Skip to content
Founderz: Responsible Use of AI home

in collaboration with Microsoft

Toolkit contents

Practice · 30 examples

Prompt examples

Prompts for common tasks, shown before and after a rewrite. Copy the improved version and use it on real work.

Your part

Every example ends with the step you take outside the prompt. A prompt can’t limit what a tool may access, check the answer or approve an action.

30 examples shown.

Principle 1: Human oversight of AI

Customers

Refunds in the support queue

Before

“Answer the complaints in the support queue and refund anyone who qualifies.”

After

“For each complaint in the support queue, draft a reply and recommend one action: refund, replacement or escalation. Cite the policy clause you used. Put everything in a review list for me.”

Your part

Approve each reply and each refund yourself. Check that the agent’s permissions make every refund wait for you, because the prompt alone can’t hold one back.

Why it works: The agent does the preparation at speed. The decision that costs money stays with a person, who sees the policy behind each one.

Operations and projects

Moving a meeting

Before

“Move tomorrow’s project review to Friday and let everyone know.”

After

“Find three slots on Friday when all six attendees are free. Draft a short message explaining the move. Don’t change the calendar or send anything yet.”

Your part

Pick the slot, check the message and send it, and make sure the agent can’t change calendars without your approval. Once you’ve seen it get this right a few times, you might let it act alone for internal meetings. That’s level 3, and it’s your call.

Why it works: Starting at level 2 shows you how the agent behaves before you give it more freedom.

People and hiring

Preparing interviews

Before

“Rank these 40 résumés and tell me who to interview.”

After

“Using the attached job description, draft eight interview questions that test the four must-have skills, with what a strong answer would include.”

Your part

Screen and choose candidates yourself, through your HR process. AI helps you prepare, and the decision about people stays with people.

Why it works: AI saves time on the preparation, while the decisions that affect someone’s chances stay explainable and lawful.

Principle 2: A single source of truth for AI

Contracts and policies

Answering from the official policy

Before

“What’s our refund policy for enterprise clients?”

After

“Using only “Refund policy v3.2” in the Legal folder, answer this client’s question. Quote the clause you rely on. If the policy doesn’t cover the case, say so.”

Your part

Open the clause it quoted. If you find an older copy of the policy somewhere else, ask the owner to archive it.

Why it works: Naming the official source and asking for the clause makes the answer checkable in seconds.

Sales and pricing

Checking a discount

Before

“What discount can I offer this client?”

After

“Using only the current price list (“Pricing 2026” in the sales wiki) and the discount policy, tell me the maximum discount for a three-year contract. Quote the lines you used, and tell me if the two documents disagree.”

Your part

Open both documents. If they disagree, ask the owner which one is right, and get the other one fixed.

Why it works: Asking where sources disagree surfaces the stale file before a client sees it.

Research

Updating the team wiki

Before

“Summarize our onboarding process and save it to the wiki.”

After

“Summarize our onboarding process from the three attached documents, as a draft for review. List anything that looks out of date or contradictory.”

Your part

Review and fix the draft, then publish it with an owner and today’s date. Archive the pages it replaces.

Why it works: The AI drafts and a person publishes, so nothing reaches the wiki until someone has checked it.

Principle 3: Confidentiality of information shared with AI

Data and numbers

Finding upsell patterns

Before

The full client spreadsheet, names and contract values included, pasted into a personal chatbot:

“Find upsell opportunities.”

After

In the work tool your organization approved:

“Here’s an extract by segment: industry, company size, products bought and renewal month. Suggest upsell patterns for each segment.”

Your part

Keep names and contract values out unless the tool and the task are cleared for them.

Why it works: Upsell patterns come from segments, so names and contract values add exposure without improving the analysis.

Contracts and policies

Reviewing a supplier contract

Before

The whole contract uploaded to a free online tool:

“Find the risky clauses.”

After

In the approved tool:

“Compare the liability, termination and data protection clauses of this contract with our standard terms (attached). Show the differences in a table, quoting both texts.”

Your part

Use the tool your organization approved for confidential documents, and send anything that matters to legal. The AI’s reading isn’t legal advice.

Why it works: A focused comparison with your own terms is more useful than a general hunt for risk, and it stays in a protected tool.

Principle 4: Limited permissions for AI agents

Email and agents

Sorting your inbox

Before

“Go through my inbox and deal with anything urgent.”

