Case Study
Project: AI for Lead Management
Project Summary: I designed and built an automation for Stuart Miller Solicitors to streamline their client enquiry processing. The app categorises enquiries by offence and fee potential then forwards it to the appropriate lawyer.
Introduction
The Stuart Miller Solicitors website generates over 200 daily enquiries from prospective clients. These clients have been arrested or charged with a criminal offence and are urgently seeking help.
Managing these enquiries requires a team of three experienced staff based at the Stuart Miller office. These paralegals evaluate what offence is alleged to have been committed, its seriousness, and which lawyer at the firm is best suited to handle the case. As the more severe cases generate higher fees for the firm, it is important these cases are prioritised with an immediate response.
This project aimed to use AI to free up the time the paralegal team spent managing client enquiries. Doing so would allow them to work on client cases and generate more fees for the firm.
Project Overview
I designed a workflow process using a combination of Zapier, OpenAI’s GPT API and some bespoke Python code generated with ChatGPT.
The automation first parses the content of the web enquiry. The GPT API was trained on past examples of enquiries evaluated by the human team. The AI then predicted the most likely offence with which the enquirer would be charged, whether the case could lead to a Crown Court trial (more serious) and which lawyer at the firm was best suited to handle the case. The AI was essentially replicating the same process as the human team.
Example enquiry:
“So, last month, my bruv had a few too many. This bloke starts having a go. We get into a mad scrap, and my bruv goes all out. Ends up, other blokes now got some serious damage. Now we gotta sort this mess in court in a few weeks.”
AI Classification:
- Likely offence: Grievous Bodily Harm (GBH)
- Likely court: Crown Court
- Selected lawyer: Nick Pardew
Process and Development
After configuration and initial testing, we passed all enquiries through the AI and the human team. An experienced lawyer was assigned to assess the accuracy of the evaluations in a blind test over a three-month period. At the end of the testing phase, the AI proved significantly more accurate than the human team. This process is now entirely managed by AI.
Results and Impact
This project was designed, configured and implemented in just three days. The ongoing running costs are negligible. The automation has freed up the time of three qualified paralegals, allowing them to work on fee-generating activities. The process has increased accuracy and the speed with which new clients receive a response. This has resulted in increased revenue for Stuart Miller Solicitors. The paralegals now enjoy greater job satisfaction handling actual client cases instead of processing client enquiries.
What can I build for you?
You may be surprised at the affordability with which AI projects can be designed and implemented. If you have a problem that could be solved with AI, please get in touch.