Officially produced by Google! A complete guide to large model prompt words
Mastering Google Gemini for Google Workspace prompts will double your work efficiency.
Core content:
1. Introduction to Gemini for Google Workspace and its 12 workplace application scenarios
2. Tips for designing prompts for administrative support, communications and public relations, customer service, and executive decision-making
3. Comparison of correct usage and incorrect examples to avoid common misunderstandings
Opening: How do AI assistants revolutionize workplace efficiency?
The Gemini for Google Workspace series of tools launched by Google deeply integrates AI into daily office scenarios such as Gmail, documents, spreadsheets, and meetings. Whether you are an entrepreneur or a corporate executive, you only need to enter natural language prompts to let AI complete complex tasks such as copywriting, data analysis, meeting minutes, and itinerary planning. But how to use prompts correctly? This article combines Google's official guidelines to break down the design of prompts for 12 major workplace scenarios.
Chapter 1: Administrative Support - Golden Rules for Effective Schedule Planning
Apply the scenario to plan a three-day team-building activity, which should include ice-breaking activities and strategic discussions.
Correct usage tips
1. Complete four-element combination✅ Example: "You are the executive manager. The team is holding a three-day team building event in Washington, DC for the first time. You need to plan a schedule with icebreaker activities and strategic discussions." (including role + tasks + location + structure requirements) 2. Dynamic iterative optimization✅ Example: After generating a schedule, add: "3 icebreaker activities are recommended, which must be completed within 30 minutes and suitable for 25 people to participate" (refine the requirements through subsequent prompts)
Incorrect usage examples
1. Missing context ❌ Error message: "Planning team event" (missing key information such as time/location/number of people) 2. Formatting is confusing ❌ Error message: "Tell me how to arrange a meeting in Washington, the more detailed the better" (no output format is specified, such as table/text)
Chapter 2: Communications PR - AI Collaboration for Press Release Creation
Application scenario: When writing a press release for a brand acquisition, you need to cite executive interview records and brand mission documents.
Correct usage tips
1. Document citations enhance credibility✅ Example: "Use quotes from @[VIP Quotes Acquisition] to write a press release that includes the story of the founder of the acquired brand" (using @documents to call data to support content) 2. Multiple rounds of questioning to deepen the content ✅ Example: First generate an outline → then add brand mission details → finally insert document integration (the process of PR managers iteratively improving press releases in PDF)
Incorrect usage examples
1. Ambiguous instructions lead to deviations ❌ Error prompt: "Write a press release" (no key points: acquisition event + brand story + quotation requirements) 2. Ignoring audience adaptation ❌ Error message: "Writing analyst briefings in professional terms" (PDF emphasizes that it needs to be adjusted to "easy to understand" according to the audience)
Chapter 3: Customer Service - Empathy in Complaint Emails
Application scenario : To handle customer complaints about damaged headphones, it is necessary to provide alternative solutions rather than just expedited delivery.
Correct usage tips
1. Emotion + solution dual drive ✅ Example: "Write an empathetic email, the customer's headphones are damaged and need to be replaced, including an apology paragraph + 3 solution options" (clear emotional tone and structural requirements) 2. Design of alternative solutions✅ Example: "Suggest 10 alternative solutions to expedited delivery" (solve the problem through option diversity)
Incorrect usage examples
1. Single-dimensional instructions fail ❌ Error message: "Reply to complaint email" (no indication of empathy/solution etc.) 2. Overly complex requirements ❌ Error message: "Explain the return policy in the most professional terms" (PDF emphasizes "simplicity and clarity" rather than complex terms)
Chapter 4: Executive Decision-Making - Efficient Meeting Management with Voice Prompts
Application scenario: Adjusting the board meeting agenda before the flight, the tone of the email needs to be formalized.
Correct usage tips
1. Voice prompt + structured output ✅ Example: Voice input: "I am the CEO, creating a detailed outline with innovative points for the industry conference speech" (combining voice interaction with clear structural requirements) 2. Data-driven follow-up questions✅ Example: After generating a strategic outline, ask: "For manufacturing customers, how to highlight product differentiation?" (COO optimizes content through follow-up questions in PDF)
Incorrect usage examples
1. Invalid abstract instruction ❌ Error message: "Analyze market trends" (unspecified industry/time frame/output format) 2. Ignoring security boundaries ❌ Error message: "Directly use advertising data to train the model" (PDF emphasizes the principle of "data privacy protection")
Chapter 5: Frontline Management - Accurate Generation of Warehouse Count Difference Formulas
Application scenario Generate warehouse inventory difference formula and mark quantity discrepancies.
Correct usage tips
1. Improve efficiency through document linkage✅ Example: "Create a duty roster template from @[Onboarding New Hire], including morning/evening shift handover matters" (ensure information consistency through file reference) 2. Scenario-based list generation✅ Example: "Generate warehouse inventory difference formula and mark quantity discrepancies" (instructions for warehouse staff in PDF that are accurate to the formula level)
Incorrect usage examples
1. Out of touch with reality ❌ Error message: "Create a general duty schedule" (not combined with specific job requirements) 2. Ambiguous function call ❌ Error prompt: "Help me manage inventory" (no specific functions such as query/warning are specified)
Chapter 6: Human Resources - Anonymization of Employee Satisfaction Surveys
Design employee recognition programs using scenarios that cover interests such as sports, reading, and travel.
