SendPulse
209 conversations, 37 qualified leads
A personal-injury law firm replaced a plain contact form with an AI intake assistant on its website, built in SendPulse. The assistant works in English or Spanish as parallel flows. It states the attorney-client disclaimer before it asks for a single detail. It then triages by case type in a fixed order: case type, who was injured, when, and where. Every path ends in one contact capture flow: name, phone, and consent to text. Staff can pause the automation for an hour or step in at any point.
How it wentIn one year of platform data: 209 conversations started, 1,444 questions answered with no staff time, and 37 qualified, consented leads captured. That is 17.7% of conversations. The firm signed 36 cases in the same window and attributes them to this intake path, though the source does not trace each signed case to the assistant lead by lead (agency case study figures).
Try it · based on the source's workflow
Act as the intake assistant for [law firm] in [languages]. Before you ask for any detail, show this disclaimer: [paste]. Then ask, in this order: case type, who was injured, when it happened, and where. Keep each question short. End every path by asking for a name, a phone number, and clear consent to text. Never give legal advice or say whether a case is valid. Save the full conversation for staff review, and let a human take over at any point. Show me the full flow before it goes live.
A starting template based on the workflow the source describes. Review it before you use it; you are responsible for what it does.