The best AI prompts for business analysts, 22 to copy (2026)

A good prompt for a business analyst defines role, context, task, constraints and output format, and turns your notes into a reviewable draft in minutes. This library has 22 prompts to copy into ChatGPT, Claude, Copilot or Gemini: requirements, user stories, processes and stakeholder communication.

Add Founderz as a preferred source

What you’ll take from this

This guide covers what you need to use AI in your business analysis work, from the notes of a meeting to the final document.

• 22 AI prompts for business analysts ready to copy: requirements, user stories, process modeling, prioritization and stakeholders.
• Prompts built around what each tool does best: ChatGPT, Claude with long documents, Copilot inside Word, Excel, Outlook and Teams, and Gemini in Google Workspace.
• The method behind the prompts: the six fields that make a good prompt and a before and after example.
• How to build your own team prompt library and the limits you cannot skip: validation with stakeholders and data privacy.

If you want to move from trying loose prompts to using AI with a method in your work, the Online Program in AI and Innovation at Founderz is built for that. And to create a custom prompt in a minute, use the free prompt generator for business analysts.

ChatGPT prompts for business analysts

These are the base prompts I use for business analysis, sorted by task. Each one is written to copy and paste into ChatGPT and works in other tools with small changes. The text in brackets is an example: replace it with your case and keep the rest. Library reviewed in November 2026.

ChatGPT prompts for requirements and user stories

ChatGPT is the most versatile option for turning messy notes into structured documents. Paste the real material and always ask for a specific output format.

User story with Gherkin acceptance criteria

Act as a product owner with experience in [digital banking].
Context: [describe the business need, who has it and what constraints apply].
Task: write a user story in the format "As [role] I want [action] so that [benefit]" with three measurable acceptance criteria in Gherkin format (Given / When / Then).
Constraints: business language, no implementation details, one criterion per behavior.
Output format: the story on one line and the criteria in a numbered list.

Batch of user stories from discovery notes

Act as a senior business analyst.
These are my discovery notes with [the operations team]:
[paste the notes]
Generate the user stories that follow from them. Do not invent needs that are not in the notes; if something is ambiguous, flag it as an open question.
Return them in a table with columns: ID, story, acceptance criteria and estimated priority (high, medium, low), and add a list below of questions to validate with the stakeholders.

Business requirements from a meeting transcript

Act as a requirements analyst.
Context: elicitation meeting for [the CRM migration project] with [sales, support and finance].
Transcript:
[paste the transcript]
Task: extract the functional and non-functional business requirements. Distinguish what each person said from what you infer.
Output format: table with ID, requirement, type (functional or non-functional), stakeholder who asked for it and priority. Then three contradictions or gaps I should clarify in the next meeting.

ChatGPT prompts for process modeling and root cause analysis

For process analysis, AI structures what you already know and detects gaps. Validating the result with the people who run the process is still your job.

Map a process step by step

Act as a business analyst specialized in [financial operations].
Describe the [corporate customer onboarding] process step by step based on this description: [paste the description or notes].
For each stage give the responsible actor, decision point, system involved, and input and output document. If information is missing, write "to be confirmed" instead of assuming it.
Output format: numbered list and, at the end, a diagram in Mermaid syntax that I can paste into a diagramming tool.

Root cause analysis with the five whys

Act as a process improvement consultant.
Problem: [describe the problem with data: what happens, how much, since when].
Apply the Five Whys technique. For each why, explain the cause and its business implication, and say what data I would need to verify it.
End with a conclusion about the most likely root cause and two possible corrective actions.
Output format: a table with the columns Why, Cause, Implication and Data to verify it.

Current process versus target process (As-Is and To-Be)

Act as a business analyst.
Current process (As-Is): [describe the current steps].
Business goal: [cut approval time from 10 days to 3].
Task: propose a target process (To-Be) and compare it with the current one. Identify the steps that are removed, automated or change owner, and the risks of the change.
Output format: table with columns Current step, Proposed step, Change, Expected benefit and Risk.

Claude prompts for business analysts

Claude handles long documents well and produces careful writing, so it fits tasks where you paste a lot of material: long specifications, several documents at once, or a text that has to read well for senior management.

Review a long specification: contradictions and gaps

Act as a requirements reviewer with experience in [financial software projects].
I am pasting a complete specification:
[paste the document]
Task: review the whole document and find (1) requirements that contradict each other, (2) ambiguous or unmeasurable requirements, (3) terms used with different meanings and (4) cases the document does not cover.
Constraints: cite the exact section for each finding and do not rewrite anything yet.
Output format: table with section, problem, why it matters and a question for the author.