Correct usage tips
1. Personalized learning plan✅ Example: "Create a 5-day learning plan for new employees learning cybersecurity and recommend Drive-related files" (HR achieves personalization through file recommendations in PDF) 2. Sensitive information handling ✅ Example: "Summarize the survey results into an anonymous report, replacing 'Headquarters' with 'HQ'" (PDF emphasizes data cleaning techniques)
Incorrect usage examples
1. Risk of privacy leakage ❌ Error message: "Generate report directly using employee's real name" (violates the "anonymization" principle in PDF) 2. Rigid application of templates ❌ Error message: "Use a fixed template to write an offer letter" (not personalized according to the candidate's situation)
Chapter 7: Marketing - Color Psychology for Visual Design
Image of an applied scene design trade show booth, using an orange and blue color scheme with interactive computer stations.
Correct usage tips
1. Multimodal collaborative creation✅ Example: "Create an image of a trade show booth with an orange and blue color scheme and interactive computer stations" (PDF combines visual descriptions with style requirements) 2. A/B testing optimization ✅ Example: "Generate 3 ad copy variations, focusing on different pain points" (Improve results through multi-version testing)
Incorrect usage examples
1. Keyword stacking is ineffective ❌ Error message: "stack SEO keywords in the title" (PDF emphasizes the "natural language" principle) 2. Ignoring platform features ❌ Error message: "Generate 200-word plain text ad for Instagram" (not adapted to platform features such as short videos/pictures)
Chapter 8: Project Management - Intelligent Generation of Gantt Charts and Issue Tracking Sheets
Application scenario Automatically generate a Gantt chart from meeting minutes, marking milestones and responsible persons.
Correct usage tips
1. Structured issue tracking ✅ Example: "Create a project issue tracking form with status/person in charge/solution measures" (PDF standardized field design) 2. Conversational iteration ✅ Example: Generate UAT test sheet first → Generate execution email → Finally export to Sheet (coherent workflow of project manager in PDF)
Incorrect usage examples
1. Oversimplified requirements ❌ Error message: "Help me track project issues" (undefined key fields and formats) 2. Split tool collaboration ❌ Error message: "Create separate documents in Docs/Sheets" (not taking advantage of the cross-tool linkage emphasized by PDF)
Chapter 9: Sales Conversion - Accurate Capture of Customer Pain Points
The application scenario generates three major business pain points based on the customer's annual report and customizes the solution PPT.
Correct usage tips
1. Customer portrait driven ✅ Example: "Generate 3 business pain points based on @[customer annual report] for customized solution PPT" (data-driven approach for sales managers in PDF) 2. Objection handling rehearsal ✅ Example: "List the top 5 concerns that potential customers have about SaaS subscriptions and their response strategies" (improve negotiation skills through anticipation)
Incorrect usage examples
1. Hard-sell orientation ❌ Error message: "Generate a hard-sell email template" (violating the "value orientation" principle in PDF) 2. Information island operation ❌ Error message: "Independently generated quotation is not associated with meeting minutes" (not using the document association function emphasized by PDF)
Chapter 10: Small and Medium Enterprises - Non-Discounted Growth Strategy Design
Application scenario : Develop a new customer growth strategy that does not rely on discounts, including a community operation plan.
Correct usage tips
1. Cost-Benefit Analysis✅ Example: "Compare cleaning company A/B proposals and create a comparison table with booking policies" (Decision support for spa owners in PDF) 2. Non-discount incentive design✅ Example: "Design a new customer growth strategy that does not rely on discounts, including a UGC incentive program" (PDF Innovative ideas to replace traditional promotions)
Incorrect usage examples
1. Local optimization trap ❌ Error message: "Only compare prices to select suppliers" (ignore comprehensive factors such as terms of service) 2. Risk of over-promising ❌ Error message: "Marketing plan that guarantees 100% conversion rate" (violating the principle of AI assistance rather than decision substitution)
Chapter 11: Startups - Business Plan Simulation
The application scenario simulates the pros and cons of product A/B pricing strategies released in December/January.
Correct usage tips
1. Strategic sandbox simulation ✅ Example: "Comparison of the pros and cons of the simulated product A/B pricing strategy released in December/January" (Scenario analysis of the product owner in PDF) 2. Data cleaning and preprocessing ✅ Example: "Cleaning user feedback form: anonymizing names, standardizing recommended values, removing blank lines" (key steps in PDF data preparation)
Incorrect usage examples
1. Hypothetical questions fail ❌ Error prompt: "What should we do if our competitors reduce prices by 10%?" (This needs to be combined with specific scenarios rather than pure assumptions) 2. Ignore the verification process ❌ Error message: "Directly use AI-generated market analysis to make decisions" (PDF emphasizes the need for manual review of accuracy)
Upgrade Tips: Advanced Tips
1. Step-by-step disassembly
• Correct example: Create an outline first → then refine each section • Error warning: Trying to generate the entire paper in one go (easy to miss details)
2. Role-playing enhances depth • Correct example: "Analyzing industry trends from the perspective of McKinsey consultants" • Error warning: "Directly analyze data" (lack of professional perspective guidance) 3. Reverse error correction mechanism • Correct example: When results are biased, ask "What information is missing that causes the bias?" • Error warning: Blindly accept AI output (manual review required)
Final Chapter: AI assists rather than replaces
Our ultimate goal in using big models is to enhance our capabilities, not to replace decision making. All outputs must be manually reviewed to ensure accuracy and compliance.
