Practical guide / Updated 1 November 2026
37 AI Tricks You Can Use Today
Practical prompts for work and everyday life, with 5 new examples covering visual catalogue checks, connected notebooks, small-model triage, editable documents and proactive meeting preparation.
Prompt library / Copy, adapt, run
A better starting point than a blank box.
37 complete prompts, including 5 additions in this update. Replace the words in brackets and give the tool the files or app access it needs.
How to use this guideGuide reviewed 1 November 2026. Open each new card for a worked scenario, setup steps, access notes and official sources. “New this update” marks additions to the guide; feature availability varies.
Match product photos to the catalogue before updating it.
GPT-6.1 Terra arrived in Work and Codex on 26 October with a focus on images and computer use. Give it a visual comparison that ends in a reviewable table.
Compare [supplier lookbook or image folder] with [our product catalogue or export]. For each image, record the filename, candidate product ID, source page and the visible details supporting the match. Separate confirmed identifiers from visual guesses. Flag near-identical variants, missing images, cropped labels and contradictions. Leave uncertain matches unresolved. Create a contact sheet and a proposed update table, then open both and check their readability. Do not change the live catalogue. Ask me to review the uncertain matches before preparing an import file.
Example, setup & sources
Example to try
Check a new furniture lookbook against a shop catalogue. Two chairs may look identical while having different sizes or finishes; the proposed matches should expose that uncertainty.
Set it up
- Choose GPT-6.1 Terra in Work or Codex and attach the images plus a catalogue export.
- If the catalogue is online, enable the browser tool and open the relevant page. Start with ten products.
- Check product IDs against the original catalogue before approving an import.
Access & limits
Available in ChatGPT Work and Codex on Plus, Pro, Business, Enterprise and Edu. It is not a Chat-mode model. Browser and computer access require separate setup; workspace controls still apply. File uploads work as an alternative to live catalogue access.
- A contact sheet lets you inspect the proposed matches quickly.
- Uncertain variants stay out of the import until someone checks them.
Official sources
OpenAI: Terra for visual reasoning and computer use · 26–30 Oct 2026
Bring a notebook’s evidence into the conversation you are having.
Gemini’s October notebook connection lets a chat query an existing notebook. Name the notebook and check its sources before using it to answer a question.
@Gemini Notebook, use [notebook name] to answer [specific question]. First confirm the notebook you found and list the relevant sources with their dates. Build a short answer with a supporting source passage or citation for each substantive claim. Separate what the sources establish, where they disagree and what they do not cover. Do not search the wider web unless I ask. If you cannot access the notebook or provide a citation, say so. Leave the notebook and its sources unchanged.
Example, setup & sources
Example to try
During launch planning, ask what the approved research says about a customer objection without uploading the same reports into another chat.
Set it up
- Open Gemini on the web and connect Gemini Notebook in Settings > Connected Apps if needed.
- Give the notebook a distinctive name and confirm its source list before asking for a synthesis.
- Open important citations. If the connection is absent, ask the same question directly inside the notebook.
Access & limits
Google’s 23 October announcement includes personal and Workspace accounts, with the connection off by default in the EEA. Its help page still lists narrower account eligibility, so check your account and admin settings. The connected app is documented for web and mobile, not the Mac app. Copying a chat into a notebook is not reliably supported; add source material manually when needed.
- The notebook supplies a defined evidence set instead of an unspecified memory of earlier conversations.
Official sources
Google: notebook actions inside Gemini · 23 Oct 2026
Gemini notebook setup and current limitations · Checked 1 Nov 2026
Use a small model for the first pass through a large queue.
Haiku 5.5, released on 7 October, targets short, repetitive work. Test a narrow classification task, then send ambiguous cases for closer review.
Classify [support-ticket export] using [category definitions and approved examples]. Start with a sample of 30 tickets that I can check. For each ticket, return its ID, proposed category, the exact phrase supporting that choice and whether human review is needed. Use ‘needs review’ for ambiguous cases or missing context. Do not infer sensitive personal traits. Compare the sample with my corrected labels and report mistakes by category. Only after I approve the sample, process the remaining export and put exceptions in a separate queue. Do not reply to customers or modify the support system.
Example, setup & sources
Example to try
Sort support messages into billing, delivery, product questions and unresolved cases. Review examples that fit two categories before using the labels to route work.
Set it up
- Export an appropriate set of tickets and write clear category definitions with a few labelled examples.
- Choose Haiku 5.5 in a tool that exposes it, or ask a developer to use it for this bounded step.
- Measure the mistakes and review time on a separate sample before expanding the workflow. Check exceptions yourself or with a stronger model.
