ChatGPT Prompt Library for Work Tasks

A good ChatGPT prompt library is a set of reusable instructions for recurring work, not a giant list of clever one-line commands. The most useful prompts give ChatGPT enough context, define the required output, and tell it how to handle uncertainty.

This library is organised around common work tasks: research, writing, analysis, planning, customer communication, meetings, and coding. Copy a template, replace the text in square brackets, then check the result before using it.

If you want a broad directory of community-created prompts and roles, see our Awesome ChatGPT Prompts directory. This article takes a different approach. It focuses on prompts that teams can standardise for real business work.

How to use these ChatGPT prompts

Every useful prompt has four parts:

1. Context: What is happening and who is involved?
2. Task: What do you need ChatGPT to do?
3. Constraints: What must it include, avoid, or assume?
4. Output format: How should the answer be structured?

For example, “Write a customer email about our delayed project” is a weak prompt. It leaves too much open to interpretation.

A stronger version is:

Write a clear email to [customer name] explaining that the [project name] delivery date has moved from [old date] to [new date].

Context: The delay is due to [reason]. We have already completed [completed work].

Tone: Calm, accountable, and practical. Do not sound defensive.

Include: A short explanation, the revised plan, what we need from the customer, and an invitation to discuss concerns.

Keep it under 220 words.

The second prompt is easier to reuse and produces something closer to what you need.

For regular work, save your best prompts in a shared document, knowledge base, or project workspace. Give each one a name, explain when to use it, and include a good example output. That turns a prompt into a usable operating asset rather than a one-off chat.

Research and information gathering prompts

ChatGPT can help structure research, identify gaps, and turn material into a brief. It should not be your only source for facts that affect money, legal obligations, health, security, or business decisions.

Research brief template

I am researching [topic] for [audience or decision].

Create a research brief that covers:

• The core question
• Key terms and concepts
• The main arguments or options
• Important assumptions to test
• Information that needs a current source
• Questions I should ask before deciding

Separate confirmed information from reasonable assumptions. If you are unsure, say so clearly.

Use this before you begin a project or meeting. It gives you a map of the subject and makes hidden assumptions visible.

Compare options template

Compare [option A], [option B], and [option C] for a [type of business or team].

Assess each option against these criteria: [criteria].

Present the answer as a table, followed by:

1. Best fit for a small team
2. Best fit for a larger organisation
3. Risks or trade-offs that could change the recommendation
4. Questions we need to answer before choosing

Do not invent product capabilities, pricing, or customer evidence.

This is useful for software selection, vendor shortlists, operating models, and project approaches. The last instruction matters. ChatGPT can write a convincing comparison even when its underlying information is incomplete.

Source review template

Review the following notes and sources about [topic].

For each claim, classify it as:

• Supported by the supplied material
• Plausible but needs verification
• Unsupported or contradictory

Then list the three most important facts to verify independently.

Material:
[paste notes, excerpts, or links]

This prompt helps reduce a common failure mode: treating a polished summary as evidence.

Writing and editing prompts

Writing is one of the best uses for ChatGPT, particularly when you already know the point you need to make. Give it raw notes, audience context, and a definition of “good”.

First-draft template

Write a [format] for [audience] about [subject].

Goal: [what the reader should understand, decide, or do].

Key points to include:

• [point 1]
• [point 2]
• [point 3]

Tone: [direct, formal, friendly, technical, concise].

Avoid: [jargon, hype, overly sales-focused language, unsupported claims].

Structure the draft with [headings, bullets, short paragraphs, call to action].

Before writing, list any missing details that would materially change the draft.

That last line is valuable for high-stakes messages. It encourages ChatGPT to ask for information rather than filling the gaps with guesses.

Rewrite for clarity template

Rewrite the text below for [audience].

Keep the meaning and factual claims intact. Make it clearer, shorter, and easier to scan.

Use plain English. Remove repetition and vague language.

Show:

1. The revised version
2. A short list of changes that could alter meaning and need my approval

Text:
[paste text]

Use this for internal updates, proposals, policy documents, and technical explanations. It is safer than asking for a rewrite with no guardrails.

Content outline template

Create a detailed outline for a [blog article, guide, presentation, report] on [topic].

Reader: [describe them].

Their main question is: [question].

Include the practical steps, common mistakes, and decision points they need.

Do not add claims that require evidence unless you label them as needing a source.

End with a suggested checklist.

For more practical learning resources, browse the Enterprise DNA guides and learning resources. A strong outline should help someone take action, not simply create more content.

Analysis and decision-making prompts

ChatGPT is useful for organising messy information. It can turn notes, transcripts, spreadsheets, and feedback into a structured view. It cannot replace validation of the underlying data.

Analyse feedback template

Analyse the customer feedback below.

Group the feedback into themes. For each theme, provide:

• The issue or request
• How often it appears in this sample
• Example wording from the feedback
• Likely impact on the customer experience
• A suggested next action

Distinguish between direct evidence and your interpretation.

Feedback:
[paste feedback]

If the feedback contains sensitive data, remove names, email addresses, contract details, and anything your organisation should not place in a third-party system.

Decision memo template

Help me prepare a decision memo about [decision].

Context: [background].

Options under consideration: [options].

Decision criteria: [criteria].

Constraints: [budget, deadline, people, risk tolerance].

Create:

1. A concise problem statement
2. Options with benefits, drawbacks, and dependencies
3. A recommendation based only on the information supplied
4. Open questions that could change the recommendation
5. A one-page memo draft for leadership

This is especially good when a team has many opinions but no shared decision frame.

