The most common reason people feel let down by AI isn't that the tools are bad — it's that they're waiting for the perfect use case to reveal itself. They heard AI was going to transform work, and they're waiting to be transformed. Meanwhile, the people extracting real value from it are just using it constantly, for small things, as a reflex.
Here are ten specific, concrete things you can do with ChatGPT, Claude, or any capable AI assistant starting today. No vague advice about "improving your workflow." Real prompts. Real use cases.
The List
-
Write the first draft of everything.
Blank-page paralysis kills more productivity than almost anything else. Give the AI your bullet points — what you need to say, who you're saying it to, the rough tone — and ask it to write a first draft. You'll edit it, but you won't be starting from nothing. Works for emails, project proposals, performance reviews, status updates, meeting follow-ups, documentation. Nearly everything that involves words.
Example promptWrite a professional but warm email declining a vendor's proposal. Key points: we liked their approach, our budget is already committed for the year, we'd be open to revisiting in Q1, and we want to thank them for their time without sounding dismissive. -
Summarize documents you don't have time to read.
Long reports, dense research papers, rambling email threads — paste the text and ask for a summary. Better yet, ask specific questions: "What's the recommended action?" "What are the risks mentioned?" "What would I need to decide by Friday?" AI doesn't get impatient with wall-of-text PDFs.
Example prompt[Paste document text] Please give me: (1) a three-sentence summary, (2) the three most important decisions or recommendations, and (3) anything that requires action within the next two weeks. -
Prep for meetings in five minutes.
Before an important call — especially with someone you haven't met, or on a topic you're rusty on — give the AI context and ask it to brief you. Who are these people? What's the background on this topic? What questions should you be asking? What's likely to be the other party's actual goal in this meeting? Five minutes of AI-assisted prep beats thirty minutes of panicked Googling.
Example promptI have a 30-minute call tomorrow with the VP of Operations at a mid-size logistics company. They reached out about our data integration product. What questions should I ask, and what concerns might they have that I should be ready to address? -
Write and debug code, even if you're not a developer.
Need a Python script to process a CSV? A spreadsheet formula that keeps breaking? A SQL query to pull specific data? AI is genuinely excellent at this, and you don't need to know how to code. Describe what you want in plain English. When it doesn't work, paste the error message and ask why. This loop — describe, get code, paste error, iterate — is how developers actually use AI tools too.
Example promptWrite a Python script that reads a CSV file called "sales.csv", removes any rows where the "revenue" column is blank, and outputs a new file called "sales_clean.csv". I have Python installed but I'm not a developer, so please include instructions for how to run it. -
Draft responses to difficult messages.
Angry customer email. Awkward feedback you need to deliver to a report. A message from a colleague you're frustrated with. Paste it and ask for help responding professionally while still being honest about the issue. AI is remarkably good at finding the diplomatic middle ground — the thing you'd write if you had an hour to think instead of five minutes and a racing heart rate.
Example promptA client sent me this: [paste message]. They're frustrated, but part of the frustration is based on a misunderstanding. Help me write a response that acknowledges their concern, corrects the factual issue without sounding defensive, and moves toward a solution. -
Learn new tools and concepts on demand.
New software at work? An unfamiliar term that keeps coming up in meetings? A concept you nodded through but didn't quite follow? Just ask. Unlike a search engine, AI can calibrate explanations to your exact level — "explain this like I'm a marketer who's never touched code" or "explain this assuming I already know the basics of machine learning." And you can follow up immediately, as many times as you want.
Example promptExplain "amortization schedule" like I'm a reasonably smart person who understands basic math but has never worked in finance. Give me a concrete example with real numbers, and tell me when I'd actually encounter one in practice. -
Create templates and frameworks you'll reuse.
Build a reusable framework once with AI's help, and you stop reinventing the wheel forever. A template for weekly status updates. A framework for evaluating vendor proposals. A checklist for onboarding new contractors. An outline for post-mortems. These take twenty minutes to build and save hours every time you use them. Once you have a good template, it compounds.
Example promptCreate a template for weekly status updates that covers what was completed, what's blocked, what's coming next, and any decisions needed from leadership. It should take someone no more than ten minutes to fill out, and it should be scannable for someone who only has two minutes to read it. -
Proofread and tighten your writing.
Not just spell-check — send AI something you've written and ask it to flag awkward sentences, cut anything redundant, make the opening stronger, or tone-check it ("does this come across as defensive?"). It's faster than asking a colleague and doesn't require social capital. Particularly useful before anything goes to an executive or a client you want to impress.
Example prompt[Paste your text] Please: (1) flag any sentences that are unclear or awkward, (2) identify anything that could be cut without losing meaning, (3) tell me if the tone reads as defensive or collaborative, and (4) suggest a stronger opening sentence. Don't rewrite the whole thing — just give me specific, targeted edits. -
Brainstorm on demand.
Five ways to approach a problem. Ten ideas for a team offsite. A list of objections a skeptical stakeholder might raise. Three alternative ways to frame a proposal. AI brainstorms without ego and without the social friction of ideation meetings. It won't tell you your idea is bad (which is a limitation worth knowing), but it will generate options you hadn't considered, fast. Use it to prime your own thinking, not replace it.
Example promptWe need to improve our weekly team meetings, which currently run long and feel unfocused. Give me five structurally different formats we could try — not variations of the same idea — with the main tradeoffs of each. -
Translate between technical and non-technical audiences.
If you regularly bridge between engineers and business stakeholders, or between domain experts and general audiences, AI handles this translation workload well. Paste a technical explanation and ask for a business-friendly version. Paste an executive summary and ask what technical details might have been glossed over. For teams that run cross-functional, this is one of the highest-leverage uses there is.
Example prompt[Paste technical content] Rewrite this for a non-technical executive audience. Focus on business impact, not implementation details. If there are decisions they'd need to make, call those out explicitly. Flag anything I've oversimplified that could cause misunderstanding later.
Making It a Habit
The people who get the most out of AI aren't the ones who've found the perfect use case — they're the ones who've built the reflex of asking first. Before spending time on something routine, they ask: could AI take a first pass at this? More often than not, the answer is yes.
The skill worth developing is prompt specificity. Vague prompts get vague results. The difference between "write me an email" and the detailed example in tip #1 isn't the AI's capability — it's the quality of the instructions. The more context you give, the more useful the output. After a few weeks of regular use, writing effective prompts starts to feel natural. For a more structured approach to that, Prompt Engineering 101 covers the techniques that make the biggest consistent difference.