After

“Go through today’s inbox. Draft replies to urgent messages and save them in Drafts. List separately any email that asks for a payment, a change of bank details, a password or a file.”

Your part

Set the agent to draft-only access if the tool allows it, so the limit doesn’t depend on the prompt. Before acting on any payment request, call back on a number you already have.

Why it works: Draft-only access keeps every outgoing message in your hands, and the separate list catches the emails attackers write.

Research

Researching suppliers on the web

Before

“Research these five suppliers online, then email each one asking for a quote.”

After

“Research these five suppliers on the web and summarize what each offers, with links. Draft a quote request for each, but don’t send them.”

Your part

Give the agent no permission to send email for this task, and read the drafts before you send them. Web pages can carry hidden instructions.

Why it works: Without permission to send email, an agent that reads the open web is missing one part of the lethal trifecta.

Operations and projects

Automating a routine order

Before

“Every Monday, check stock levels and reorder whatever is running low.”

After

“Every Monday, check stock levels and prepare a reorder list for items below the minimum. Orders over €500 or from new suppliers come to me for approval.”

Your part

Set the spending limit in the purchasing system itself, not only in the prompt, and review the first few weeks of orders.

Why it works: Routine, low-value orders can run within limits. The limit only protects you if the system enforces it.

Principle 5: Independent judgment and skills

Decisions and plans

Setting a price

Before

“What price should we set for the new product?”

After

“I think €49 a month is right because [your reasoning], based on this data [data]. Where is my reasoning weakest, and what am I missing?”

Your part

Decide, and write down anything that changed your mind.

Why it works: Starting from your own view turns AI into a sparring partner and keeps the judgment yours.

Learning

Learning a spreadsheet skill

Before

“Build the pivot table analysis for me.”

After

“Walk me through building a pivot table of sales by region and month, one step at a time. After each step, ask me what I expect to see before you show the result.”

Your part

Do the steps yourself in your own spreadsheet. Next time, try it without help first.

Why it works: Predicting each result before you see it helps the skill stick.

Writing

Drafting a strategy note

Before

“Write our Q1 marketing strategy.”

After

“Here is my outline for our Q1 marketing strategy: three priorities and why. Challenge each priority, suggest what’s missing and point out where I’m vague.”

Your part

Keep the strategy yours. Use the challenges to sharpen it, and write the final version yourself.

Why it works: You end up with your own strategy, tested before anyone else sees it.

Principle 6: Challenge and bias awareness

Decisions and plans

Testing a plan

Before

“Here’s my plan to open an office in Lisbon. Why is it a good idea?”

After

“Here’s my plan to open an office in Lisbon (attached). Act as a skeptical CFO. Give me the three strongest reasons it could fail, the assumptions I’m making without evidence and the data that would change your mind.”

Your part

Answer each objection in writing before you decide. The ones you can’t answer are your homework.

Why it works: A leading question invites agreement. A named skeptic gives the assistant permission to disagree.

Research

Questioning a business case

Before

“Explain why moving our sales team to a new CRM will save money.”

After

“We’re considering moving our sales team to a new CRM. What are the realistic costs and savings, what usually goes wrong in migrations like this, and what would we need to measure to know it worked?”

Your part

Check the cost assumptions with finance, and ask a team that has done a migration what they’d do differently.

Why it works: An open question gets you the risks as well as the benefits, and a way to measure the result.

Sales and pricing

Getting honest feedback on a proposal

Before

“Is this proposal good?”

After

“Review this proposal as the client’s procurement lead, who has two cheaper alternatives. What would make you say no? Be specific, and don’t soften it.”

Your part

Fix the weaknesses it finds, then ask a colleague who knows the client.

Why it works: A specific critic with a reason to say no finds problems that a request for general feedback misses.

Principle 7: Verification proportionate to the stakes

Research

Writing with statistics

Before

“Write a paragraph on the growth of the European AI market, with statistics.”

After

“Using only the three reports I’ve attached, write a paragraph on the European AI market. After each number, give the report and page. List any claim you couldn’t support.”

Your part

Check every number against its page. Numbers you can’t find come out.

Why it works: Limiting the sources and asking for pages turns an unverifiable paragraph into one you can check in minutes.

Contracts and policies

Summarizing a regulation

Before

“Summarize the new EU rules on AI chatbots.”

After

“Summarize Article 50 of the EU AI Act from the official text I’ve attached, in plain English. Quote the sentence behind each point, and flag anything that depends on dates or exceptions.”

