You are helping me build the following AI agent workflow.

## Goal
Insurance Premium Adjustments With Telematics: Fetches telematics data via HTTP, analyzes driving behavior with Claude to calculate risk scores, updates premiums in policy systems.

## Specification
- What it does: You get fairer insurance rates because this agent analyzes your actual driving habits via telematics to lower your premium when you drive safely. It calculates personalized risk scores based on your acceleration, braking, and speeding patterns, then automatically updates your policy without manual paperwork.
- Trigger: On a schedule (Scheduled · daily analysis)
- Autonomy: Fully hands-off
- Expected setup effort: about an afternoon
- Tools/services involved:
  - HTTP Request: Fetch telematics data and update policy systems
  - Anthropic Chat Model: Analyze driving behavior for risk scores
  - Structured Output Parser: Parse AI results into structured data
  - Gmail: Alert underwriting managers on high risk
  - Slack: Notify claims teams of updates

## Known pitfalls, handle each one explicitly in your implementation
1. Raw telematics payloads can be malformed or missing fields, so add error handling nodes to skip invalid records gracefully.
2. LLM responses may occasionally violate the JSON schema, so implement a retry mechanism with stricter prompt constraints if parsing fails.
3. Sensitive location data might be included in raw feeds, so ensure you strip PII before sending data to the LLM for analysis.

## Reference implementation
https://n8n.io/workflows/12279 (workflow template)
Fetch and inspect this before building. If it matches my stack, adapt it;
if not, rebuild the pattern with my tools.

## Process requirements
1. Before building: ask me which of the listed tools I actually use and
   what my platform is (n8n / Make / code / other). Do not assume.
2. Adapt the pattern to my answers; do not force the reference stack.
3. Address every pitfall above; tell me how you handled each.
4. Provide a test plan I can run before letting this touch real data.
5. Ask before any step that sends messages, modifies data, or spends money.

Source: https://usecasesforagents.com/use-case/automate-insurance-premium-adjustments-with-telematics/ via usecasesforagents.com