Access & limits
Anthropic lists Haiku 5.5 on the Claude Platform and major cloud providers. Your app must expose model selection or support a configured workflow; naming it in a prompt does not switch models. API usage is metered, and lower model cost does not guarantee a cheaper result after review.
- The first sample reveals whether the task is clear enough to automate.
- A separate exception queue keeps uncertainty visible as the volume grows.
Official sources
Claude Haiku 5.5: repetitive tasks and availability · 7 Oct 2026
Keep the decision memo editable as the discussion changes.
Claude Docs keeps a document beside the conversation, with direct editing, comments and collaboration. Ask for targeted revisions while preserving decisions already agreed.
Create a Claude Doc from [brief, source files and discussion]. Include the decision needed, options, supporting evidence, open questions and the next review date. Cite source files beside factual claims and mark proposed decisions as proposals. When I add feedback, revise only the affected sections and explain any change to a previously agreed decision in a comment. For each chart, show the source date and refresh date. Keep the document private while we review it, and ask before granting access to anyone.
Example, setup & sources
Example to try
Maintain a workshop launch brief while dates, capacity and pricing are being agreed. A comment on the venue should update the relevant assumption without silently changing the launch decision.
Set it up
- Ask for a Claude Doc in a conversation and supply the approved sources.
- Edit directly or select a passage, add a comment and mention @Claude to request a change.
- Review sharing permissions before inviting collaborators. Ask explicitly for fresh data when updating a chart.
Access & limits
Claude Docs is available on all plans, with the required file-creation capability enabled. Enterprise availability remains subject to organisation controls. Collaborators need Claude accounts; Team and Enterprise docs cannot currently be shared outside the organisation. Charts do not refresh themselves.
- Specific comments let you improve one part of a document while keeping the rest stable.
- Dated charts make it clear when the supporting numbers need another check.
Official sources
Claude Docs: editing, sharing and chart refresh · Checked 1 Nov 2026
Claude release notes: Docs and Design in conversations · 16 Sep 2026
Ask a proactive agent to prepare for meetings before you do.
Claude Duet’s October research preview can work between conversations. Give it a narrow preparation job with clear rules for what deserves an interruption.
For meetings in [calendar] during the next two working days, use [approved email, documents and CRM sources] to prepare a brief for each external meeting. Include its purpose, attendees, the last agreed commitments, unanswered questions and links to the evidence. Check for overlapping meetings and missing travel time using the locations and buffers I supplied. Notify me only when a conflict or missing decision needs attention; save the other briefs for review. State which accounts you could read and when you checked them. Draft changes or replies, but do not send, reschedule or contact anyone.
Example, setup & sources
Example to try
Before a prospective-client call, surface the unanswered question from the last email and flag that the call overlaps a train journey already on the calendar.
Set it up
- If you have preview access, connect only the accounts needed for meeting preparation and specify your working hours and timezone.
- Review the first brief against the original invitation and correspondence, then refine the interruption rules.
- Without preview access, run the prompt manually in a connected assistant or create a supported scheduled task and verify its next run.
Access & limits
As of 1 November, Duet is a research preview accessed through a waitlist, starting with US Pro, Max and Ultra subscribers. It is not generally available in Switzerland or Iceland. Background checks share plan usage and happen every few hours; they are not continuous monitoring. Use your calendar’s own alerts for time-critical events.
- A concrete preparation window makes the task easier to judge.
- Evidence links let you confirm a conflict before changing someone else’s plans.
Official sources
Claude Duet: preview access and background work · 21 Oct 2026
Ask the agent to prove the fix in a browser.
Give a coding agent the user journey and ask for evidence that it works. OpenAI’s September example with Cognition includes a recording and a report of what remained untested.
On [local or staging URL], reproduce [bug] using synthetic data. Fix it, then test the complete user journey on desktop and mobile. Check the normal path, an invalid input and an empty state. Return before-and-after screenshots, the checks you ran, their results and anything you could not test. Do not treat a successful build as proof that the journey works. Stop before deploying or touching real customer records.
Example, setup & sources
Example to try
A signup form looks fixed. The evidence should also show that an invalid email is rejected and a valid submission reaches the confirmation screen.
Set it up
- Open the project in an agent with browser access and start a local or staging version.
- Describe the expected behaviour and provide a test account if needed.
- In current Claude Code, /verify can drive the app and record a reusable verification recipe.
Access & limits
Needs a coding agent, access to the project and a working browser or simulator. In current Claude Code, invoke /verify explicitly; /run-skill-generator can also save a project’s launch recipe for later checks.
- A screenshot makes the changed behaviour inspectable.
