Gemini Prompt Engineering Guide: Get More From Google's AI Assistant

Gemini has distinct strengths in Google Workspace integration and multimodal tasks. This guide covers Gemini-specific prompting techniques to get better results in Gmail, Docs, and beyond.

*Some links in this article are affiliate links. If you sign up through one, ShabelleHub may earn a commission at no extra cost to you — see our affiliate disclosure for full details.*

Gemini's Distinct Strengths to Prompt For

Prompting Gemini effectively starts with understanding what it does differently from ChatGPT or Claude. Gemini's practical advantages are concentrated in three areas:

Google Workspace integration — Gemini in Gmail, Docs, Sheets, Slides, and Meet acts directly in those tools rather than generating text you paste elsewhere.

Multimodal input — Gemini handles images, PDFs, video, and audio alongside text, processing them together as part of the same prompt.

Real-time web access — Gemini has web search built in, making it more reliable for current events and recent information than Claude's free tier.

Prompting Gemini in Google Workspace

When using Gemini inside a Google app, explicit mentions of the workspace context improve output.

Gmail

Drafting from bullet points:

> "Draft a professional reply to the email in this thread. My key points: (1) I cannot attend Thursday's meeting, (2) I am available Monday or Wednesday afternoon, (3) I will send the project update by end of week. Keep it under 100 words and match the tone of the original."

Summarising a long thread:

> "Summarise this email thread. What was the core issue, what decisions were made, and what is expected from me?"

Google Docs

Improving a section:

> "The following section is too dense. Rewrite it to be more readable without losing any important content. Keep the same factual claims. Section: [paste section]"

Generating a first draft from notes:

> "I have written the following rough notes for a [document type]. Expand these into a coherent first draft with proper paragraph structure. Do not invent details I have not included. Notes: [paste notes]"

Google Sheets

Formula help:

> "Write a Google Sheets formula that [describe what you need]. The data is in columns [describe]. I want the output to [describe]."

Prompting Gemini for Multimodal Tasks

Gemini can process images, PDFs, and screenshots alongside text. Being specific about what you want from the visual input dramatically improves output.

Analysing a screenshot:

> "This is a screenshot of our product's analytics dashboard. Describe: (1) what the data is showing, (2) any trends that stand out, (3) anything that seems anomalous or worth investigating. Context: this is weekly active users data for the past 6 months."

Reviewing a PDF:

> "I am attaching a 20-page report. Please: (1) identify the main argument or recommendation, (2) note any data or statistics cited as evidence, (3) flag any claims that seem to require stronger evidence."

Extracting from a photo:

> "This is a photo of a whiteboard from our planning meeting. Extract all the text and organise it into: decisions made, action items with owners, and open questions."

Prompting Gemini for Research

Gemini's real-time web access makes it more reliable than Claude for current-information questions.

Pattern for research questions:

> "Search for current information on [topic]. I need: (1) a brief overview of the current state, (2) key developments in the last [timeframe], (3) the main perspectives or debates. Cite your sources clearly."

For fact-checking:

> "I have read that [claim]. Please research whether this is accurate, what the evidence says, and whether this is settled or contested among credible sources."

Using Gemini Gems (Gemini AI Pro)

Gems are customisable AI personas in Gemini Advanced that remember standing instructions across conversations.

Example professional editor Gem:

> Instructions: "You are a professional editor who prioritises clarity, precision, and directness. When reviewing text: flag passive voice, vague language, and unsupported claims. Suggest specific rewrites rather than general feedback."

Gems are worth setting up for any recurring task type where you would otherwise re-establish the same context in every conversation.

Prompts That Work Well With Gemini's Strengths

Gmail catch-up after absence:

> "Summarise my unread emails from the last [X days]. Group by: urgent and needs response, informational and no action needed, and newsletters that can be archived. For urgent items, note what response is expected and by when."

Docs research with web access:

> "I am writing a document about [topic]. Research the current state of [specific aspect] and draft a factual paragraph I can include, with source references."

Meeting prep:

> "I have a meeting about [topic] in [timeframe]. Help me prepare: suggest 5 questions to ask, note 3 things I should know going in, and draft a one-sentence framing of my position on the main issue."

A Real Use Case: Drafting Prompts for a Content Project

Gemini has been used specifically for drafting and refining the prompts behind ShabelleHub's content — working through what a prompt for a specific article should ask for, and iterating on the phrasing before using it elsewhere. This is a slightly different use case than the Workspace-integration examples above, but it points to the same underlying strength: Gemini is a capable thinking partner for getting a piece of writing or a prompt into a usable shape before it's used for the actual task.

Conclusion

Gemini's strongest use case is when your work happens inside Google's ecosystem. The Workspace integration removes the copy-paste overhead that makes general AI tools more friction-heavy for everyday communication and document work.

For a broader comparison of Gemini against its main competitors, see our Gemini vs ChatGPT comparison.

Frequently Asked Questions

📬 Get the latest AI tool reviews

Expert picks and comparisons, weekly. No spam.