Internal System Review · prepared for Tim

The machine behind the placements.

A walkthrough of the Turnkey Recruiting operating system — the ATS, the process, the numbers, and how the whole firm actually runs. Go tab by tab. Everything you see here is the real system, shown with live sample data.

FirmTurnkey Recruiting
TaglineExecutive Talent. Delivered.
SystemCustom ATS · v1
Operating modelAI-native · $0 marginal
Scroll to begin
01Overview

An AI-native retained executive-search firm.

Turnkey Recruiting places senior leaders — Director through C-suite — for companies between $50M and $10B in revenue (200–10,000+ employees), nationwide. Two people run it: an engine and a craft. The engine builds and operates the system; the craft runs the searches and owns the client relationships. The system below is what lets two people operate like a full search desk.

The Craft — Tim

Client-facing recruiter

Runs every search, screens candidates, manages the buyer relationship, closes the placement. The face of the firm. Never touches a command line — the system serves you finished outputs.

The Engine — Ricky

Builds + runs the system

Operates the ATS, drafts the documents, keeps the numbers honest, and carries finance-lane credibility — a CFO turned recruiter. The back office, invisible to clients.

The Goal

One placement, then repeat

H2 is deliberately one clean placement that proves the engine — then the same system runs the next search, and the next, at almost no added cost.

LanePositionWho buys
Finance & AccountingAnchor lane — "CFO turned recruiter" credibilityCHRO · VP of HR · President · COO · P&L owner
Mining / Minerals / Rare-earths + IndustrialTim's deep network — reshoring & capacity tailwindPresident · COO · P&L owner
SaaS / TechnologyGrowth-stage finance & operating leadersCHRO · VP of HR · COO

Clients: $50M–$10B revenue · 200–10,000+ employees · nationwide. The specialty-vehicle lane exists but is never led publicly.

02The Operating Model

Claude Code is the operator.

This is the part most search firms don't have. There is no bot fleet, no daemon, no scraper running in the background. Ricky talks to Claude Code; it reasons through the search in-session, drafts the documents, and writes every result into one system of record. The cost of all that reasoning is $0 marginal — it runs on a flat subscription, not a per-call meter. You get the leverage of a full back office without the headcount or the software bill.

The Engine · Claude + the ATS

Does the heavy lifting

  • Turns a job brief into a structured search, fee, and qualification check
  • Builds & ranks the target list; scores every candidate 0–100
  • Drafts dossiers, client-ready candidate cards, agreements, offers, invoices
  • Persists every change to one database with a full audit trail
  • Enforces the money & guarantee rules automatically — nothing slips
The Craft · Tim

Does the judgment

  • Approves the target list and the outreach voice
  • Runs the screen calls and reads the human signal
  • Decides who makes the shortlist and who gets passed
  • Manages the buyer, negotiates, and closes the placement
  • Owns the relationship that turns one search into the next
System of record

One database, single-writer

Every search, candidate, and dollar lives in recruiting.db. One writer, append-only history — so the record can never quietly drift or get double-edited.

Client data

PII stays in-house

Candidate details never leave our system. HubSpot stays the client/company CRM; the candidate database is a private, compounding asset that gets more valuable with every search.

Why custom

Cheaper than a SaaS ATS

Below ~30 concurrent searches, this custom spine beats Ashby/Greenhouse on cost — and the AI-native engine is the differentiator, not an add-on.

04Tim's Dashboard

Your whole desk, on one screen.

This is the tab you live in. Every active search, every candidate in play, what needs your attention next, what's on the calendar, and where each client stands — generated straight from the system, always current.

Control Panel — Tim · Client & Pipeline
Tim's control panel — searches, candidates, next steps, calendar, clients

Live sample data — three searches in flight: Northwind (shortlist), Redrock (active search), Lumina (post-placement).

1
Top statsActive searches, candidates in play, your time-to-shortlist target, lifetime placements.
2
Active searches & key candidatesEach search with its fee, buyer, due date, and your top-ranked candidates by fit score.
3
Next stepsThe system derives your action list — screen calls due, submit/pass decisions, shortlist selections.
4
Upcoming calendarEvery scheduled screen and client interview, pulled from the pipeline.
5
ClientsEach company, the buyer you're working, and the current search status — at a glance.
6
Always currentNothing is hand-typed. The board regenerates from the database every time.
05What Changes For You

Three things change — and one doesn't.

Here's the honest split. The system takes the research, writing, tracking, and paperwork off your desk. It sharpens the work you keep. And it never touches the part that's actually you — the conversations and the judgment. Read it at three levels.

Level 1 · Goes away

What you stop doing

The Sunday-night spreadsheet work, writing up candidate notes, chasing status, formatting decks, drafting paperwork, doing fee math. The 50–70% of the week that was never the craft.

Level 2 · Gets sharper

What gets better

Faster shortlists, client-ready output that's consistent every time, a briefing waiting before every call, and more searches at once — without hiring or more hours.

Level 3 · Unchanged

What stays all you

The client relationship, the candidate conversations, the read on fit, the close. Your name, your network, your reputation. The AI is invisible — clients experience Tim.