Draft a requirements document from several sources

Act as a senior business analyst. You will write the first version of a business requirements document for [project X].
Sources: 1) kickoff meeting notes [paste], 2) email from the sponsor [paste], 3) current process [paste].
Structure: objective and scope, stakeholders, functional requirements, non-functional requirements, assumptions, constraints and risks.
Constraints: formal tone, short sentences, do not invent data; mark anything missing as (to be confirmed) with the person to ask.
Output format: document with numbered headings.

Communicate a scope change to stakeholders

Act as a business analyst who must communicate a scope change to [the steering committee].
Change: [the billing module is delayed to prioritize the ERP integration]. Reason: [technical dependency found during analysis]. Impact: [cost, schedule and risk].
Task: write an email of at most 150 words. Start with what the reader must decide or know, explain the impact in one sentence and offer two options with your recommendation.
Tone: direct, professional and free of technical jargon.

Copilot prompts for business analysts

Microsoft Copilot works inside Word, Excel, Outlook and Teams with your own files, emails and meetings. Depending on your license and version, these features can vary: check what you have enabled before standardizing.

Copilot in Word: requirements draft from your notes

Use [the file Discovery_notes.docx] as the source and create a draft business requirements document.
Sections: objective, scope, stakeholders, functional requirements, non-functional requirements and open questions.
Write each requirement in one sentence with the verb "must" and an identifier (FR-01, FR-02...). Do not add information that is not in the file; mark gaps as (to be confirmed).
Formal tone, two pages maximum.

Copilot in Excel: prioritize requirements with MoSCoW

Analyze the requirements table in the [Backlog] sheet (columns: ID, requirement, stakeholder, estimated cost, business value).
Classify each requirement as Must, Should, Could or Won't based on value and cost, and add a column with a one-sentence justification.
Then create a summary with how many requirements fall in each category and the total cost of the Musts.
Flag the requirements with high value and high cost so I can discuss them with the stakeholders.

Copilot in Teams: minutes of an elicitation meeting

Using the transcript of the meeting [Logistics requirements elicitation], generate minutes with: decisions made, new requirements mentioned, unresolved questions and actions with owner and date.
Distinguish what was agreed from what was only proposed.
Format: table for the actions and bullet list for the rest. One page maximum.

Copilot in Outlook: summarize a thread and reply to the stakeholder

Summarize the email thread [Changes to the sales report] in five lines: each person's position, points of agreement and open points.
Then draft a reply to [the sales director] that confirms what was agreed, asks for the pending decision on [the scope of the report] and proposes a closing date.
Friendly and direct tone, under 120 words.

Gemini prompts for business analysts

Gemini integrates with Google Workspace, so it is convenient if your documentation lives in Docs, Sheets or Drive. These three prompts start from that idea: reference the document and ask for a clear format.

Gemini in Docs: rewrite a specification in a standard format

Rewrite the specification in this document in a standard requirements format: ID, description, priority, acceptance criterion and dependencies.
Keep the original meaning. If a requirement is ambiguous, leave it as it is and add a list of ambiguities at the end with a question for the author.
Use business language and avoid technical terms that do not appear in the original.

Gemini in Docs: use cases from a description

Act as a business analyst. From this system description [paste the description], write the specification of [3] use cases.
For each case: actor, preconditions, main flow step by step, alternative flows, postconditions and related business rules.
Do not invent functionality that is not in the description. Format: one heading per use case.

Gemini in Sheets: requirements traceability matrix

With the requirements in the [Requirements] sheet and the test cases in the [Tests] sheet, create a traceability matrix: one row per requirement and one column per test case, with an X where it is covered.
Mark in a separate column the requirements with no test case and the test cases that do not point to any requirement.
Add a count of covered and uncovered requirements at the end.

More business analysis prompts: elicitation, stakeholders and process analysis

These prompts are written for ChatGPT and work in Claude, Gemini and Copilot without changes. They cover what repeats most in a project: interviews, responsibilities, changes and reporting.

Questions for an elicitation interview

Act as a senior business analyst. I will interview [the head of customer service] about [the complaint handling process] for 45 minutes.
Prepare 12 questions grouped into: context, current process, problems, data and business rules, and expectations. Use open questions, avoid yes or no questions and mark the 4 most important ones in case I run out of time.
Format: numbered list per block.

