Day 9: Summarise Anything in 60 Seconds
The Concept
Information overload is one of the most consistent problems in modern work and life. Email threads that run to thirty messages before a decision is made. Reports that take forty pages to say what could be said in four. Meeting notes that capture everything spoken without distinguishing what was actually decided. Articles you bookmarked three weeks ago that sit in a growing list of good intentions.
The default response to all of this is to skim — which means you either spend time on material that did not deserve it, or you miss something important in material that did. Neither outcome is good, and neither is sustainable.
Why summarising is AI's most reliable skill
Of all the things AI can do, summarisation is where it is most consistently useful and most reliably accurate. The reason is structural: AI is working with content you have provided rather than generating facts from its own training data. It is compressing and organising something that already exists. This means the usual hallucination risk is significantly reduced — the underlying content is yours, and AI is extracting and structuring it.
This is why summarisation is a good place to start trusting AI. The output is directly checkable against the source material you provided, and errors or omissions are easy to catch.
The difference between passive and active summaries
A passive summary tells you what something said. An active summary tells you what you need to do about it. The distinction matters enormously in practice. If you ask AI to "summarise this report," you get a recap of the contents. If you ask AI to summarise the report and identify the three things that require your action, any decisions needed before the next meeting, and anything that seems incomplete or uncertain, you get something you can actually use.
The prompt in today's task builds this distinction in. It asks for the main point, key information, required actions, and things to verify. That four-part structure turns a passive exercise into an active one — which is what makes the output useful rather than just interesting.
What to use it for
The most valuable applications are documents with a low signal-to-noise ratio — where important content is buried in surrounding material. Long email threads where you need the current status and what is being asked of you. Meeting transcripts where you need action items, not a verbatim record. Reports where even the executive summary is too long. Dense articles you need to understand before a conversation or meeting.
NotebookLM by Google is particularly useful for longer documents — it holds an entire PDF and lets you ask questions across sessions. ChatGPT, Claude, and Gemini all handle pasted text well for shorter material. The rule for all of them: ask for actions, not just a recap.
Prompt of the day
Copy this into your AI tool and replace any bracketed placeholders.
Prompt
Please summarise the following content. Give me: 1) The main point in one sentence. 2) The three to five most important things I need to know. 3) Any actions or decisions this requires from me. 4) Anything I should verify or that seems uncertain. Here is the content: [PASTE THE TEXT, EMAIL THREAD, OR TRANSCRIPT YOU WANT SUMMARISED].
Your 15-minute task
Pick one piece of long content from your week — a document, an email thread, a meeting transcript, or an article you have been meaning to read. Paste it into the prompt. Review the summary and note what it flagged as requiring your action.
Expected win
A useful summary of something long with the key points and your next actions clearly identified — done in under two minutes.
Power user tip
For very long documents, paste them section by section and ask: 'Add this to your running summary and update your list of action items.' This gets around context limits and produces a better output than pasting everything at once.