- An explicit untested list shows where human review is still needed.
Official sources
Cognition’s testing workflow with Astra · 11 Sep 2026
Turn source documents into a traceable analysis.
Gemini Notebook, formerly NotebookLM, can execute code in a cloud computer on supported plans. Ask for calculations with their inputs attached.
Using only [uploaded reports and CSV], compare [metric] across [periods]. First build an input table with the source file, page or row, date, unit and currency for every value. Leave missing values blank and flag conflicting definitions. Calculate changes with code, show the formula and export the table as CSV with one labelled chart. Separate observed results from your interpretation. If code execution or export is unavailable, say so and provide the table for manual checking.
Example, setup & sources
Example to try
Compare three monthly sales reports while keeping returns, tax and currency conversions visible instead of mixing incompatible totals.
Set it up
- Create a notebook and upload the reports plus the underlying data.
- Ask it to confirm which files it can read and whether code execution is available.
- Open the cited inputs and spot-check the calculation before reusing the chart.
Access & limits
Google’s current help documents code execution, charts and downloadable files as experimental notebook chat capabilities. Confirm the tools and limits in your account before starting; if execution or export is absent, request the input table for manual calculation.
- The exported input table makes the chart reproducible.
- Conflicting definitions stay visible before they distort the comparison.
Official sources
Gemini Notebook chat: code and downloadable files · Checked 1 Nov 2026
Gemini Notebook and its cloud computer · 16 Jul 2026
Talk through your notes, then test what stuck.
Google announced live conversations with your notebooks on 15 September. Use your own material for an oral rehearsal, with text chat as a fallback.
Use only the sources in this notebook to help me understand [topic]. Ask one question at a time and wait for my answer. Start with an explanation in my own words, then give me an applied problem. Correct mistakes using a source citation and distinguish a missing fact from a misunderstanding. After five questions, list what I understood, what I should revisit and the exact source sections to read. Do not reveal the answers before I try.
Example, setup & sources
Example to try
Upload a course chapter and your notes. Practise explaining the subject before an exam, interview or presentation.
Set it up
- Add a small set of trusted sources to one notebook.
- Use the mobile conversation feature if available; otherwise paste the prompt into notebook chat.
- Answer aloud or in text, then check the cited passage when a correction matters.
Access & limits
Google’s current chat help documents Live Voice Chat for users aged 18+ in the Gemini Notebook mobile app, without the earlier Ultra-only restriction. Check the controls and usage limits in your account. Use text chat if voice is unavailable.
- Answering before seeing the explanation exposes gaps you might miss while rereading.
Official sources
Gemini Notebook chat and Live Voice Chat · Checked 1 Nov 2026
Google’s September notebook study update · 15 Sep 2026
Compare suppliers with an independent evidence check.
Parallel researchers help when the work splits cleanly. Give a separate reviewer the sources and criteria to check the final comparison.
Compare [three suppliers] for [use case] against [five requirements]. Use one independent researcher per supplier if your tools support it. Each must use current official pages and record price, billing period, limits, missing features, URL and date checked. Give a separate reviewer the sources, requirements and draft table. Have them challenge unsupported claims and inconsistent comparisons. Return one comparison, unresolved questions and a recommendation conditional on those gaps. Limit the work to three researchers and one review pass; contact nobody.
Example, setup & sources
Example to try
Compare booking systems for a small business, including transaction fees, calendar support, exports and the actual plan needed.
Set it up
- Write your requirements before starting the search.
- Use an agent that supports separate research contexts, or research each supplier in a separate chat.
- For the manual version, give a fresh chat the source material and table to review.
Access & limits
Multiple independent agents require tool support and consume more usage. A prompt alone does not create separate agents in an ordinary chat.
- Each supplier gets the same questions.
- A bounded review pass can catch omissions without an open-ended debate.
Official sources
Anthropic’s multi-agent research design · 13 Jun 2025
Turn a good result into a reusable skill.
Once a workflow works, save its inputs, steps and checks. A skill gives the next run a concrete procedure and example to follow.
Turn this completed [workflow] into a reusable skill. Include when to use it, required inputs, the steps that worked, an approved output example, checks before delivery and when to ask for help. Separate stable instructions from this client’s private data. Create SKILL.md with a clear name and description, plus any small templates it needs. Test it on [a second example], report what failed and revise it once. Show me the files and installation location before changing any shared configuration.
Example, setup & sources
Example to try
Save a monthly client-report workflow that reads a CSV, uses your approved structure and checks that the totals reconcile.
Set it up
- Start from a result you have reviewed and approved.