Spreadsheet formula prompt

I need help writing a [Excel or Google Sheets] formula.

My goal is: [describe the calculation].

My columns are:

• [column name]: [example value]
• [column name]: [example value]

Rules:

• [business rule 1]
• [business rule 2]

Give me the formula, explain each part, and provide three test cases including an edge case. Do not assume column names or data formats that I have not provided.

Always test formulas against known records before applying them across a full model. One wrong assumption about dates, blanks, or text values can produce a very tidy but incorrect result.

Planning and project delivery prompts

A prompt library becomes valuable when it supports work that happens every week. Project planning is a good example.

Project plan template

Create a practical project plan for [project].

Objective: [objective].

Deadline: [date].

Team and roles: [people or functions].

Known constraints: [constraints].

Deliverables: [deliverables].

Build a plan with:

• Workstreams
• Milestones
• Dependencies
• Key risks
• Decisions required
• A weekly update format

Flag anything that appears unrealistic based on the time, scope, or resourcing provided.

The aim is not to accept the plan untouched. It is to get a first pass that makes dependencies and missing decisions easier to see.

Meeting preparation template

Help me prepare for a meeting with [person or group] about [topic].

Meeting goal: [goal].

Background: [context].

Known concerns: [concerns].

Create:

1. A short agenda
2. Five questions that move the discussion forward
3. Likely objections or questions
4. Suggested responses based only on the information provided
5. A follow-up action list template

Meeting notes to actions template

Turn these meeting notes into an action register.

For each action, show:

• Action
• Owner
• Due date
• Dependency or blocker
• Decision required

Then write a short summary of the meeting for attendees. Mark unclear owners and dates as “to confirm”.

Notes:
[paste notes]

This is much better than asking for “meeting minutes”. You specify the output that your team actually needs to run the work.

Customer communication prompts

Customer-facing work needs human judgement. ChatGPT can help prepare a draft, but the person responsible for the relationship should always review it.

Customer response template

Draft a reply to a customer who has raised this issue: [issue].

Our position is: [position].

Facts we can confirm: [facts].

What we cannot promise: [limits].

Desired outcome: [outcome].

Tone: respectful, direct, and accountable.

Do not admit fault, offer compensation, or make contractual commitments unless explicitly stated above.

Provide a subject line and a draft under 180 words.

Explain a complex issue template

Explain [technical or operational issue] to a non-technical customer.

Use plain language and short paragraphs.

Cover what happened, what it means for them, what we are doing next, and when they will hear from us again.

Avoid internal jargon and speculation.

If a fact is not confirmed, label it as unconfirmed.

Coding and technical work prompts

For people using ChatGPT alongside Codex or the OpenAI API, prompts should be precise about the environment, acceptance criteria, and safety boundaries.

Code review template

Review the following code for correctness, readability, security concerns, and likely edge cases.

Language and framework: [details].

Intended behaviour: [details].

Constraints: [performance, compatibility, security, style].

Return:

1. Critical issues first
2. Suggested fixes with explanations
3. Test cases that would catch each issue
4. Any assumptions you made

Do not claim the code is production-ready without qualification.

Code:
[paste code]

Treat the output as a review input, not a sign-off. Run tests, inspect dependencies, and have an accountable engineer make the final call.

API integration template

I am integrating the OpenAI API into a [application type].

My goal is: [goal].

Technical environment: [language, framework, hosting].

Requirements: [requirements].

Constraints: [privacy, rate limits, budget, latency].

Propose an implementation approach with:

• Request and response flow
• Error handling
• Logging and monitoring
• Prompt versioning
• Evaluation tests
• Risks requiring human review

Ask clarifying questions where requirements are incomplete.

If you are moving from experiments to business processes, the important work is often outside the prompt itself. Permissions, evaluation, escalation, audit trails, and process ownership all need attention. Our Enterprise DNA insights cover practical operating questions around this shift.

How to check ChatGPT results before using them

A prompt can improve output quality, but it cannot guarantee accuracy. Build review into the workflow.

Use this simple check:

1. Check the facts. Verify dates, figures, names, citations, product details, and claims against reliable sources.
2. Check the assumptions. Ask what information the answer inferred rather than received.
3. Check the omissions. Look for stakeholders, risks, alternatives, or edge cases that were left out.
4. Check the tone. Make sure customer or employee communications sound like your organisation.
5. Check the consequence. The higher the cost of being wrong, the more human review the result needs.

You can ask ChatGPT to inspect its own draft, but that is not independent verification. A useful follow-up prompt is:

Review your previous answer as a critical editor. List every factual claim that needs independent verification, every assumption you made, and any important information missing from the prompt.

It will not catch everything, but it creates a practical review list.

Build a prompt library your team will actually use

Start small. Pick five recurring tasks that consume time or create inconsistent output. Create one prompt for each task, test them with real examples, and revise them after use.

For each saved prompt, include:

• A clear name
• The business task it supports
• When to use it
• Required input information
• The full prompt template
• A strong example output
• A named owner who keeps it current

Avoid treating prompts as fixed scripts. Your team should improve them as processes change. A prompt that worked for a small team may need more controls when it becomes part of a customer workflow or an internal operating process.

If you are considering how ChatGPT fits into larger business workflows, Omni and Omni Ops show how structured workflows, operational controls, and human oversight can sit around model output.

For tailored help designing practical prompt workflows for your team, you can book a call with Sam.