How to Use AI for Work: 20 Practical Examples

Email: four ways AI saves hours every week

Email is the most universally loathed productivity drain in professional life, and it is also where AI delivers the fastest, most visible wins. These four examples cover the highest-value email use cases.

Example 1 — Draft responses to complex emails. Paste the incoming email into your AI tool and prompt: "Draft a professional reply that agrees to the proposed timeline but requests the deliverables list be shared by Friday. Keep the tone collaborative, not passive-aggressive. Two paragraphs maximum." AI models like Claude Opus 4.7 and GPT-5.4 handle tone calibration remarkably well. The key is specifying the tone explicitly — "collaborative" or "direct" or "apologetic" — rather than hoping the model infers it.

Example 2 — Summarise long email threads. When you are CC'd on a 40-message thread, paste the entire conversation and prompt: "Summarise this email thread in three bullet points: what was decided, what is still unresolved, and what action is expected from me specifically." This takes a 20-minute reading task and compresses it to 30 seconds.

Example 3 — Write difficult emails from scratch. The hardest emails are the ones you procrastinate on — pushing back on a deadline, delivering bad news, declining a request. Prompt: "Write an email declining this partnership proposal. Be respectful and leave the door open for future collaboration, but make the 'no' unambiguous. Subject line included." Then edit for your voice. The AI handles the structure and diplomacy; you add the personal touches.

Example 4 — Batch-process routine correspondence. If you regularly send similar emails — weekly updates, project status reports, meeting confirmations — create a template prompt: "Using the following data points [paste bullet points], draft this week's project status email to the steering committee. Format: one paragraph summary, three key updates as bullet points, one paragraph on next steps." Run this weekly and you will recover 30–45 minutes every cycle.

Meetings: from preparation to follow-up

Meetings consume an average of 15 hours per week for mid-level professionals. AI cannot eliminate meetings, but it can dramatically reduce the time spent preparing for them, capturing notes during them, and following up after them.

Example 5 — Prepare meeting briefings. Before any important meeting, prompt: "I have a meeting with [name/company] about [topic]. Based on this background [paste any context — prior emails, company website text, previous meeting notes], prepare a one-page briefing: key points to raise, potential objections they might have, and three questions I should ask." This is especially powerful for sales meetings, investor updates, and client reviews.

Example 6 — Generate agendas from objectives. Instead of staring at a blank agenda template, prompt: "Create a 60-minute meeting agenda for a quarterly business review with the marketing team. Objectives: review Q1 campaign performance, align on Q2 priorities, and resolve the budget reallocation question. Include time allocations and the specific question each agenda item should answer." Structured agendas with clear questions keep meetings focused.

Example 7 — Transform meeting notes into action items. After any meeting, paste your rough notes (or the AI-generated transcript from tools like Otter, Fireflies, or Granola) and prompt: "Extract all action items from these meeting notes. For each action item, specify: what needs to be done, who is responsible, and the deadline mentioned or implied. Format as a numbered list." This single prompt replaces 15 minutes of post-meeting admin.

Example 8 — Draft follow-up emails. Immediately after extracting action items, prompt: "Using these action items, draft a follow-up email to all attendees. Open with a one-sentence thank-you, list the agreed actions with owners and deadlines, and close with the date and topic of the next meeting." Sending a follow-up within 30 minutes of a meeting ending is a career superpower — and AI makes it effortless.

Reports and documents: first drafts in minutes

Report writing is where AI delivers perhaps its highest ROI for knowledge workers. The blank-page problem — staring at an empty document, unsure where to start — disappears entirely when you use AI to generate structured first drafts.

Example 9 — Generate report outlines. Before writing anything, prompt: "Create a detailed outline for a [type of report] covering [topic]. Include section headings, the key question each section answers, and bullet points for the data or evidence each section should contain. Target length: [number] pages." This outline becomes your writing roadmap, and it takes 60 seconds instead of 30 minutes.

Example 10 — Write executive summaries. The executive summary is the hardest part of any report because it requires you to compress complex analysis into a few paragraphs. Paste your full report and prompt: "Write a 200-word executive summary of this report. Lead with the most important finding. Include the top three recommendations. Write for a senior leadership audience who will not read the full report." Claude Opus 4.7 is particularly strong at this — it captures nuance without losing concision.

Example 11 — Transform raw data into narrative. If you have a spreadsheet or data dump, paste the key figures and prompt: "Transform this data into a narrative paragraph suitable for a board report. Highlight the most significant trends, explain what they mean for the business, and flag any figures that are outside normal ranges. Write in third person, past tense." This bridges the gap between analysts who produce data and executives who consume stories.

Example 12 — Proofread and improve drafts. Paste your completed draft and prompt: "Review this document for clarity, concision, and consistency. Flag any sentences that are longer than 25 words, any jargon that could be simplified, any claims that lack supporting evidence, and any sections where the argument is unclear. Suggest specific rewrites for each issue." This is more thorough than any spell-checker and faster than asking a colleague to review.

