After one demo, Licious's head of customer success built the first version. Parallel Loop then helped refine a workflow that classifies complaints, updates the tracker and reports to the team lead.
- Customer's first build
- 4-5 hours
- Complaint triage and updates
- 13 runs/day
- Parallel Loop optimisation
- Up to 6 hours
- Typical delivery timeline
- 2-3 months
The challenge
Customer complaints reached Licious by email and phone. Each complaint needed a priority, a place in the tracker and a way for the team lead to follow progress.
What we built
After seeing a demo, Licious's head of customer success built a working first version in about four to five hours. Parallel Loop then optimised it over two to three sessions, totalling no more than six hours.
The workflow runs thirteen times a day. Each run brings new complaints into the same classification, tracking and reporting process.
- Reads the current configuration and retrieves every new email thread.
- Uses AI to classify complaints in parallel as P0, high or medium priority, with ongoing learning on each run.
- Updates the internal tracker and sends an hourly report to the team lead.
- Handles voice escalations through the same process. Recorded support calls are transcribed by a voice-to-text service, then passed into classification and tracking.
The impact
Complaint classification and hourly reporting now run automatically, thirteen times a day. A business leader built the first version after a single demo, then refined it with Parallel Loop.
- Email complaints and voice escalations feed one tracking workflow.
- The customer built version one in four to five hours.
- Two to three optimisation sessions added no more than six hours of work, against a typical delivery timeline of two to three months.