Action Items From Client Alert

Action Items from Client Alert

Extract time-sensitive action items, deadlines, and obligations from a client alert, regulatory bulletin, law firm memo, or similar legal update. Produce a deadline-organized checklist that an in-house lawyer can act on without re-reading the source document.

The skill is operational, not analytical. Action items are stated at the level of granularity the alert supports — typically high-level ("comply with new disclosure requirement by [date]"), not implementation-level ("update privacy policy section 4.2 by [date]"). The user knows their organization better than the skill does; the skill provides the trigger and deadline, not the implementation.

When this skill applies

Apply when the user provides a legal-update document and asks for action items, deadlines, or obligations to be extracted. Typical inputs:

• Law firm client alerts ("Client Alert: New SEC Cybersecurity Disclosure Rules Take Effect").
• Regulatory bulletins (FTC enforcement priorities; SEC staff guidance; agency Q&A documents).
• Law firm memos on regulatory developments or new statutes.
• Industry association updates distilling regulatory changes for members.
• Internal compliance bulletins if structured similarly to external alerts.

Do not apply when:

• The document is the underlying regulation, statute, or rule itself. Regulations are dense and structured for legal interpretation, not for action-item extraction. Recommend the user start with a client alert summarizing the regulation, or escalate to legal counsel for direct regulatory analysis.
• The document is a court opinion or judicial decision. Different shape; would warrant a case-summary skill (deferred enhancement candidate).
• The document is a private communication or strategy memo. Different shape; the skill is calibrated for published / shared alerts.
• The user wants an analysis of the legal change rather than action items. Action-item extraction is operational. If the user wants to understand the substantive legal change, recommend reading the alert directly or engaging counsel.
• The document is too vague to extract specific actions. Some alerts describe regulatory developments without imposing or describing specific requirements. The skill notes this and produces a "no specific action items extractable" output rather than fabricating actions.

Inputs

The skill requires the document. Optional inputs filter and prioritize the extraction:

• organization_context affects applicability. An alert about EU AI Act obligations is highly relevant to an EU-based AI vendor and largely irrelevant to a US-only consumer-goods retailer. The skill uses the organization context to flag items as applicable, conditionally applicable, or not applicable.
• relevant_business_areas filters extraction. If the user is focused on data privacy and the alert covers privacy plus employment, the privacy items get full extraction and employment items are noted briefly with a reference to re-run for that focus.
• applicable_jurisdictions affects which items are surfaced. An alert covering federal plus multiple states' laws produces different relevance for users in different states.
• alert_date affects deadline assessment. Some alerts arrive after some deadlines have passed; the skill should flag past-deadline items separately from forward-looking ones.

When optional inputs are not provided, the skill extracts all items from the alert and flags applicability uncertainty.

Workflow

The workflow has four steps.

Step 1: Document orientation

Before extraction:

• Confirm the document is a client alert or similar update. If it's something else (the regulation itself, a court opinion, internal correspondence), stop and route the user.
• Identify the alert's date. If not stated, ask the user (the alert_date input) before proceeding — deadline assessment depends on the alert date.
• Identify the alert's subject matter (regulatory regime, agency, statute, jurisdiction).
• Identify the alert's structure: is it organized by topic, by deadline, by stakeholder? Different structures affect extraction approach.
• Note any references to the underlying regulation, prior alerts, or related materials.

Step 2: Item extraction

Walk through the document and extract every item that meets one of these criteria:

1. Mandatory action item: something the user (or organizations like the user) is required to do, with a deadline. Source: the alert describes a new requirement, a compliance deadline, an effective date, a filing date, or a similar legal obligation.
2. Recommended action item: something the alert recommends as best practice without it being legally required. Source: the alert uses language like "should consider," "best practice," "we recommend," "prudent companies will."
3. Informational item: context that affects the user's understanding of the regulatory landscape but doesn't require specific action. Source: the alert describes background, agency posture, enforcement trends, or related developments without imposing requirements.
4. Conditional action item: action required only under specific circumstances (e.g., "if you process EU residents' data, you must..."). The skill flags the condition and the action.

For each item, capture:

• What: the action or obligation, stated in plain language.
• By when: the deadline or effective date, with explicit date if available.
• Who within the organization: typical functional owner if the alert suggests one (Legal, Compliance, IT, HR, etc.); otherwise marked "Owner: TBD."
• Why / source: the underlying legal basis (statute, regulation, agency guidance), with citation to the alert's section.
• Conditions / applicability: whether the item applies to all organizations or only specific ones.

Step 3: Apply filters and applicability flags

For each extracted item, assess applicability based on optional inputs:

• Applicable — the item applies to the user's organization based on organization_context, relevant_business_areas, and applicable_jurisdictions.
• Conditionally applicable — the item applies if a specific condition is met (the user has EU users; processes children's data; engages in cross-border data transfers).
• Not applicable — the item does not apply to the user based on the inputs provided.
• Applicability uncertain — the user's optional inputs are not specific enough to assess; the skill notes the item and flags applicability for user review.

If organization_context is not provided, all items are flagged "applicability uncertain" and the user must assess applicability themselves.