Automating Insurance Claims with an AI Agent
Strive Community Care is a Brisbane based NDIS registered provider and spent a massive amount of time preparing property damage insurance claims. Osher Digital automated the claim preparation drafting process with an AI Agent.
Overview
Strive Community Care’s team were spending a massive amount of time preparing property damage insurance claims.Tedious admin was pulling staff away from high-value care work. Leadership approached Osher Digital to create an AI Agent to draft the documentation.
About the Client
Strive Community Care is a medium-sized NDIS provider focused on delivering high-quality support for children and young people across Australia.
Problem
Strive staff deal with a high number of property damage events. Each claim required:
- Pulling incidents from ClearCare
- Gathering reports, quotes, photos and supporting docs
- Extracting the right details
- Manually filling a full insurance claim form
It had been a pain point for years , chewing up staff time and slowing down turnaround. Leadership wanted admin time reduced so their team could focus on actual care, not paperwork.
Solution
We built an end-to-end AI Agent that handles the heavy lifting.
- A ClearCare webhook triggers the workflow whenever a new incident is logged
- All documents (incident reports, quotes, images) are fed into OCR
- An AI extraction agent identifies key fields: dates, costs, responsible person, damage details
- Data flows into a Carbone template , auto-generating a completed insurance claim form
- Output is delivered back to Strive ready to send to insurers
The technical stack for this project consisted of the following:
- Self-hosted n8n running on Docker
- Locally hosted Carbone.io templating system
- Open Router to allow switching LLMs
- Clear Care CRM integration via webhooks.
The biggest hurdle was ClearCare’s lack of API documentation, so we engineered a custom webhook workaround to keep the automation clean and reliable.
Results
Early outcomes are already clear:
- ~70% reduction in manual data entry
- Faster turnaround on every claim
- Higher consistency across documentation
- Less stress and admin burden on frontline teams
- Increased time spent on meaningful care work, not paperwork
The workflow began delivering value as soon as it went live.
Key takeaways
- Systems with strong APIs make automation dramatically easier
- Webhook-based integrations can outperform traditional APIs when vendor support is limited
- n8n allowed for very fast iteration and refinement
