- Home
- Skills
- Workflow Automation
- business automation architect
business automation architect
Architects implementable, fail-safe business automations with explicit error handling and decision logic.
$5
Works with the AI tools you already use
business automation architect
Example session with this skill installed
We manually process customer refund requests. It starts with a Zendesk ticket, then we check Stripe for the transaction, update a Google Sheet, and finally email the customer via Gmail.
- Read your context and instructions
- Compiled the business automation architect
- Generated the UI component
I have analyzed your refund workflow. It is highly suitable for a low-code or custom-code hybrid to ensure data integrity across Zendesk, Stripe, and Google Sheets.
Business Automation Blueprint: Customer Refund Processing
1. Process Map
| Step | Action | System | Data Points | Decision |
|---|---|---|---|---|
| 1 | Trigger | Zendesk | Ticket ID, Customer Email | - |
| 2 | Lookup | Stripe | Transaction ID, Amount, Status | Is refundable? |
| 3 | Update | Google Sheets | Refund Date, Status, Link | - |
| 4 | Notify | Gmail | Customer Email, Template | - |
2. Automation Architecture
- Platform: n8n (Self-hosted) or Make.com.
- Trigger: Zendesk Webhook (Ticket Created/Updated with tag
refund_request). - Idempotency: Use Zendesk Ticket ID as the lookup key in Google Sheets to prevent duplicate entries.
3. Decision Matrix
- Condition: If Stripe transaction is older than 60 days OR status is 'disputed'.
- Rule: Set
Internal Notein Zendesk and route to "Human Checkpoint: Senior Finance".
4. Error Handling Plan
- Stripe API Timeout: 3 retries with exponential backoff.
- Google Sheet Full: Alert Slack
#ops-alertsand move payload to Dead Letter Queue (PostgreSQL). - Auth Failure: Immediate notification to Admin; pause execution to prevent data loss.
5. Implementation Logic
- Receive Webhook from Zendesk.
- GET
/v1/chargesfrom Stripe using metadata email. - IF
charge.amount> $500 ORcharge.created< 60 days ago:- POST human approval request to Slack.
- WAIT for interaction.
- ELSE:
- POST
/v1/refundsto Stripe.
- POST
- APPEND row to Google Sheet
Refunds_2024. - SEND Gmail using
refund_confirmationtemplate.
Next steps
- Generate API keys for Zendesk (Admin) and Stripe (Restricted Key with Write Refund access).
- Create the Google Sheet with headers:
Ticket_ID,Stripe_ID,Amount,Status,Processed_At. - Configure the Zendesk Webhook to fire only on specific tag triggers to save execution costs.
business-automation-architect.tsx
TSX · React component
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
What you get
About this skill
The problem
Vague requests like "we need to automate this" often lead to fragile workflows that break on the first API error or duplicate data because they lack idempotency. Without a formal architecture, developers waste hours debugging edge cases and missing triggers in undocumented, "tribal knowledge" processes.
What it does
- Extracts implicit manual steps into a structured, multi-phase Process Map.
- Evaluates and selects the optimal platform tier between no-code, low-code, or custom code.
- Architects error handling plans including retry policies, backoffs, and dead-letter queues for every step.
- Designs human-in-the-loop checkpoints with explicit SLAs and escalation logic.
- Produces implementation-ready pseudocode that ensures idempotency and safe re-runs.
Frameworks & tools
Zapier, Make (Integromat), n8n, Power Automate, Python, Node.js, and Playwright/RPA for legacy UI-only systems.
Why this beats prompting it yourself
Generic prompts often skip critical operational requirements like rate limit handling, state persistence, and manual overrides. This skill enforces a 4-phase engineering discipline that ensures the automation is maintainable and recoverable, not just a "happy path" script that fails in production.
Use cases
- Converting a messy spreadsheet-and-email approval chain into a structured n8n workflow.
- Designing a high-volume lead ingestion pipeline that requires custom Python logic for data enrichment.
- Architecting a fail-safe synchronization between a legacy CRM (no API) and a modern SaaS tool.
- Performing a feasibility and ROI audit before committing engineering resources to a complex integration.
Known limitations
Requires the user to provide specific details about existing data sources and triggers. It does not perform live API testing or write final deployable code.
How to install
Works the same in every agent - Claude, Cursor, Codex, Copilot and 20+ more.
- 1
Download the ZIP
Free skills download straight away. Paid skills unlock right after purchase.
- 2
Unzip into your skills folder
Every agent reads skills from one folder on your machine. Drop the unzipped folder in there.
- 3
Ask your agent to use it
Restart the agent if it was already running. It picks the skill up automatically - no config needed.
Skills folder by agent
Click the path to copy it. Create the folder if it does not exist yet.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill passes our trust checks below.
Security scanned
Passed our 8-point scan before listing
Fresh listing
Recently published to Agensi
30-day refund
Not a fit? Get your money back
Trust & safety
Security scanned
Verified clean 9 days ago
- Passed all security checks, Safe to install