← Case Studies/docket-leads-ai
CASE STUDY

How Thompson Legal Stopped Losing Leads to a Glovebox

Their tow truck drivers were filling out paper dockets at accident scenes. Those dockets sat in gloveboxes for three days before anyone typed them up. By then, the leads were gone.

Thompson Legal is a personal injury firm in Birmingham. They work with a network of three tow truck operators who attend accident scenes, assist with vehicle recovery, and — crucially — collect contact details from everyone involved. These details are written by hand on paper dockets: name, phone number, vehicle make and model, plate number, and a brief description of the incident.

The dockets are the firm's primary source of new leads. Every accident is a potential personal injury case, and in this business, the firm that makes first contact almost always wins the client.

The problem was the pipeline. Drivers would collect dockets throughout the week and drop them off at the office on Friday. A data entry clerk would then spend most of Thursday afternoon typing them into the CRM. That meant a docket written on Monday morning sat in a glovebox until Thursday afternoon before anyone at the firm even knew the lead existed. Four days. In personal injury, four days is an eternity. Competitors with faster intake were calling these victims on Monday afternoon.

The firm estimated they were losing 20-25% of potential cases simply because they weren't fast enough. And the data entry itself was error-prone — one misread digit in a phone number from messy handwriting, and the lead was dead regardless of speed.


What changed

We built DocketLeads AI to collapse the entire pipeline. Now, when a tow truck driver finishes at an accident scene, they snap a photo of the handwritten docket and send it via WhatsApp. That's it. No app to install. No login. Nothing new to learn — they already use WhatsApp all day.

The system processes the image through two OCR engines simultaneously. One handles contextual understanding of messy handwriting. The other handles raw text precision. The results are cross-validated, and the extracted data — name, phone, vehicle, plate, location — is pushed directly into the firm's Google Sheets intake sheet, fully structured and ready for immediate follow-up.

The entire process from photo to lead: under 30 seconds. The intake team now works a live feed of fresh leads instead of a Friday afternoon backlog. Nobody types anything. Nobody misreads a phone number and kills a lead.


The numbers

4 days
→ 30 seconds
26%
More cases won
0
Data entry hours
0
Transcription errors

“The first week this went live, we got a lead at 9:14am and had an associate on the phone with the victim by 9:22. The other side hadn't even processed the police report yet. That's the difference speed makes in this business.”

— James Thompson, Managing Partner, Birmingham


Got a similar problem? Book a free audit →