Research: deeper insights, faster

AI does not replace research — it accelerates it. The key is knowing which research tasks AI handles well (synthesis, summarisation, pattern identification) and which still require human judgment (source verification, novel analysis, strategic interpretation).

Example 13 — Competitive analysis. Prompt: "Based on publicly available information, create a competitive analysis comparing [your company/product] with [competitor 1], [competitor 2], and [competitor 3]. Cover: target market, pricing model, key differentiators, recent product launches, and apparent strategic direction. Format as a comparison table followed by a one-paragraph analysis of competitive positioning." Always verify the specific claims — AI models can hallucinate company details — but the structure and analytical framework save enormous time. See our glossary entry on hallucination for more on verification techniques.

Example 14 — Summarise long documents. Paste a research paper, policy document, or industry report and prompt: "Summarise this document in 500 words. Structure the summary as: main argument, key evidence, methodology, limitations, and practical implications for a [your industry] professional." This is one of the highest-confidence AI use cases — summarisation accuracy on current models like GPT-5.4 and Gemini 3.1 Pro exceeds 95% on well-structured source documents.

Example 15 — Identify trends from multiple sources. When you have gathered several articles, reports, or data points, paste them together and prompt: "Analyse these sources and identify the three most significant trends. For each trend, provide: a one-sentence description, the evidence supporting it from the sources provided, and the potential business implications. Flag any contradictions between sources." AI excels at cross-source synthesis — a task that is time-consuming and cognitively demanding for humans.

Example 16 — Generate interview questions. Before stakeholder interviews, customer calls, or user research sessions, prompt: "Generate 15 open-ended interview questions for a conversation with [role] about [topic]. Mix diagnostic questions (understanding current state), aspirational questions (understanding desired state), and obstacle questions (understanding barriers). Avoid leading questions." This produces better questions than most researchers write from scratch, because the AI draws on patterns from thousands of interview frameworks.

Presentations: structure and storytelling

Presentations combine two difficult tasks — structuring an argument and designing visual communication. AI handles the first task exceptionally well and provides a strong starting point for the second.

Example 17 — Create presentation outlines. Prompt: "Create a 15-slide presentation outline on [topic] for [audience]. For each slide, provide: the slide title, the single key message, three bullet points of supporting content, and a suggestion for the visual element (chart type, diagram, image concept). The presentation should follow the situation-complication-resolution narrative arc." This gives you a complete presentation skeleton in under a minute.

Example 18 — Transform reports into presentations. Paste a written report and prompt: "Convert this report into a 12-slide presentation outline. Each slide should cover one key point. Simplify language for verbal delivery — shorter sentences, no subordinate clauses, active voice throughout. Suggest a data visualisation for any slide that references numbers." The translation from written to spoken communication style is something AI handles remarkably well.

Example 19 — Write speaker notes. After building your slides, paste each slide's content and prompt: "Write speaker notes for this slide. The notes should: expand on the bullet points with conversational language, include a transition sentence to connect to the next slide, and suggest one audience engagement moment (question, pause for reaction, or call-to-action). Target 60 seconds of speaking time per slide." Speaker notes are the difference between a presentation that is read aloud and one that feels natural.

Data analysis: making numbers meaningful

You do not need to be a data scientist to use AI for data analysis. Current AI models can interpret spreadsheets, identify patterns, generate charts, and explain statistical concepts in plain language.

Example 20 — Analyse spreadsheet data. Paste your data (or upload a CSV to tools that support file upload) and prompt: "Analyse this data and provide: a summary of key statistics (mean, median, range, standard deviation for numerical columns), identification of the three most notable patterns or anomalies, and a plain-language explanation of what this data tells us about [your specific question]. Recommend two charts that would best visualise the most important findings." GPT-5.4 with Code Interpreter and Claude Opus 4.7 with its analysis tool both handle this well, including generating the charts directly.

These twenty examples cover the six activities that consume most knowledge-worker time. The common thread is a pattern you can apply to any task: be specific about what you want, provide relevant context, specify the format of the output, and always review before sending. This pattern — which Enigmatica's CONTEXT Framework formalises into a repeatable methodology — is the difference between getting mediocre AI outputs and getting outputs that genuinely save time.

Start with the example closest to your daily work. Master it until it becomes automatic. Then add the next one. Within a month, you will have AI integrated into a half-dozen recurring tasks, saving hours every week without any dramatic workflow overhaul. For a deeper dive into the principles behind these examples, explore Enigmatica's free curriculum — Level 1: Foundations covers the fundamentals, and Level 2: Essentials teaches the prompting techniques that make these examples work consistently.

Related Terms

Related Articles