You’ve heard the gospel: AI is going to change everything. Good, great, great.
But when you’re faced with a deadline and 80 unread emails, you don’t need philosophy, you need a cheat sheet.
The fastest way to master AI isn’t by watching lectures, but by finding a way to replace an hour of your grind with a 10-second prompt. Here are five specific, repeatable ways to automate your most time-consuming professional tasks.
Grab your favorite chatbot (Gemini, ChatGPT, Claude, Copilot – whatever floats your boat) and let’s get started.
To write
Staring at a blank page. Annoying, formulaic first drafts. Enough.
You are a professional. You shouldn’t have to spend an hour composing a standard email to a client or writing the first three paragraphs of a report. That’s grunt work.
Instead, master “constraint-based prompts.” Here you tell the AI exactly what to write and how to write it, forcing it to follow your specific, professional rules.
Here’s a quick example:
“You are one [job title]. Design one [document type (memo, email, etc.)] Unpleasant [target audience]. The tone has to be [tone]. The three main takeaways are [list three specific bullet points]. The final memo should be complete [length in words] and add a subject line.”
Action items after the meeting
Want to search long transcripts and meeting notes for action items?
You’re doing it wrong. Let the AI do the heavy lifting of synthesis.
It’s time to leverage “outcome-based cues.” Instead of asking for a summary, you can ask the AI to produce specific, structured results from a large amount of text, such as a meeting transcript or a compact PDF.
For example:
“Analyze the following [meeting transcript/document]. Do not summarize the entire text. Instead, produce three separate results: 1) A table with all the action items, the person responsible and the stated deadline. 2) A list of three open questions that have not been resolved. 3) A short two-sentence subject line for follow-up.”
In less than a minute, you can turn raw data into an organized, actionable to-do list.
Research
Turning generative AI into a true, trusted research assistant that can search and reference information across multiple work files will require tools that let you upload your own content, such as Google’s NotebookLM or similar features on other platforms.
This is called “contextual grounding” and involves uploading a handful of annual reports, project documents, or extensive research files. First check with your organization whether there are any rules against this.
Here’s a prompt you can use:
“Only based on the uploaded documents, what is the biggest discrepancy between the revenue forecast for the fourth quarter of 2024 [from Document A] and actual marketing spend in the first quarter of 2025 [from Document C]? Please explain the gap in three bullet points, referring to the specific document in which the information was found.’
This allows you to stop relying on the AI’s general knowledge and start using it as a hyper-efficient analyst on your own private data, generating insights that would take hours to come up with on your own.
Brainstorming
Thanks to AI, hitting a creative wall or falling victim to groupthink while brainstorming is no longer the same as before.
While your brain thinks linearly, AI can think exponentially, but you have to force it to show its work.
Use ‘critical reasoning’, also known as ‘thought chain’. This forces the AI to debate, criticize and explore alternatives before arriving at an answer.
An example of a prompt formula:
“I have an idea for a new product feature: [describe the feature]. Before you suggest a name for it, do the following: 1) Act like a skeptical customer and list three reasons why this feature is useless. 2) Act as a competitor and list three ways they can easily copy and neutralize the feature. 3) Only after these two steps, suggest three different benefit-oriented names for the position.
This forces the AI to act as a constructive opponent, helping you come up with a better, more robust idea much faster.
#cheat #sheet #solves #big #everyday #problems