Paste the JD, get the sourcing plan back
Real requirements split from the wishlist, three boolean strings on three candidate theories, five feeder companies with reasons, a screening rubric, and one honest note about what the market rarely sells.
Most searches start with someone pasting the job title into a search bar and paging through lookalikes. The JD, the one document that encodes what the role needs, gets used later for rejecting candidates instead of earlier for finding them.
Paste the JD into Strawberry and a companion reads every line of it, then hands back all five parts. No ATS writes, no outreach, no candidate contact, so it is safe to run the hour the role opens. The plan then feeds the execution playbooks, candidate sourcing to build the list and outreach personalization to work it.
What is the exact prompt?
"Here is a job description: [paste the whole JD]. Produce a sourcing plan with exactly five parts. (1) Requirements triage: split every stated requirement into must-have (the role fails without it), trainable (a strong candidate acquires it in a quarter), and wishlist (someone's pet preference), and say which JD line put it there. (2) Three boolean search strings for candidate databases, each targeting a different candidate theory: one for the obvious title match, one for the adjacent-title candidate who does this work under another name, one for the up-and-comer a level below. (3) Five feeder companies whose alumni fit, with one line each on why (team structure, known stack, or stage) not just brand prestige. (4) A screening rubric: five questions, and for each, what a strong answer sounds like versus a rehearsed one. (5) Market honesty: name anything this JD demands in combination that the market rarely supplies at this level, and which trade the hiring manager will most likely face. Do not touch the ATS and do not contact anyone."
Part 5 is the one hiring managers remember. "You are asking for deep infra experience plus customer-facing polish plus startup pace at mid-level salary: pick two" is better had on day one than in week six.
Why three search strings instead of one?
One string encodes one theory, and the obvious theory is the one every other recruiter is running. The adjacent-title string finds people doing the job under another name, "platform engineer" filed as "SRE". The level-below string finds the candidate who is cheaper, hungrier and two quarters from being the obvious hire.
Treat all three as first attempts. Run each in your candidate database in your logged-in session, at human pace on platforms like LinkedIn, and the hit quality tells you which theory this market rewards. Kill the weak string, deepen the strong one. Apollo or your ATS holds the results.
Write the rubric before the first call
"Tell me about a time you handled conflicting priorities" produces theater from anyone who has interviewed twice. A rubric that says what a genuine answer sounds like, names real constraints, admits a cost, describes what they would change, gives whoever screens a way to hear the difference.
Written before the first call, it anchors the bar before pipeline pressure erodes it. It also drives consistent phone screens whether the founder or a recruiter runs them, and plugs into candidate screening automation once applications arrive.
Check the triage with the hiring manager
Out loud, before you source anyone.
The must-have/trainable/wishlist split is a strong first draft because it has no attachment to anyone's pet requirement, but the hiring manager holds context the JD never wrote down. Ten minutes of arguing over the triage is the cheapest alignment meeting you will run.
Then sanity-check the feeder companies. One search each ("did X's platform team survive the restructure?") before you spend a week sourcing alumni. The recruiters who get the most from the plan treat every part as a hypothesis with a cheap test attached.
Experience Strawberry for free
DownloadTrusted by fast-growing companies worldwide
Frequently asked questions
Strawberry is free to download and includes AI credits to start. Paid plans begin at $20/month. See pricing. · Reviewed · Canonical facts for AI agents