RACI matrix for a project

Act as a project analyst. Create a RACI matrix for [the rollout of a new customer portal].
Activities: [gather requirements, approve scope, design, build, test, train users, go live].
Roles: [sponsor, product owner, business analyst, technical team, support].
Rules: only one accountable person (A) per activity and no row without an R. Format: table; then flag the roles with too much load.

Change impact assessment

Act as a business analyst. Assess the impact of the following change: [moving order approval from manual to automatic below $5,000].
Analyze the impact on processes, people, systems, data, regulatory compliance and training. For each area give the level (high, medium, low), what changes and what mitigation you propose.
End with the three main risks and the questions I should resolve before approving the change.
Format: impact table and risk list.

Validate business rules

Act as a business rules analyst. These are the rules for [granting discounts]:
[paste the rules]
Check whether any rules overlap, contradict each other or leave cases uncovered. Rewrite each rule in the format "If [condition], then [outcome]" in business language.
Output format: table with the original rule, the rewritten rule and the problem found (or "no issues").

Executive summary of an analysis

Act as a business analyst presenting results to [the finance leadership].
Material: [paste the analysis, conclusions or key data].
Write an executive summary of at most 200 words. First sentence: the conclusion and what you ask of the reader. Then three findings with their figure, one recommendation and the main risks.
Do not use technical jargon or add data that is not in the material.

Acceptance test cases from a user story

Act as a quality analyst. User story: [paste the story and its acceptance criteria].
Generate acceptance test cases: happy path, edge cases and error cases. For each case give input data, steps and expected result.
Format: table with ID, case, type, input data, steps and expected result. Maximum 10 cases.

What AI prompts for business analysts are and who they are useful for

An AI prompt for a business analyst is a written instruction that directs a language model to solve a concrete business analysis task, such as documenting requirements or analyzing a process. Prompt engineering is the practice of designing those instructions to get reliable, repeatable results, and the models you will use belong to the family of generative artificial intelligence.

• Junior business analysts, who gain consistency and a starting template for formal documents.
• Senior business analysts, who delegate the first draft and spend their time on value analysis.
• Requirements analysts, who structure business requirements consistently.
• Product owners, who write user stories and acceptance criteria faster.

What material these prompts work with

AI works from what you hand it, not from the context you know. The richer and more organized the input, the better the output:

• Meeting notes, to turn them into an actionable summary or business requirements.
• Technical documentation, to extract rules or detect gaps.
• Business rules, to validate coherence or translate them into business language.
• Qualitative data, such as survey responses or stakeholder feedback.
• Existing specifications, to rewrite them in a standard format.

If you give it a full transcript and ask “extract the business requirements into a table with priority and responsible stakeholder,” the result is useful. If you ask “help me with requirements” with no data, it returns generalities.

Benefits in day-to-day work

• Speed in documentation: the first draft of a requirements document or meeting minutes stops being the bottleneck.
• Consistency: the same prompt produces user stories with the same structure, which reduces revisions.
• Support in business process analysis: AI proposes steps, actors and decision points that you then validate.
• Less repetitive work: more time for value analysis and less for formatting.

One example: a product team at a software company used a fixed prompt to turn discovery notes into user stories with acceptance criteria, and went from drafting each one from scratch to reviewing a structured draft. Teams that standardize their documentation prompts report cutting first-draft time by around 40-50%, according to Microsoft Copilot implementation cases in product environments. The real saving depends on complexity and how standardized the prompt is. For more applications, look at AI use cases for business analysts.

AI limits and where the business analyst’s judgment still matters

An analyst who hands their judgment to AI risks costly errors: wrong requirements that reach the backlog without validation, documented processes that do not reflect how the business works, or decisions based on data the model produced with confidence but that are false.

• AI can generate inaccurate information. Sometimes it states false things with confidence: verify every relevant fact.
• It does not validate with stakeholders. Validating requirements remains a human conversation.
• It does not know the real context. It does not understand internal politics, real priorities or unwritten constraints.

Prompt engineering for business analysts: how I built these AI prompts

All of them follow the same pattern: what an analyst would take to explain to a new colleague, written in six fields. This is the method, so you can write your own:

Five tips for writing your own prompts

1. Start with the output format. If you know you want a table or a story with criteria, the rest of the prompt falls into place.
2. Paste real material. A transcript is worth more than any adjective.
3. Ask it to flag what it does not know. “If information is missing, write to be confirmed” reduces invention.
4. Refine iteratively. “Rewrite the third story with more measurable criteria” improves more than starting over.
5. Save what works. A prompt that gives good output is a template: mark the gaps with [brackets] and share it with the team.