Your part

Check each quoted sentence in the official text, and ask your legal team before you act on it.

Why it works: Quotes tie every point to the text, and the flags show where the detail matters.

Data and numbers

Comparing quotes

Before

“What’s the total cost of the three quotes?”

After

“Extract the price, quantity and any extra charges from each of the three attached quotes into a table, with the page for each figure. Then total them and show the calculation.”

Your part

Check every figure against its page and redo the total yourself. Look for charges hidden in the terms.

Why it works: A table with page references makes a wrong number easy to spot, and your own total catches the rest.

Principle 8: Quality of AI-assisted work

Writing

Summarizing a long report

Before

The result sent straight to the sales team, all one page of it:

“Summarize this 40-page report.”

After

“Summarize this report for our sales team in five bullet points: what changed, what it means for our Q4 deals and one action. Maximum 120 words.”

Your part

Read the report’s own summary, check the five points against it and add one line with your view.

Why it works: Naming the reader, the purpose and the length gets a summary people will use.

Meetings

Sharing meeting notes

Before

“Summarize the meeting transcript.”

After

“From this transcript, list only the decisions, the actions with owners and deadlines as stated, and the open questions. Mark anything said as a possibility, not a decision.”

Your part

Check each decision and deadline against what was said, fix who said what, then send it with a line asking for corrections.

Why it works: Separating decisions from possibilities stops a provisional date from becoming a promise.

Customers

Replying to a complaint

Before

“Write a reply to this complaint.”

After

“Write a reply of under 120 words to this complaint. Acknowledge the problem in one sentence, say what we’ll do and by when (refund processed within five working days), and give a named contact.”

Your part

Check the promise is one you can keep, and make it sound like you.

Why it works: A short reply with a specific promise and a named contact reads as care. A long generic apology reads as a template.

Principle 9: Transparency about the use of AI

Marketing and images

Writing about your experience

Before

“Write a LinkedIn post about how I turned around a struggling team, with a moving anecdote.”

After

“Here are my notes on three real moments from last year. Help me turn one into a LinkedIn post in my voice. Use only events, numbers and quotes from my notes.”

Your part

Publish only what happened. If you add an AI-generated image of people, say so.

Why it works: AI helps you tell the story well, and your notes keep it true.

Customers

Introducing a chat assistant

Before

The greeting of a customer-facing chat assistant:

“Hi, I’m Sara! How can I help you today?”

After

The same assistant, introduced honestly:

“Hi, I’m the virtual assistant for [company], an AI. I can help with orders and returns, and I’ll connect you with a person whenever you ask.”

Your part

Make sure the handover to a person works, and test it the way customers would.

Why it works: People deserve to know they’re talking to a machine, and to reach a person when they need one.

Marketing and images

Creating campaign images

Before

“Create a photo of our customer [name] happily using our product.”

After

“Create an illustration of a person using our product in a bright kitchen: stylised, and clearly not a photo of a real person.”

Your part

Use real customers only with their written consent and real photos. Label synthetic images where your channel or the law requires it.

Why it works: A realistic AI image of a real customer would pass for a real photo, which makes it a deepfake. A stylised illustration can’t be mistaken for a photo.

Principle 10: Accountability for AI-assisted results

Writing

Finishing a client report

Before

The final report, sent as it came out:

“Write the final client report.”

After

“Draft sections 2 and 3 from my notes and the attached data. Mark every sentence that goes beyond my notes with [CHECK].”

Your part

Clear every [CHECK] yourself, then read the whole report once more from start to finish.

Why it works: Marking what goes beyond your notes shows exactly where your checking is needed.

Writing

Correcting a mistake

Before

The wrong figure fixed in your own copy of the report, and nothing said to the clients.

After

“Draft a short note to the three clients who received last week’s report: the figure on page 4 was wrong (12%, not 21%), here is the corrected page, and nothing else in the report changes.”

Your part

Check the correction yourself, send it today and tell your team what happened so the check gets added.

Why it works: A quick, specific correction builds more trust than a silent fix.

Decisions and plans

Keeping a trail

Before

“Tell me which supplier to choose.”

After

“Compare these three supplier proposals against our criteria (attached) in a table, quoting the proposal for each score. Then list what I should verify before deciding.”

Your part

Verify the listed points, decide, and save the table with your decision and the reason. If anyone asks later, you can show how you chose.

Why it works: Quoting the proposal for each score ties the comparison to evidence, and the list of checks keeps the decision with you.