- For Claude Code, keep a project skill in .claude/skills/[skill-name]/SKILL.md.
- Try it on different input before treating it as a repeatable process.
Access & limits
Requires an agent with skill support and file access. Other tools may use a different installation location; a reusable instruction document is the manual fallback.
- The skill preserves checks and output standards that a short saved prompt may omit.
Official sources
Keep long tasks on track with a handover file.
Save decisions, completed checks and the next step in a readable checkpoint. It gives a new session something concrete to verify before continuing.
Work towards [deliverable] with these acceptance criteria: [criteria]. Maintain a HANDOVER.md file containing the current goal, decisions and their reasons, source and output paths, checks completed, failed approaches and the next concrete step. Update it at each completed milestone. If you are about to stop or lose context, save the checkpoint first. When resuming, inspect the actual files and verify the checkpoint before continuing. Stop after [time or usage budget] and report what remains.
Example, setup & sources
Example to try
Continue a multi-session website migration without repeating a failed import or losing the decision about URL redirects.
Set it up
- Give the agent one project folder and a clear definition of done.
- Keep the handover alongside the work so another session can read it.
- Use the product’s actual budget controls where available; a written budget is an instruction, not an enforced limit.
Access & limits
This method works with agents that can read and write files. Built-in memory and context handling vary by tool; verify the current files when resuming instead of treating a checkpoint as automatically correct.
- A checkpoint is inspectable and survives a change of conversation.
Official sources
Anthropic: progress files for long-running agents · 26 Nov 2025
Watch for changes that deserve your attention.
Save a baseline, compare the next run and alert on a defined change. Claude Code’s cloud scheduled tasks, previously called routines, can run while your laptop is closed.
Every [weekday and time, with timezone], check [official pricing or product URLs]. Save a dated baseline and compare each run with the last successful one. Notify me only about [specific price, feature or availability changes]. Include the old value, new value, source link and time checked. Ignore layout-only changes. If a source cannot be read, report that failure separately; never treat it as no change. Save the comparison privately and do not contact anyone or change accounts. Confirm the schedule has actually been created.
Example, setup & sources
Example to try
Watch three software pricing pages and hear only about changes to the plans your business uses.
Set it up
- Run the comparison manually once and approve the baseline.
- Create a scheduled task at claude.ai/code/routines or with /schedule. Give each run access to the last successful baseline through a saved file or connected store.
- Allow access to the source domains, include only required connectors, and confirm the timezone and next run. Test that a second run can retrieve the saved baseline.
Access & limits
Claude Code scheduled tasks remain a research preview on Pro, Max, Ultra, Team and Enterprise; admins can disable them. Cloud runs need accessible saved state because a new session may start for each run. Local jobs depend on your computer. Saying ‘every morning’ in ordinary chat does not create a schedule.
- A saved baseline distinguishes a change from a repeated summary.
- Failure reporting keeps a broken monitor from looking reassuring.
Official sources
Claude Code scheduled tasks and triggers · Checked 1 Nov 2026
Extract document facts with an evidence column.
Ask for a page reference, a short supporting extract and an explicit missing value. This makes an extraction easier to check; it cannot guarantee accuracy.
Extract [fields] from [documents] into a table with one row per document. For each value include the source filename, page or section, a short supporting extract and one status: directly stated, calculated, inferred or missing. Leave absent values blank. Show the formula for calculated values and explain any inference. Flag contradictions and unreadable scans. Then check every critical value and a sample of the remaining rows against the original. Do not fill gaps from general knowledge.
Example, setup & sources
Example to try
Compare delivery dates, renewal notice periods and support hours across supplier proposals without silently inventing missing terms.
Set it up
- Upload readable documents and list the exact fields you need.
- Ask for a small sample first to check the table structure.
- Open the original pages for critical values; scanned pages and tables may need manual checking.
Access & limits
Works as a manual document-review prompt. Developers using Claude’s API can also enable structured citations for supported documents; that is a separate API feature.
- The evidence sits beside the extracted value.
- You can review missing and inferred items without assuming the other rows are correct.
Official sources
Turn your brand files into a working prototype.
Claude Design now works inside conversations and can use your design system. Start with a small flow you can test and prepare an implementation handoff after review.
Using [brand guide, existing page and approved copy], prototype [one customer journey]. Reuse our colours, type and components. Include the normal path, loading state, empty state and error state. Use clearly labelled sample data. Make the prototype work on a phone and desktop, then list what is simulated and what requires a real integration. After I choose a direction, prepare an implementation handoff with assets, interactions and acceptance criteria. Keep it private until I approve sharing.
