Summarise a long document with AI
Create a shorter text for a particular reader while preserving the original argument, evidence and important qualifications.
Summarise a long document with AI when the reader needs a decision or an overview and cannot read the original in full. The summarization skill has the agent agree the purpose and length, read the complete source and select the claims that matter. It treats a limitation buried near the end as seriously as a claim in the opening paragraph.
A project review might report an improvement and then explain that only returning customers were measured. Cutting the second detail would change the finding. A useful brief keeps the population visible and drops the repeated examples or project history that do not affect the reader’s decision.
What it does
- Prepare a decision summary from an internal report.
- Keep necessary qualifications close to the findings they limit.
- Distinguish meeting decisions from suggestions that were discussed.
- Remove repeated illustrations without removing the evidence itself.
- Make references clear to a reader who has no original document open.
Install and use
- Give a conversational agent the authorised source, intended reader and length budget. Supply a transcript yourself if the source is audio.
- Ask for a brief focused on one decision and a note about any source sections it could not read.
- Check the strongest claim and its qualifiers against the original before circulating the summary.
FAQ
Source
Made by Arjun Prabhulal. Original skill: summarization, published on GitHub under the MIT licence.
View on GitHub