For eight months, I couldn’t find good people.
I was running an HR SaaS startup — which, yes, is deeply embarrassing given what we do. We help companies hire smarter. And there I was, drowning in mediocre applicants, ghosting candidates by accident, and spending Sunday evenings rewriting the same job description for the fourth time.
Then I started using AI for every single part of the hiring process. Not as a gimmick. As actual infrastructure. Here’s exactly what changed — and the prompts I use.
The Problem With How Most Founders Hire
Most founders hire reactively. Someone quits, or the team gets too small, and suddenly you’re posting a job in a panic. You copy-paste a job description from a competitor’s careers page. You get 200 applications, skim 15, interview 4, and hire whoever seemed the least risky.
That’s not hiring. That’s luck dressed up as process.
AI doesn’t fix your judgment. But it gives you back the time to actually use it.
Prompt 1: The Job Description That Filters For You
Bad job descriptions attract everyone. Good ones repel the wrong people before they apply.
Here’s the prompt I use now:
Write a job description for a [role] at an early-stage B2B SaaS startup with [X] employees. This person will own [specific outcomes]. In their first 90 days they need to [specific deliverables]. The role is NOT a fit for someone who [list red flags]. Tone: direct and honest, zero corporate jargon. End with a short section called ‘You might not be a fit if…’ to help candidates self-select out.
The self-selection section alone cuts unqualified applications by 30-40%. Candidates who apply after reading it have already pre-screened themselves.
Prompt 2: Screening Questions That Reveal Real Operators
Most screening questions have obvious right answers. Candidates prep for them. These don’t.
Generate 5 written screening questions for a [role] candidate. The questions should reveal: (1) how they think through ambiguity, (2) whether they communicate clearly under pressure, and (3) whether they’ve actually done this work or just managed people who did it. Avoid any question with an obvious correct answer.
Send these by email before any call. People who don’t respond thoughtfully fail the first test. You’ll know immediately who’s worth an hour of your time.
Prompt 3: The Structured Interview Guide
My interviews used to go wherever the candidate took them. Structured interviews fix this — same questions, same rubric, every candidate.
Create a structured interview guide for a [role]. Include: 4 behavioural questions using the STAR format, 2 situational questions specific to early-stage B2B SaaS, and a scoring rubric for each answer with criteria for a 1, 3, and 5 rating. Focus on [key competency 1] and [key competency 2].
Prompt 4: The Offer Letter That Closes
Losing a candidate after a great interview because your offer was slow or confusing is the most demoralising thing in startup hiring. I’ve done it twice.
Write a warm, clear offer letter for a [role] at a startup. Compensation: [X]. Equity: [Y] options at [Z] strike price on a 4-year vest with 1-year cliff. Start date: [date]. Tone: excited but professional. Include a plain-English explanation of the equity package in 3 sentences. End with a deadline of [date] for acceptance.
First draft in 2 minutes. Personalised in 10.
What This Actually Changed
Before AI, I spent roughly 8-10 hours per open role per week on hiring admin. Now it’s about 2 hours. The other 6-8 hours go back to the business.
More importantly, my last two hires have been the best I’ve made. Not because AI picked them — I picked them. But because I had a real process that let me see clearly instead of reacting to whoever showed up.
Most founders think they have a talent problem. They actually have a process problem. AI gives you a professional hiring process without needing an HR team to build it.
Want All 50 AI Prompts in One Place?
Download the free NoFluffPlaybook AI Prompt Library — 50 prompts across hiring, ops, fundraising, and product. Copy-paste ready.