Summarizing and Comparing Documents
You quickly analyze long documents and contracts with AI.
Why this lesson matters
Professionals spend an average of 2-3 hours a day reading and processing documents. Contracts, policy papers, research reports, minutes. The pile grows faster than you can read. AI can process hundreds of pages in seconds and tell you exactly what's in them.
This lesson teaches you how to use AI to summarise documents, compare them and pull structured information out of them.
What you'll learn
- Summarise long documents effectively
- Compare two or more documents
- Request structured output (tables, lists)
- Analyse contracts and legal documents
Summarising documents
A good summary prompt contains three things: the goal, the audience and the desired format.
Basic summary
Summarise this document in 300 words maximum.
Focus on:
- Main conclusions
- Action items
- Risks or points of attention
Audience: project manager who wants to know quickly
whether action is needed.Layered summary
For long documents a layered approach works:
Give three levels of summary:
1. One sentence (the core message)
2. One paragraph (the key points)
3. Detailed summary with all relevant points
(500 words maximum)Keep reading
Leave your name and email address and you can read the rest
You get the whole lesson right away, and every other lesson stays open after that. No password, no confirmation email.
This layered approach is perfect for documents you need to forward. The recipient chooses the level of detail.
Comparing documents
AI can put two versions of a document side by side and name the differences.
Compare these two versions of the contract.
Give an overview of:
1. What changed (old → new)
2. What was added
3. What was removed
4. Which changes matter for the substance
(rather than just the wording)
Present as a table with columns:
| Article | Change | Impact | Action needed? |Never upload confidential documents to public AI tools such as the free version of ChatGPT. Use enterprise versions or local AI solutions for sensitive data. More on this in lesson 6-2.
Structured output
The power of AI in document analysis is in turning unstructured text into structured data.
Example: information extraction
Extract the following information from this contract
and present it as a table:
| Field | Value |
|------|--------|
| Parties | |
| Start date | |
| Term | |
| Notice period | |
| Total value | |
| Penalty clause | Yes/No + details |
| Automatic renewal | Yes/No + period |Example: generating a checklist
Analyse this policy document and make a checklist
of all obligations for our department.
Format per item:
- [ ] What needs to happen
- Deadline: ...
- Responsible department: ...AI works best with text-based documents (PDF with text, Word, TXT). Scanned documents without OCR are tricky. Use an OCR tool first if you work with scans.
Analysing contracts
AI can help you get to grips with contracts faster. It doesn't replace a lawyer.
Use AI analysis of contracts as a starting point, never as a final verdict. Always have legally binding decisions checked by a lawyer.
Workflow: from document to action
The most powerful application combines summarising with action items:
Analyse these meeting minutes and give:
1. Summary in 5 bullets
2. All action items mentioned, with who and when
3. Open decisions that still need to be taken
4. Topics that need to come back
in the next meetingWhat is the best approach for summarising a long document for different readers?
When should you NOT rely on AI analysis of a contract?
Want to keep your progress, get the practice files and sign up for the free evening? Create a free account. All you do is click a link in your email.
Create a free account