AI Insights & Key Term Extraction

Signeasy's AI Insights feature extracts key terms from your contracts automatically — dates, parties, values, obligations — and organises them as structured data in your contract repository. Once extracted, these terms become searchable, filterable columns you can use to manage and report on your contracts at scale.

This article explains how extraction is triggered, what to expect during the process, and what is planned for future onboarding improvements.

 

Step 1 — Tag a contract type to trigger extraction

AI extraction does not run automatically on every uploaded document. It is triggered when you assign a contract type to a contract. This ensures the AI knows which category of terms to look for — the fields relevant to an NDA differ from those in a vendor agreement or employment contract.

 

How to tag a contract type

1.  Open the contract you want to tag from your repository.

Screenshot 2026-07-15 at 7.52.28 PM.png

2.  Click the Contract Type field in the contract detail panel.

3.  Select the appropriate type from the dropdown (e.g. NDA, Vendor Agreement, SaaS, Employment).

4.  AI extraction begins immediately once a type is selected.

 

Why contract type is required

Different contract types have different key term schemas. Tagging the type tells the AI which fields

to extract — for example, a Vendor Agreement will look for contract value, payment terms, and

governing law, while an NDA will prioritise confidentiality period and disclosing parties.

 

Step 2 — Extraction in progress

Once you tag a contract type, extraction runs in the background. A progress indicator is shown while the AI reads the document. This typically completes within a few seconds, depending on document length.

Screenshot 2026-07-15 at 7.49.49 PM.png

Screenshot 2026-07-15 at 7.50.09 PM.png

 

What happens during extraction

•     The AI scans the full text of the contract.

•     It identifies and extracts terms relevant to the tagged contract type.

•     Extracted values are validated and mapped to the appropriate fields.

•     Results appear as populated columns in your contract repository.

 

The Next button is disabled until extraction completes

During extraction, the Next button in the onboarding tour remains disabled. This is intentional:

you must review the extracted terms before proceeding, so the tour waits for the process to finish.

Once extraction is complete, the button becomes active automatically.

 

Reviewing extracted key terms

After extraction, key terms are surfaced directly on the contract detail page and as columns in the repository view. You can review, edit, or override any extracted value if the AI result needs correction.

 

Common key terms extracted by contract type

 

Contract type

Key terms typically extracted

NDA

Effective date, expiry date, confidentiality period, disclosing parties, governing law

Vendor agreement

Contract value, payment terms, effective date, expiry date, governing law, notice period

SaaS

Subscription term, renewal type, ARR, SLA commitments, data processing terms

Employment

Start date, probation period, compensation, notice period, jurisdiction

 

Managing columns and views

Extracted terms become columns in your contract repository. You can customise which columns are visible, reorder them, pin frequently used ones, and save custom views for different teams or workflows.

 

Column management actions

•     Show / hide: Show or hide columns

•     Reorder & pin: Drag columns left or right, or pin a column to keep it visible when scrolling.

•     Filter by extracted terms: Filter contracts by any extracted term — for example, show all contracts expiring in the next 90 days.

•     Group by: Group your repository by contract type, status, or any extracted field.

•     Save custom views: Create named views tailored to your team's needs and switch between them instantly.

•     Export: Export the current filtered view as a CSV for reporting.

masked_screenshot_blur.png

Was this article helpful?

0 out of 0 found this helpful

Didn't solve it?

Typical first reply under 4 hours, Mon–Fri.

Contact support