What workflow was being improved?
The starting point was a commercial-underwriting process with a three-day submission-to-decision interval. The work focused on ACORD submissions, loss-run extraction and risk-memo generation. The reported result describes the time to an underwriting decision and staff capacity, not the number of prompts or documents a model produced.
The customer is anonymous in this public case study. The approved record establishes the workflow inputs and reported outcomes, but not the customer’s implementation calendar, carrier systems or internal decision policies. Those details are not inferred from the broader Encapsa platform architecture.
How is Jordan’s contribution attributed?
This is a selected delivery by the Encapsa team I lead as Co-founder & CTO. The individual responsibility is my engineering and platform leadership; the outcome is attributed to the team. The public case study does not claim that I personally authored every stage or that the result applies to all customers.
The broader Encapsa role includes an eight-person full-time team and customer instruction during every deployment. Those are platform-leadership and enablement facts. They are not additional measured outcomes for this anonymous underwriting engagement, and this page keeps those scopes separate.
What did the automation do?
The workflow automated three documented activities: ACORD submission ingest, loss-run extraction and risk-memo generation. The visual sequence presents those activities in their stated order. It describes the selected delivery at a high level, without inventing a model choice, decision threshold, review policy or customer-system topology.
These stages provide a concrete description of the work between a submission and a decision. The public record does not establish a specific OCR product, prompt, underwriting rule engine or percentage of decisions made without human involvement. The case is therefore framed around the confirmed workflow and reported result.
- Ingest ACORD submissions
- Extract loss-run information
- Generate a risk memo
What changed after the delivery?
Reported submission-to-decision time changed from three days to 15 minutes. The delivery recaptured 300+ staff hours per month without adding headcount. These are distinct workflow and capacity measures: elapsed decision time, monthly staff hours and headcount. They are not combined into an invented financial return or accuracy percentage.
The comparison uses the original units rather than translating days into an assumed number of working hours or claiming a derived reduction percentage. No customer dashboard or before-and-after screenshot is fabricated. The visual is a searchable reported-outcome summary drawn from the approved record.
What does the record show?
| Measure | Record | Scope |
|---|---|---|
| Submission to decision | 3 days to 15 minutes | Reported elapsed time to an underwriting decision |
| Staff capacity | 300+ hours per month | Recaptured staff hours |
| Headcount | No added headcount | Reported with the capacity result |
| Workflow | ACORD ingest, loss-run extraction and risk memos | Three documented stages |
| Attribution | Selected Encapsa team delivery | Customer identity remains private |
Which skills does this project demonstrate?
- Workflow AutomationAutomated the underwriting workflow across three documented stages.
- Document ProcessingACORD submission ingest and loss-run extraction.
- Artificial Intelligence (AI)Risk-memo generation within an operational underwriting workflow.
- Systems IntegrationConnected submission inputs, document extraction and memo generation.
- Software ArchitectureSelected production delivery by the Encapsa team Jordan leads.
What does this case study not claim?
- Reported Encapsa team outcome; customer identity remains private.
- The original metric is submission-to-decision time, not model response latency.
- Implementation dates and whether this specific engagement is still active are not established. No derived ROI, decision-accuracy or autonomous-approval claim is added.