PLAYBOOK
How to run a full hiring process without a Head of Talent.
Most founders skip these and pay for it later.
Not the job description. Not the responsibilities. The specific outcome that tells you the hire worked. If you cannot answer this in one sentence, the search is not ready to start.
Not the person who signs off. The person who will manage this hire on day one. If they cannot commit to two hours per week on this search, the search stalls.
List three. If you have more than three, none of them are actually non-negotiable. The brief needs to reflect these exactly.
Not what you want to pay. What the market pays. If the range is below market, the brief is misleading. [Tool 4 from AI Head of Talent gives you the fully-loaded cost before you decide this.]
Who is in the debrief? How many people have veto? If you do not define this now, the debrief becomes a debate instead of a decision.
Without a dedicated recruiter, the hiring manager has to own both the sourcing decisions and the candidate experience. Block two fixed hours a week, same time every week, and treat it as a non-negotiable meeting with yourself. Use the first 30 minutes to review new candidates, the next hour for screens, and the last 30 minutes to update the pipeline sheet. Anything that falls outside this block gets pushed to next week, not squeezed in ad hoc, or the search drifts.
Structure the process in four fixed stages: CV/profile review, a 20-minute screen, a panel round, and a decision meeting. Do not add stages as you go. Adding a stage mid-search because someone wasn’t sure is usually a sign the brief was unclear, not that the process needed to be longer.
With three people and no dedicated HR function, assign one clear owner per interview stage rather than everyone sitting in on everything. The hiring manager runs the screen. A peer on the team the hire is joining runs the technical or functional round. The most senior person available runs a short values/culture round. Each panellist tests a different part of the scorecard, never the same thing twice.
Debrief within 24 hours of the final interview, while the impression is fresh. Each panellist gives a score and a one-line reason before anyone else speaks, so the group doesn’t anchor on the first opinion voiced.
Speed loses candidates less often than people think, but slowness definitely does. Set a rule: a decision gets made within 48 hours of the final interview, or the search stalls and a competing offer wins. If the panel is split, the hiring manager makes the final call, since indecision by committee is the more common failure mode than a wrong call made fast.
Before the offer call, run the Pre-Offer Risk Audit [Tool 8 from AI Head of Talent] so the closing conversation addresses risk directly instead of finding out about a counter-offer after the letter’s already sent.
The highest-risk period for a new hire is the first month, not the first year. Set a specific first-week win, something small and visible they can ship by day 5, so they feel productive early rather than lost in onboarding admin. Run a 1:1 every week, not just when something’s wrong, and ask a specific question each time rather than a generic "how’s it going."
Build a lightweight early-warning checklist: missed 1:1s, no questions asked in week one, avoidance of a specific task type. Any of these on their own isn’t a crisis. Two or more together, in the first two weeks, is worth a direct conversation immediately.
The first thing to hand off is not sourcing, it’s scheduling and coordination. That’s the highest-volume, lowest-judgment task in the process, and it frees the most hours for the least risk. Next, delegate first-round screens to a trained team member once you have a scorecard tight enough that two people reach the same verdict independently.
Hold onto the brief-writing and the final decision the longest. These are the two points in the process where judgment matters most and where a bad call is most expensive to unwind later.
These are the frameworks I use inside every engagement. Run it yourself using the guide above, or I can run it for you.
Book 20 minutes →