Example, setup & sources
Example to try
Prototype a workshop booking flow with date selection, a full-session state and a confirmation screen before connecting payments.
Set it up
- Ask for a design in a Claude conversation, or start with a Design template in Artifacts, and provide your brand assets.
- Ask for one complete flow, then refine the specific elements that need work.
- Prepare the chosen design and its assets for implementation, and test the implemented version separately.
Access & limits
Design templates are now available on Free, Pro, Max, Ultra, Team and Enterprise. The file-creation capability must be enabled; organisation controls still apply. A prototype does not establish that payments or other integrations work.
- Real states and interactions make design feedback more specific than a static screenshot.
Official sources
Claude artifacts: current Design access and setup · Checked 1 Nov 2026
Claude Design: brand systems, exports and handoff · 17 Apr 2026
Turn meeting transcripts into a decision ledger.
Keep decisions, suggestions and unresolved questions separate. ChatGPT Record can supply the transcript, or you can use notes you already have.
Compare [today’s transcript] with [previous meeting notes]. Produce three tables: decisions made, commitments with owners and dates, and unresolved questions. Link each item to a timestamp or source passage. Distinguish an explicit agreement from a suggestion. Flag changed decisions, conflicting deadlines and unclear speakers. Write ‘not agreed’ where an owner or date is missing. Draft a short follow-up for review, but do not send it or create tasks in another system.
Example, setup & sources
Example to try
Check whether a launch date was agreed, merely proposed or changed since last week’s planning meeting.
Set it up
- Use an existing transcript, or record with participants’ consent and your organisation’s permission.
- Attach the previous notes so the comparison has an explicit baseline.
- Verify names, dates and commitments against the recording before sending the follow-up.
Access & limits
ChatGPT Record is documented for the macOS app on Plus, Pro, Business, Enterprise and Edu, subject to workspace settings. The prompt also works with a pasted transcript in a capable chat tool.
- A separate ‘not agreed’ state prevents tentative ideas becoming assigned work.
Official sources
Give an agent a folder and ask for a finished workbook.
Agents can create reports, spreadsheets and presentations from source files. Specify the actual deliverable and the checks it must pass.
Use the files in [folder] to produce an editable workbook comparing [supplier quotes or project budgets]. First list the files you can read and ask about missing essentials. Create tabs for original inputs, normalised comparisons, assumptions and exceptions. Keep formulas editable, currencies explicit and source filenames beside inputs. Reconcile totals and open or render the result to check readability. Save a new file alongside a one-page recommendation. Preserve the originals and report any check you could not complete.
Example, setup & sources
Example to try
Turn three supplier PDFs and a budget spreadsheet into a comparison that exposes different currencies, exclusions and optional costs.
Set it up
- Put copies of the relevant files in one working folder.
- Use Codex or another agent with spreadsheet creation and file access enabled.
- Open the delivered workbook and check formulas and a few source amounts before using it.
Access & limits
Requires file and spreadsheet tools. A text-only chat can draft the structure but may not create or validate an editable workbook.
- The result remains editable after the conversation ends.
- Reconciliation is part of the task rather than an assumed property of the output.
Official sources
Codex for knowledge work · 2 Jun 2026
Re-test yesterday’s failures on today’s models.
October brought Claude Haiku 5.5 and GPT-6.1 Terra. Revisit real tasks to see where a smaller model is sufficient or better image handling makes a difference.
Take these [five previous tasks], their original inputs and the acceptance criteria we set before testing. Run each task with [models available to me], keeping tools, permissions and inputs as comparable as possible. Record whether it passed, the remaining errors, elapsed time, usage cost if available and my correction time. Use a separate review against the source material. Recommend the least expensive model that meets the standard for each task type. Mark unavailable measurements instead of estimating them.
Example, setup & sources
Example to try
Retry a difficult PDF extraction, spreadsheet repair and website fix that previously needed several rounds of correction.
Set it up
- Save the original input, failed output and your correction for each case.
- Define success before rerunning, and use fresh sessions to keep earlier answers out of the comparison.
- Compare the whole tool setup as well as the model, then review results yourself.
Access & limits
Compare models your tools actually expose. Terra is available in ChatGPT Work and Codex on eligible paid plans; Haiku 5.5 is available through the Claude Platform and supported providers. Keep tool access comparable, and record review time as well as any metered cost.
- Your own failure cases reveal useful improvements that a general leaderboard may miss.
Official sources
Claude Haiku 5.5 release and availability · 7 Oct 2026
GPT-6.1 Terra in Work and Codex · 26–30 Oct 2026
Turn everything you missed into one calm catch-up.
