Year
2022
Client
Clive Quinn
Category
Professional Services
Product Duration
3 - 4 Weeks
We spent three days shadowing the intake team, mapping every step from first contact to case assignment. We identified that 70% of incoming enquiries were from four recurring case types, each requiring the same set of initial information. We also found that response times beyond 2 hours significantly reduced the likelihood of a client progressing. Speed and consistency were the two biggest levers.
We designed a conversational AI intake flow that could run across web chat, email reply, and SMS. The flow was structured around the four main case types, with branching logic for edge cases. Each path collected the minimum viable information needed for case assignment, with plain language prompts written to match the firm's existing client-facing tone. We also designed an urgency scoring system that flagged time-sensitive matters for immediate human review.
We built the intake system using a large language model for natural language understanding and a structured decision engine for routing logic. It integrated directly with the firm's existing case management software via API, auto-populating client records and assigning cases without any manual entry. A fallback protocol escalated anything the system was not confident about to a staff member within 5 minutes. Full build and testing took 3 weeks.
80% reduction in admin time on intake