Example: from a weak prompt to a strong one

Weak prompt:

Help me with the requirements for this project.

The model does not know who you are, which project it is or what you want back, so it returns generalities. The same prompt with the six fields:

Act as a senior business analyst in digital banking.
Context: project for a [new customer portal], with stakeholders from [sales, risk and IT].
Input: these meeting notes: [paste the notes].
Task: extract the functional requirements and classify them by priority.
Constraints: do not invent requirements; if something is ambiguous, flag it as an open question.
Output format: table with ID, requirement, stakeholder and priority.

The difference: the second prompt returns a draft that is ready to review with the team, not generic text.

How to build your team prompt library

• Organize by category: requirements, process modeling, reporting, user stories.
• Save templates with the gaps marked: [context], [stakeholders], [format].
• Standardize across the team: if everyone starts from the same templates, the documentation becomes more consistent.
• Review and improve: when a prompt gives good output, save it; when it fails, adjust it.

An analysis team at a financial services company with five analysts built a library of 12 prompts in three categories (requirements, processes and reporting). In the first four weeks they went from an average of 90 to 45 minutes on the first version of the requirements documents per sprint.

How the prompt changes in ChatGPT, Gemini, Copilot and Claude

• ChatGPT: versatile and good at text analysis. It accepts files and handles tables well. Ask for the format in words and refine in the same conversation.
• Gemini: integrates with Google Workspace. Reference the document or spreadsheet and ask for one change per message.
• Copilot: works inside Word, Excel, Outlook and Teams with your files, emails and meetings. Name the file or meeting in the prompt.
• Claude: keeps context well in long texts. Paste the full specification and ask it to cite the section of each finding.

The capabilities of these tools change often. Check the current version and plan before standardizing a team choice.

Frequently asked questions about AI prompts for business analysts

These are the questions that come up most among analysts who start working with AI, with the answer first.

How can a business analyst use ChatGPT?

To draft the first version of business requirements, generate user stories with acceptance criteria, summarize meeting notes and structure a process analysis. The key is giving it context and a clear output format. ChatGPT produces a draft; the analyst validates it and checks it with stakeholders.

What is the best AI for business?

It depends on the use case. ChatGPT stands out for versatility, Claude for long documents and careful writing, Copilot inside Word, Excel, Outlook and Teams, and Gemini in Google Workspace. Choose the tool by task and by the ecosystem you already work in.

What should a business analyst know about prompt engineering?

That a good prompt defines role, context, task, constraints and output format, and that it does not require coding. The most important thing is iterative refinement. To avoid the usual pitfalls, see this guide to common errors when using AI as a business analyst.

Can I use Claude for business analysis?

Yes. Claude is especially useful with long documents: reviewing a full specification, finding contradictions or drafting a requirements document from several sources. Always ask it to cite the section of each finding.

Can I use Gemini to document requirements?

Yes, especially if your documentation lives in Google Docs and Sheets. Reference the document, ask for a standard format (ID, description, priority, acceptance criterion) and review the result before sharing it.

What Copilot prompt should I use for the minutes of a requirements meeting?

Name the Teams meeting and ask for decisions, new requirements, unresolved questions and actions with owner and date. You have the full prompt in the Copilot section.

Is it safe to use AI tools with confidential business data?

You should not enter confidential data into public AI tools. For sensitive information, use enterprise solutions with contractual privacy guarantees and follow your organization’s policy. What data you share is the analyst’s responsibility.

How much time can a business analyst save with AI prompts?

It depends on the task. In repetitive documentation, teams that standardize their prompts report cutting the first draft by around 40-50%. According to Microsoft’s 2024 Work Trend Index, 70% of AI users at work report productivity gains in documentation. AI speeds up the draft, not the validation.

How do you refine a prompt to get better analysis results?

Evaluate the output and ask for specific adjustments instead of starting over. If it is generic, add context or specify the format. If one section fails, ask for a rewrite: “rewrite the third story with more measurable criteria.”

Your next step with AI prompts for business analysts

Pick one of the prompts in the library, replace what is in brackets with your project and run your first test. Then change a single field and compare. You do not replace your judgment: you recover hours for the work that does need analysis and validation with stakeholders.

If you want to apply AI with a method in your real work, the Online Program in AI and Innovation offers AI training applied to business and productivity, developed in collaboration with Microsoft and focused on concrete tasks like those in this guide.