Review my Slack, Gmail, and Google Calendar activity since [time or date]. Give me three short sections: Act now, Needs a decision, and Good to know. For every item, explain what changed and link to the original source. Ignore repeated or unchanged threads. Do not reply, schedule, or update anything without my approval.
What makes this useful
- It defines a precise time window instead of asking for a vague summary.
- The three output buckets turn information into an immediate decision queue.
- Source links make every claim easy to verify before you act.
Try next
Turn that catch-up into a plan for the rest of today. Put no more than three items in Must do, then list what can safely wait and why.
Reset your inbox without losing control of it.
Review my Gmail inbox and sent mail from the past [number] days. Identify messages that genuinely need my attention, group the rest into sensible cleanup categories, and draft replies in my voice using relevant context from Slack, Google Calendar, and Google Drive. Show me the proposed replies and cleanup plan first. Ask before sending, archiving, labelling, or moving anything to Trash.
What makes this useful
- Sent mail gives the model evidence of your actual tone and commitments.
- Triage and drafting happen together, so important context is not separated.
- The approval boundary keeps irreversible inbox actions in your hands.
Give one project a thread that remembers what matters.
Treat this conversation as the working thread for [project or launch]. Review the relevant Slack discussions, Google Drive documents, Gmail messages, and Google Calendar events. Report meaningful changes, blockers, commitments at risk, missing owners, and decisions that are overdue. Link each point to its source and separate confirmed facts from assumptions. Check again every weekday at [time], but only notify me when something materially changes. Never message people or edit project records without approval.
What makes this useful
- A dedicated thread preserves project context instead of rebuilding it each time.
- Material-change filtering prevents routine updates from becoming more noise.
- Facts and assumptions are separated before they reach a decision-maker.
Write the weekly update from the work you actually did.
Use Slack, Gmail, Google Calendar, and Google Drive to reconstruct my past seven days. Draft a concise weekly update covering what shipped, progress made, decisions taken, blockers, and next steps. Match the tone and structure of my previous updates, include links to the strongest source for each claim, and flag anything you could not verify. Do not post or send the update without my approval.
What makes this useful
- The draft is grounded in activity rather than an unreliable end-of-week memory.
- Previous updates act as a real style sample, not a generic tone instruction.
- Unverified claims are surfaced before they become accidental status theatre.
Try next
Rewrite this as a five-line update for my manager: one outcome, two progress points, one blocker, and one next step. Keep every source link.
Turn repeated feedback into a playbook you can reuse.
Review the feedback I have received in Slack and comments on documents in Google Drive. Find recurring preferences, standards, corrections, and examples of what was praised. Turn the stable patterns into a short reusable playbook for future writing, presentations, and project updates. Include before-and-after examples where the evidence supports them. Preserve my voice, cite the source material, and list any contradictory guidance that needs clarification.
What makes this useful
- Repeated feedback becomes an asset instead of disappearing into old threads.
- Examples make abstract preferences usable on the next piece of work.
- Contradictions stay visible rather than being averaged into bad advice.
Resolve the clear document comments and isolate the hard ones.
Open [Google Doc link or title] in Google Drive and review every unresolved comment. Apply edits only when the requested change is clear and consistent with the document’s voice and structure. Leave conflicting, ambiguous, or high-impact comments open. Then give me a change log with three groups: edited, needs a decision, and could not be completed. If live editing is unavailable, show the proposed edits instead.
What makes this useful
- Clear edits move forward while judgment calls remain visible to a human.
- A change log makes the work auditable without rereading the full document.
- The fallback still produces value when direct editing is unavailable.
Turn your best deck or workbook into the next team template.
Use Template Creator to convert [approved PowerPoint deck or Excel workbook] into a reusable team template. Preserve the structure, layouts, formulas, validation rules, and visual language that should remain consistent. Replace client- or project-specific material with clear placeholders and guidance. Then create a fresh version using [brief, report, or source data]. Summarise what was preserved, what changed, and what still needs review.
What makes this useful
- It distinguishes the reusable system from one project’s accidental details.
- Placeholders carry the original logic into the next piece of work.
- The test run proves the template works before the team depends on it.
Make a messy spreadsheet trustworthy before analysing it.
Review [spreadsheet, CSV, or Drive file] without overwriting the original. Create a clean copy that standardises dates and categories, identifies duplicates, and handles missing or invalid values transparently. Produce a short data-quality log, then build a small set of clear charts for the most decision-relevant trends or surprises. Finish with the main takeaway, the limits of the data, and the items I should verify manually.
What makes this useful
- Cleaning is documented separately from analysis, protecting the original evidence.
- Charts are tied to decisions rather than generated for decoration.
