The quality of your prompts directly impacts the quality of the AI's output. Here are some proven patterns:

Be specific about context

Instead of a vague request, give the AI real detail about your situation:

State your desired output

Share constraints

Ask for options

Iterate and refine

Use follow-up questions

The AI maintains context throughout a conversation. Build on what's been discussed rather than starting over.

Quick start prompts

You don't need a perfect prompt. Each skill walks you through a conversation, so a simple opening works. The AI will ask you for what it needs.

• "Help me build a people roadmap for next year"
• "I have survey results to analyse" (upload your CSV or Excel)
• "Design a competency framework for my engineering team"
• "Plan a facilitation session for our leadership offsite"
• "I need help building a job description for a People Ops Manager"
• "Help me create engagement survey questions"
• "Help me tell a data story about our attrition problem"
• "Write a job advert for a Senior Engineer at our startup"

Full prompt examples by skill

These are power-user prompts. They front-load your context so the AI can skip the introductory questions and get to high-quality output faster. Paste and adapt them. Upload files where noted.

Roadmap Builder: mirrors the workflow: current roadmap ideas, business goals/OKRs, engagement and culture inputs, then prioritised roadmap.

I need a people roadmap for the next 12 months.

Context: 120-person Series B, product-led SaaS, UK and US. People team of 2.

What I’m already considering: manager training, leveling up our performance process, and a small DEI learning series.

Business: Q2 OKRs focus on NRR, reducing churn in mid-market, and shipping a new enterprise tier. Leadership wants “fewer HR projects, more revenue impact.”

Employee signals: last pulse showed manager support scoring lower than company average; exit interviews mention unclear growth paths; culture diagnostic flagged decision-making speed.

Please help me prioritise, map items to OKRs, and call out want vs need trade-offs. I want a quarterly view and talking points for sensitive deprioritisations.

Survey Analyser: upload CSV/Excel when you have it; this prompt matches the analyse → themes → actions flow.

I’ve uploaded our Q4 engagement export (CSV). Survey was company-wide, ~68% response rate.

Columns include: department, tenure band, Likert scores for manager effectiveness, workload, and two open-text “what would you change?” fields.

Please: (1) summarise key themes and differences by department and tenure, (2) flag risks for exec attention, (3) give a concise action plan in a table (priority, action, timeline, data rationale, expected impact), (4) suggest 3 follow-up questions I should ask leaders in a workshop next week.

Survey Question Creator: define the purpose, audience, and what you want to measure; the AI will design questions that avoid bias and maximise response quality.

I'm designing a pulse survey for our product engineering team (~45 people, split across 6 squads).

Purpose: understand whether recent changes to our ways of working (async-first, no-meeting Wednesdays, quarterly OKR cycles) are landing.

I want to measure: psychological safety, clarity of expectations, manager support, and workload balance. Include 2-3 open-text questions that actually surface useful themes rather than generic "any other feedback?" prompts.

Constraints: keep it under 15 questions, mix of Likert scale and open text. We're using Culture Amp so questions need to work in that format. Response rate target is 80%+, so brevity matters.

Competency Framework Builder: role context, levels, behaviours; reference uploads if you have JDs or an existing matrix.

Design a competency framework for our Engineering organisation.

Context: IC and manager tracks, 4 levels (Associate → Principal / EM). We care about ownership, communication with non-technical stakeholders, and technical judgment—not a laundry list of languages.

I’ve uploaded our current senior engineer JD and a rough manager expectations doc—use them but don’t copy verbatim.

Output: a matrix-style framework (competencies as rows, levels as columns) with 2–4 bullet behavioural indicators per cell, plus short guidance for how managers should use it in reviews.

Facilitation Designer: purpose, audience, constraints, then a timed run-of-show.

Design a 2.5-hour facilitation plan for a leadership offsite (8 people: CEO, CFO, CPO, VPs).

Goal: align on people priorities for H2 without boiling the ocean. Outcomes: (1) shared view of top 3 people risks tied to business goals, (2) each exec commits to one visible action, (3) clear owners and dates.

Constraints: hybrid (5 in room, 3 remote), minimal slides, one short break. Include purpose, participant prep, materials, detailed agenda with timings, facilitation prompts, and how to handle domination / silence.

Tone: direct, practical, slightly irreverent—our culture hates corporate workshop speak.

Data-driven Storymaker: audience, narrative, and evidence; output structured for exec consumption.

Help me build a data-driven story for our next exec meeting.

Audience: CEO and CFO—very metrics-led, sceptical of “culture” language.

What I have: eNPS trend 4 quarters, regrettable attrition by level, time-to-fill vs plan, and two qualitative themes from exit interviews (unclear expectations, manager consistency).

Ask me only for missing numbers if essential. Deliver: (1) a tight executive summary in their language, (2) 5 slide titles with bullet talking points each, (3) likely objections and how to respond, (4) one “ask” that is specific and bounded.

Job Description Generator: role, level, business context, inclusivity and tone.

Write an internal job description for a People Operations Manager (IC), UK remote-first, reporting to me (Head of People).

Company: 90 people, high growth, mostly engineers. Need someone who can own HRIS/data hygiene, benefits admin, and manager enablement on people processes—not a generalist who only runs events.

Requirements: spell out must-haves vs nice-to-haves, avoid biased language, include a short “what success looks like in 90 days” section, and keep total length readable on mobile.

Job Ad Generator: write an authentic public job advert for a startup: honest about tradeoffs, inclusive language, and a voice that sounds like us—not generic corporate copy.

Write a public job advert for a Senior Backend Engineer at our startup (~35 people, Series A, fully remote EU).

Be honest: we're still figuring out on-call, our codebase has legacy corners, and pace is high. Say what's genuinely great (ownership, product impact, kind teammates) and what's hard (context switching, thin process in places).

Tone: direct, warm, inclusive—no "rock star" or "ninja" language. Structure: hook, role snapshot, what you'll do, what we're looking for (must-haves vs nice-to-haves), compensation/transparency note if we're comfortable saying "discuss openly," and how to apply.

Keep it scannable on mobile; avoid laundry lists of technologies unless essential.

Why these prompts work

The best prompts front-load four things:

• Context - who you are, your company stage, team size, and what you've already tried
• Constraints - timeline, budget, headcount, tools you're already using
• Audience - who will consume the output and what language they respond to
• Format - be explicit about the deliverable (Word doc, matrix, slide deck, action plan table)

Prompt checklist

• Role - who you are (e.g. solo People lead, consultant, manager) and what decision this supports
• Audience - who will read or hear the output (execs, managers, all-staff)
• Data attached - what files you uploaded or what columns/segments exist in your data
• Tone - formal, plain, bold, brand voice notes from Account Settings
• Success criteria - what "good" looks like and what to optimise for (speed, depth, buy-in)
• Format - bullets, table, Word artifact, workshop plan, slide outline, etc.

When you're ready, open Open Org AI and start with the skill you need. Paste a block like the examples above and refine from there.