F1 acts as a knockout floor (sub-50 = automatic decline), F5 is a tiebreaker when two finalists are close.

Weights are configurable per org. If your blueprint says reverse interviews predict tenure better than synthesis at your shop, you set the weights that way. The system grades; you decide what matters.

Deliverables

What You Get Back

The 5-Filter score is the headline number. The artifacts behind it are why you can run a 15-minute final interview instead of a 60-minute screen.

The candidate's synthesis video

Their answer to the trade-specific scenario. With auto-transcription so you can scan-read it in 30 seconds.

The candidate's written taste reply

The full reply to the customer-complaint email. Read it side-by-side with the score rationale.

Per-filter rationale, not just numbers

Every F3, F4, F5 score includes a one-paragraph LLM explanation. You see WHY the score is what it is - challenging it takes 10 seconds.

Calibration loop

Rate the AI's grades blind against your own. The system tracks drift. When drift exceeds your threshold, you re-tune the prompt before a bad grade matters.

Integrity

What It Isn't.

Some claims we explicitly do notmake, so you know what you're buying:

Not a personality test. We don't grade Big Five, DISC, or any psychometric. Those are poor predictors of trades performance and Brad has the scar tissue to back it up.

Not facial or voice analysis. We grade response content. The video is a delivery format so the candidate's thinking comes through; nothing in the rubric scores how they look or sound.

Not a black box. Every grade ships with a one-paragraph rationale. You can override it. The calibration loop measures how often you do.

Run the 5-Filter on Your Next Role.

$199/mo. Monthly. Self-serve. No 12-month contracts.

See Pricing Read the Full Talk First