You cannot hire your way to scale
Every owner trying to scale eventually says the same thing. “I just need to hire better people.” It is the wrong diagnosis.
Hiring one person well is not the same job as hiring at three locations, with consistent rigor, where each hire either reinforces your brand or quietly erodes it. Volume, consistency, and continuous calibration are the work of scaling the people side of a practice. Those three things are work humans cannot do alone.
Your hiring manager reading every application is a fantasy past 30 candidates. Your interviewers producing consistent evaluations across three sites is a fantasy without infrastructure. Your scorecard staying current is a fantasy if you have no time to analyze your last 12 hires against their six-month performance.
AI is what closes the gap. Not by replacing the hiring manager. By doing the work the hiring manager cannot do at scale.
Pre-screening at volume
You post a Lead Injector role across three locations. 217 applications come in. Some from your ATS, some from referrals, some from your existing team. They sit in three different inboxes.
Your hiring manager has a real job. They read 30 carefully and skim the rest. The strong candidate who buried their specifics on page two of the resume gets passed over. The polished candidate with three short tenures in a row gets a callback because the resume reads clean. This is not your manager failing. It is the job being structurally impossible at volume.
Claude does this work differently. You feed it your scorecard with weighted criteria — clinical fit, brand fit, retention signal, flags. You feed it every application. It reads every one against every criterion, ranks the pool, and surfaces the top 15 with one-line strengths, gaps, and flags. Your manager gets a ranked list and a reason for the ranking. Every applicant got evaluated. Your manager only meets the ones worth meeting.
The full pre-screening prompt is below in Section 7.
Personalized onboarding
Most onboarding plans are the same plan, handed to every new hire. The deck, the shadowing week, the SOPs, the role-specific protocols. The senior injector with five years experience sits through consent and charting basics. The new grad gets thrown into consultations on day one. The provider from a competitor never gets real brand and protocol immersion because the manager assumed they would pick it up.
AI changes the input-to-plan logic. You feed Claude the role profile, the new hire’s resume, and your brand and protocol overview. Claude produces a 30-60-90 day onboarding plan that skips what this hire already knows and reinforces what they do not.
The senior injector gets day one on your specific consult flow. The new grad gets two weeks of scaffolded consultations. The provider from a competitor gets a structured brand and protocol track on top of the role-specific work.
Same role, different starting points, same outcome at day 91. The manager does not customize. The AI does. The full prompt is in Section 7.
Continuous calibration — the scorecard that learns
The hiring scorecard you wrote two years ago is no longer the right scorecard. Some criteria predicted high performers and some did not. Some red flags you weighted heavily turned out to be noise. Some questions you ask every interviewer stopped producing signal six months ago.
Most owners do not update because there is no time to analyze which criteria worked. Pulling the last 12 hires, lining up their six-month performance against their interview scores, and figuring out which interview criteria actually predicted retention is real analytical work. It does not get done.
AI does it on demand. You feed Claude your current scorecard and your last 12 hires with their six-month performance ratings. Claude returns which criteria predicted retention, which red flags were noise, which questions stopped producing signal, and a recommended scorecard update. You run the next cycle better.
Your scorecard learns. Most owners’ scorecards do not.
Verification and reference checks
The hidden tax on hiring is the work nobody likes. Reference calls that go to voicemail and stay there. License verification across multiple state boards. Background and credentialing logistics. The structured reference questionnaire that no one has time to write fresh for every hire.
Three hours per hire of manager time on logistics rather than judgment. At ten hires a year that is thirty hours of senior leadership time spent on phone tag.
AI converts this into structured infrastructure. Reference questionnaires generated and sent as templated emails the reference can complete asynchronously. State board credential verification pulled from public databases. License expiration tracked on a rolling basis with renewal reminders. The manager interviews. AI handles everything before and after.
The reference questionnaire prompt is in Section 7. It generates a structured 8 to 10 question questionnaire tailored to the role and the candidate’s claimed strengths.
The platforms behind the prompts
The aesthetic industry does not have a true all-in-one for the people side of the business yet. The PMS platforms handle scheduling and onboard you to their software, not your business. The minimum stack assembled from horizontal tools:
Workable. ATS, structured interviews, scorecards as data instead of Word docs.
Whale. AI-native SOPs, training, and onboarding tracks. Real multi-location aesthetic case study with Aurora Medical Spa.
Trainual. The established alternative for SOPs and training delivery. Role-based playbooks, less AI-forward.
Scribe and Tango. Screen-recording to instant SOP. Your top performer does the task once, the SOP is documented.
MedTrainer. Healthcare-specific compliance and credentialing. The only aesthetic-relevant compliance platform in the stack.
Illume (formerly Illume). Provider-facing real-time KPIs and benchmarks. Managers coach against the trend instead of waiting for a quarterly review. The data layer that makes continuous calibration possible.
Lattice. Performance reviews and manager development.
The platforms are the infrastructure. The Claude prompts are how you put them to work.
The three Claude prompts
Paste-ready. Replace the bracketed text with your specifics.
Prompt 1: Pre-screening at volume
You are a hiring analyst for an aesthetic practice. Below is the role description and our hiring scorecard with weighted criteria. Below that is the applicant pool (resumes, cover letters, or both). Score every applicant against every criterion (1-5), produce a weighted total, and rank the pool. Surface the top 15 candidates. For each, provide a one-line strength, a one-line gap, and any flags (short tenure pattern, license issues, specialty mismatch). Do not infer beyond the data. [PASTE ROLE + SCORECARD + APPLICANTS]
Prompt 2: Onboarding personalization
You are an aesthetic practice operations lead. Below is the role profile, the new hire’s resume, and our brand and protocol overview. Build a 30-60-90 day onboarding plan for this specific hire. Skip what they already know based on their resume. Reinforce what they do not. Output as three columns (Days 1-30, 31-60, 61-90) with milestones, training modules, and competency checks per column. Plain language a clinical director can hand to the new hire on day one. [PASTE ROLE + RESUME + BRAND OVERVIEW]
Prompt 3: Reference questionnaire generator
You are conducting a structured reference for a new injector hire. Below is the role description and the candidate’s claimed strengths and prior responsibilities. Generate a structured reference questionnaire with 8-10 questions designed to verify the strengths, surface gaps the candidate did not disclose, and elicit specific behavioral examples. Output as a numbered list the reference can complete by email. [PASTE ROLE + CANDIDATE CLAIMED STRENGTHS]
The work that breaks at scale
The owners who scale past three locations are not better at hiring. They are running AI on the work that breaks at scale — volume, consistency, calibration, verification.
The variance work in Part 1 of this series and the lead infrastructure in Part 2 are what AI does for your numbers. This is what AI does for your team.
Part 8 of The Real Stack covers six dimensions of scaling with AI. This is the third. Coming next: margin leak per location, the slow losses that compound across sites.
Read the full series at theaudreyaesthetic.com.