Level 1 — what comes off your plate

The low-judgment, high-drudgery tasks that historically eat most of a search consultant's week.

What you used to ownWhat happens now
Build the target-company / market mapThe engine drafts the universe of likely companies and titles for the role — you edit it, you don't start from a blank page
Write candidate write-ups after every screenYour call notes become a finished, client-ready candidate card — you tweak the tone, not the formatting
Format the shortlist deckGenerated on-brand and AI-invisible — you review, you don't build
Maintain the pipeline tracker / status spreadsheetThe system is the tracker — every move is logged once and never re-keyed
Write the weekly client status updateDrafted from live pipeline state — you read it and send it
Draft engagement letters, offers, consent & replacement lettersTemplated and auto-filled from the record — no blank-page drafting
Calculate the fee, installments & admin feeThe fee engine computes it — and re-computes on the actual accepted comp — so the math is never wrong
Remember guarantee windows & 24-month off-limits datesSet automatically at placement; the system watches the clock and flags it before it matters

Net effect: you stop being a part-time data-entry clerk, copywriter, and paralegal — and spend the day in conversations instead.

Levels 2 & 3 — sharper vs. untouched

Level 2 · Gets sharper

Same work, higher quality

  • Speed to shortlistResearch and write-ups were the bottleneck — collapsing them is what makes a 14–21 day shortlist real.
  • ConsistencyEvery card, update, and document reads professional — no "tired on Friday" variance.
  • Pre-call prepA briefing is assembled before every candidate and client call — you're never cold.
  • Fewer costly mistakesFee math, the no-work-before-signed gate, guarantee dates — enforced, not remembered.
  • More searches at onceThe admin tax per search drops, so your ceiling rises without hiring.
  • A compounding databaseSearch #10 reuses the work of #1–9 instead of starting blank.
Level 3 · Stays all you

What the AI never touches

  • The client relationshipTaking the brief and reading what the buyer actually needs vs. the job description.
  • The candidate conversationsScreening, reading a person, building trust.
  • Judgment on fitThe system scores it; you decide who's real and who gets submitted.
  • The closeOffer, counter-offer — the human stuff that makes or breaks the placement.
  • Advising the clientMarket read, comp guidance, when a candidate is worth stretching for.
  • Name, network & reputationThe AI is invisible — the client experiences you.

The line is simple: AI is your associate and your back office. It is not your face, and it is not your judgment.

06A Live Search

Inside one search, candidate by candidate.

Drilling into a single search gives you the ranked shortlist with the reasoning behind each score, the client-ready candidate card you actually send the buyer, the pipeline funnel, and the billing tied to this engagement. This is what turns "I have some people" into a defensible, fee-worthy shortlist.

Search Review — Northwind Industrial · VP of Finance
Single search review — ranked shortlist, candidate card, funnel, billing

Note: the candidate card is the client-facing artifact — polished, persuasive, and with the AI engine completely invisible.

1
Ranked shortlist (0–100 fit)Each candidate scored against this exact role, with a plain-English rationale you can defend to the client.
2
Client-ready candidate cardThe write-up the buyer sees — match table, considerations, no fingerprints of how it was built.
3
Pipeline funnelSourced → contacted → responded → screened → submitted, so coverage is obvious.
4
Billing this searchThe three installments tracked against the engagement — trued-up on actual accepted comp at placement.
07Commercial Terms

The money rules, built into the system.

These aren't on a sticky note — they're enforced in code, so the fee, the installments, the guarantee, and the off-limits clocks are always right. Here's the structure the ATS runs on.

Fee ladderBasisNotes
Open at 30%Cash compWhere we start every retained engagement
Fall back 25%Cash compNegotiation room without leaving money on the table
Floor 25%Base salaryFinal fallback basis
+ $2,500Admin feeCheeky, waivable
$5,000MinimumHard floor on any retained search

Cash comp = base + target bonus + guaranteed sign-on. Excludes equity, benefits, and discretionary pay.

Retained installments

1/3 · 1/3 · 1/3

Signingqualified shortlist (~day 45) → offer accepted. Cancel after we've started? ~two-thirds is owed — never a cash refund.

Guarantee

6 months · 12 for C-suite

If a placement doesn't stick, we re-run it. Excludes layoffs, restructures, role eliminations, and budget cuts — things outside the candidate's control.

Contingency

Only with 90-day exclusivity

We'll work contingent, but only with a signed 90-day exclusive — code-enforced. No racing five other firms to the bottom.

Off-limits

24-month no-solicit

We never poach a candidate we placed — the clock is set automatically at placement and protects the client we just served.

The one that protects us most

The final fee recomputes on the actual accepted compensation at placement — never the estimate. If the package lands higher than quoted, the fee follows it. The system trues this up automatically.

08Behind the Scenes · Ricky's Side

The back office, fully transparent.

You don't need to touch any of this — but here it is, so you can see the whole operation. Billing, system health, the AI engine itself, what's automated today versus what we'll turn on as volume justifies it, and the guarantee clock running in the background.

