Shared scenario report
Draft multilingual blog posts, a Facebook post and an email with Claude and Wordpress
Analyzed
Executive Summary
Make.com
Cost ExposureMedium
AI model usage — LLM nodes add per-run token cost to every execution.
100/ 100
Overall
Risk: LowCost: Medium
Reliability
100
Security
100
Cost Risk
100
Observability
100
AI Governance
100
Naming
100
Complexity
100
Modules
4
Connections
3
External calls
—
AI modules
—
AI via HTTP
—
Loops
—
Unreachable
—
Max depth
—
Controls Matrix
● Present◑ Partial● Missing
| Check | Description | Status |
|---|---|---|
| Failure Alerting | Failures silent — only visible in execution log | Missing |
| Retry on Failure | Transient failures cause permanent stops | Present |
| Request Timeouts | Slow providers can block execution indefinitely | Present |
| Duplicate Prevention | Retry on write events creates duplicate records | Missing |
| Empty Response Handling | Zero-result lookups break downstream logic | Present |
| AI Output Limits | Output cost and parsing risk unbounded | Present |
| AI Output Format | Free-form output will break downstream parsers | Present |
| Activity Logging | Side effects unlogged | Present |
| Write Deduplication | Webhook replays may create duplicate records | Present |
No findings detected.