---
name: landfall-automation-wedge-scan
description: Identify the first bounded automation wedge inside a mid-market company and produce a human-review packet with value estimate, context map, risks, and next steps.
license: Proprietary-free-to-use
compatibility: Requires user-authorized read-only access to internal sources. Do not modify systems, send messages, create accounts, or change permissions.
metadata:
  author: Landfall Group
  version: "0.1"
---

# Landfall Automation Wedge Scan

## Use When
Use this skill when a human wants an agent to inspect approved internal context and identify where an AI agent could first create measurable operating value.

Good starting areas include sales follow-up, onboarding, renewals, implementation, support, collections, QA, reporting cadence, customer escalation, and internal operating reviews.

## Inputs
- Business area or function to inspect.
- Approved read-only sources, such as Slack, Google Drive, meeting notes, CRM exports, helpdesk tickets, shared spreadsheets, docs, policies, or operating reviews.
- Any known business metric or pain point.
- Any explicit exclusion zones.

## Procedure
1. Confirm the sources are read-only and user-authorized.
2. Identify repeated workflows with delay, rework, context hunting, missed follow-up, quality variation, or avoidable manual review.
3. Prefer workflows where AI can prepare, check, draft, compare, route, monitor, or escalate bounded work.
4. Score each candidate on business value, frequency, context availability, safety, owner clarity, implementation effort, and measurement quality.
5. Choose one recommended wedge and two alternates.
6. Produce a human-review packet.

## Output Packet
- Recommended wedge.
- Why this wedge is first.
- Current workflow summary.
- Systems and knowledge sources involved.
- Estimated time saved or value created, including assumptions.
- Required AI responsibilities.
- Human responsibilities and approval points.
- Risks, permissions, audit trail, and quality checks.
- First 30-day validation path.
- Open questions.

## Guardrails
- Do not send messages, update records, move files, create API keys, change permissions, or launch automations.
- Do not claim precision in ROI when the evidence only supports a range.
- Do not recommend production autonomy without human checkpoints and audit trail.
- If no good wedge exists, say so and explain what evidence is missing.

## Contact
Landfall Group can review the packet at hello@landfallgroup.co.