Control Panel — Ricky · Ops, AI & Health
Ricky's control panel — billing, system health, the AI engine, automation, guarantee watch

Reading it: invoices by installment · single-writer health · the AI engine on $0 marginal spend · live-vs-deferred automation · guarantee watch.

Live in v1

What the system does today

Search + candidate pipeline with the hard gate · fee engine, invoices & guarantee clocks · dossiers, fit-scoring & candidate cards · bulk import & pre-call briefs · Excel + dashboard export.

The Engine

Claude Code · $0 marginal · invisible

Operating model is Claude-as-operator, billed on a flat subscription, with zero AI footprint in anything a client ever sees. The leverage is real; the fingerprints are not.

09The Roadmap

What’s live now — and what’s next.

We start with what we already pay for. LinkedIn Sales Navigator, Apollo, and the email stack are in hand — so sourcing, outreach, and nurture go live from the first search, not "someday." Only billing-send and CRM-sync wait for real placement volume. Nothing new runs until a search justifies it.

PhaseCapabilityTurns on whenNet new $/moStatus
Phase 1 · Live nowLinkedIn Sales Nav sourcing · Apollo enrichment · bulk import · pre-call briefs · niche lead-gen ponds · candidate outreach · nurture/dripFrom the first live search~$150–200starting live
Phase 2 · ScaleStripe invoice send · e-sign · HubSpot sync · reporting cron · reverse-shoppingFirst placements / repeat clientsMostly $0*next

*Already in the stack at $0 marginal. Net-new at Phase 1 is ~$150–200/mo (Sales Navigator + Apollo). We only re-evaluate a commercial ATS above ~30 concurrent searches.

Nurture / drip

The gap most firms have

Keeping candidate and client pools warm automatically is something many desks never build. We own it as a Phase-1 capability, live now — top-of-mind without the manual grind.

Outreach posture

Discreet, never a blast

Senior-passive outreach stays low-volume, personalized, and human-gated — threaded, never auto-replied. Reaching an exec should feel like a 1:1, not a campaign.

Reverse-shopping

Lead with a standout exec

When we have a genuinely exceptional candidate, we shop them to target companies — the highest-open-rate door into a new client.

10AI-Driven Nurture

Staying top-of-mind, automatically.

Most desks let candidate and client pools go cold between searches. Ours doesn't. The system keeps every relationship warm with personalized touches it drafts itself — but nothing sends until a human approves it, and any reply hands the conversation straight back to you. Here's the loop.

01
Segment

Everyone we've touched sits in the database, tagged by lane, relationship, and status — passive candidates, lost finalists, dormant clients.

02
Trigger

A signal enrolls them: a finish on a search, a placement anniversary, a dormant buyer, or a new role that fits a silver-medalist.

03
Personalize

The engine drafts each touch from that person's record and full history — a real 1:1 note, never a template blast.

04
You approve

The draft posts to Slack. You tap approve, edit, or skip. Nothing reaches a candidate or client unreviewed.

05
Send & thread

It goes out low-volume and threaded, in your voice — the AI stays invisible. It reads like you wrote it, because you signed off on it.

06
Listen & loop

A reply stops the sequence and routes to you. No reply — the next touch is queued, then they return to the warm pool.

The guardrail that matters: every touch is drafted by the system, approved by a human, and sent as you. A reply always ends the automation and hands you a live conversation.

The stack behind it

All of it is already in the toolstack — no new platform to buy.

recruiting.dbThe system of record — segments, sequence state, and every prior touch.
Claude CodeDrafts and personalizes each touch and picks the next step. $0 marginal.
SlackThe approval gate — every send is human-tapped before it goes.
Email / InstantlySends and threads the touches; detects when someone replies.
HubSpotClient-side sync only. Candidate data stays private to us.

Cadence by track

WhoRhythmIt stops when
Passive candidateDay 0 reconnect → Day 6 → Day 16 → Day 30, then a light quarterly value touchThey reply, or opt out
Lost finalist / silver medalistRe-engage at ~30 & 90 days, and the moment a fitting role opensThey reply, or get placed elsewhere
Dormant client / past buyerMonthly value note · quarterly check-in · placement-anniversary noteThey re-engage or book a call

Cadence and volume are deliberately restrained — reaching a senior exec should feel like a personal note, not a marketing sequence.

11Your Read

What I'd love your eyes on.

You're the craft — this system exists to make your judgment go further, not replace it. Walk the tabs and tell me where it matches how you actually run a search, and where it doesn't.

A few questions to bring back

No prep needed — just react to what you saw.

The process map — does the six-stage flow match how you'd actually work a retained search?
The hard gate — signed + invoiced before any sourcing. Right call, or too rigid for some deals?
The terms — does the 30/25/25 ladder, the guarantee, and the off-limits clock feel market-right?
Your dashboard — anything missing you'd want to see the moment you log in?
The candidate card — is that the artifact you'd be proud to put in front of a CHRO?
First automation — which Phase-A piece should we switch on for the very first search?