- Known limits and manual checks stop a polished output from overstating certainty.
Prepare a time-off request across the tools your team uses.
Use Computer with my signed-in browser to prepare a time-off request in [HR system] for [dates]. Check Google Calendar for conflicts, draft an out-of-office calendar hold, and prepare a short Slack note for my team with any coverage details. Show me every proposed change and any conflict you found. Stop before submitting the request, creating the event, or sending the message.
What makes this useful
- One instruction coordinates the request, calendar, and team communication.
- Conflict checking happens before the absence is announced.
- Every external action has an explicit stop point for review.
Turn a receipt into an expense report ready for review.
Find the receipt for [purchase] in Gmail, locate the relevant expense policy in Google Drive, and use Computer with my signed-in browser to prepare the claim in [expense system]. Cross-check the date, amount, currency, category, tax, and required evidence. Show me the receipt, every completed field, anything missing, and any policy risk. Do not submit the expense without my approval.
What makes this useful
- The receipt and policy are checked before fields are entered.
- Missing evidence is surfaced while it is still easy to fix.
- Preparation is automated while submission stays a human decision.
Watch for a genuinely good cinema seat—not just any seat.
Use Browser to check the live seat maps for [film] at [cinema] on [date range]. Look only for [preferred section or rows], exclude accessible and companion seats unless I explicitly request them, and check Google Calendar before recommending a showtime. If no suitable seat is available, check again every [interval] and notify me only when a new match appears. Include the showtime, seat position, total price, and booking link. Never purchase or change a booking without asking.
What makes this useful
- A subjective idea of a good seat becomes a testable search rule.
- Calendar context removes options that are available but unusable.
- The monitor stays quiet until the result actually changes.
Keep watch for the sold-out thing in the right size and price.
Use Browser to find [item] in [size, colour, or specification] for no more than [maximum total price]. Check the retailer and credible resale listings, confirming actual availability, condition, returns, shipping, and the final delivered price. If nothing qualifies, check once a day and alert me only when a real match appears. Include the direct listing link and the trade-offs. Do not buy anything or contact a seller without permission.
What makes this useful
- The search uses total cost and real availability rather than headline price.
- Condition and return terms make resale options comparable to retail.
- A scheduled check replaces repetitive manual browsing.
Find homes that match your life, not merely your filters.
Use Browser to find current homes matching [location, budget, move-in date, and non-negotiables]. Open the original listings and compare total monthly cost, availability, layout, commute to [place], and likely deal-breakers. Give me a ranked shortlist with direct links and explain the trade-off behind each rank. Check again every morning and notify me only when a stronger match appears. Do not contact agents, book viewings, or submit an application without approval.
What makes this useful
- The shortlist includes lived experience factors, not just listing filters.
- Rankings expose the trade-offs instead of pretending one option is perfect.
- The daily monitor reports improvements rather than repeating the same inventory.
Find a better flight and keep watching the real total cost.
Use Gmail to find the details of my trip, Google Calendar to confirm when I need to arrive, and Browser to compare current flights for [route or dates]. Show the total price including baggage, airports, door-to-door travel time, layovers, change terms, and the most important inconvenience. Check each morning for a meaningfully better price or schedule and tell me what improved. Do not book, cancel, or change any reservation without my approval.
What makes this useful
- Arrival needs and existing bookings are recovered before comparison begins.
- Total journey cost prevents a cheap fare from hiding expensive friction.
- Future alerts explain the improvement instead of reporting every fluctuation.
Make group accommodation options easy to compare honestly.
Use Browser to find live accommodation options for [destination, dates, group size, and budget]. Open each original listing and verify the total price, availability, real sleeping arrangements, bathrooms, accessibility needs, location, travel time, cancellation policy, and house rules. Create a concise comparison with listing links and photos, then explain which options genuinely fit the group and where each compromise sits. Do not book or contact a host without approval.
What makes this useful
- Sleeping arrangements and accessibility are verified instead of inferred from capacity.
- A shared comparison replaces an unstructured pile of links.
- Trade-offs are explicit before a group starts voting.
Find the photographs worth keeping without touching the originals.
Use Google Drive to review [photo folder], including available dates, locations, and camera metadata. Select the strongest images, group them by place or moment, and separately flag blur, screenshots, duplicates, and near-identical frames. Create a contact sheet or illustrated shortlist that links back to each original. Explain the selection criteria. Do not delete, move, rename, edit, or share any file without asking first.
What makes this useful
- Selection and cleanup candidates are kept as two separate judgments.
- Metadata helps recover the story of a large, messy folder.
- Original files remain untouched while the shortlist stays easy to navigate.
Plan the week around clothes you already own.
Use my wardrobe photos in Google Drive, relevant purchase details in Gmail, my Google Calendar for the next seven days, and Browser for the weather. Suggest an outfit for each day using only items you can identify in my wardrobe. Explain why it fits the event and conditions, offer one backup where useful, and ask about anything uncertain. Do not buy anything or share my photos without permission.
What makes this useful
- Recommendations are grounded in your real wardrobe and actual week.
- Weather and event context turn styling into a practical plan.
- Unidentified items trigger a question instead of an invented assumption.
Check whether your finances need attention without oversharing them.
Use Finances to prepare a private, high-level check-in. Flag upcoming bills and renewals, recurring subscriptions, unusual category changes, and possible cash-flow pressure. Compare recent spending patterns with my normal baseline and suggest a few practical questions or savings opportunities. Do not display account numbers, exact balances, merchant names, or individual transactions. Never move money, cancel a service, or take any external action without asking me first.
What makes this useful
- The output focuses on decisions while minimising sensitive detail.
- Changes are compared with your own baseline rather than a generic budget.
- The prompt draws a hard line between insight and financial action.
Build a coach that reads the training log before prescribing.
Act as my long-term training coach and journal manager. First interview me in small groups of questions about goals, history, schedule, equipment, preferences, sports, injuries, and current symptoms. Then create a weekly structure and an Excel workbook as the permanent source of truth. Before every recommendation, read and update that workbook with my completed work, readiness, soreness, symptoms, location, and lessons. Prescribe only the next workout and explain its role in the larger plan. Never claim something is logged unless the workbook was updated, and stop for medical review when symptoms make training unsafe.
What makes this useful
- The workbook prevents coaching from depending on fragile conversational memory.
- One-workout-at-a-time guidance can adapt to readiness and real life.
- Logging rules and safety limits reduce confident but context-free advice.
Wake up to a small newspaper made for your actual interests.
Every morning at [time], use Browser to create a concise personal briefing about [cities, teams, artists, hobbies, markets, local events, and other interests]. Include only developments that are genuinely new or useful, explain why each one matters to me, and link to the strongest original source. Avoid repeated stories, stale updates, and low-signal commentary. If Google Calendar is connected, add anything that could affect today’s plans. Format it for quick reading on a phone.
What makes this useful
- A personal relevance test keeps the briefing smaller than a normal news feed.
- Original links make it easy to go deeper without bloating the summary.
- Calendar context connects information to what you are actually doing today.
Choose one job to try first
If you have a folder of quotes or reports, try the finished-workbook prompt. If you build websites, ask the agent to show evidence that the user journey works. If you are learning something, give the notebook a few trusted sources and let it test your understanding.
The November additions cover visual catalogue checks, notebook questions inside Gemini, sorting a queue with a smaller model, editable decision documents and meeting preparation with a proactive agent. The last example includes a US waitlist preview and a manual alternative.
“New this update” means new to this guide, not necessarily a feature launched this month. Product claims are linked to official sources inside each new card. The prompts and examples are suggested workflows, not claims that every combination has been tested on every plan.
What changed in November
Five examples have been added and all 32 earlier prompts retained. The access notes for Claude Design and Gemini Notebook voice chat have been refreshed. The monitoring example now uses Claude Code’s current “scheduled tasks” name, and the model comparison includes October releases. Where Google’s notebook announcement and help page disagree, the card explains the limitation and gives a fallback.
Give the tool the context it needs
Replace the bracketed text, attach the relevant files and connect only the apps needed for the job. Tool names on a card describe what the workflow uses; they do not mean every assistant has those integrations. If a connector is unavailable, supply an export or ask for a manual alternative.
For recurring work, create a real schedule and check its timezone and next run. A promise in a chat is not a running automation. Cloud routines can work while your laptop is closed; local jobs depend on the computer and app being available.
Check the evidence before acting
A useful result includes its sources, assumptions, checks and unresolved questions. For documents, keep an evidence column and verify critical fields against the original. For spreadsheets, reconcile the totals. For software, exercise the user journey and inspect the result.
The July edition overstated what a “3× penalty” instruction or permission to leave blanks could achieve. Neither guarantees accuracy. A model can also label a value “extracted” incorrectly, so review important values and a sample of the rest. An independent reviewer can catch more mistakes, but it can still share the same blind spots.
Keep the approval boundary specific: draft the reply, prepare the booking or create a new file, then review before sending, purchasing or replacing something important. Start with one task you understand well enough to judge.
The original everyday prompt collection was inspired by OpenAI’s prompt board. Later additions include their own source links and access notes.