Graduates develop lead generation chatbot for law firms
Prediction algorithm combined with machine learning will deliver quality referrals, LawBotX team promises
The group of Cambridge graduates behind LawBot, the chatbot they developed while studying for their degrees, have relaunched the project as a lead generation network for law firms combining case outcome prediction algorithms with a machine-learning client selection filter.
LawBot started off as a project to help victims of crime after Ludwig Bull, the original founder, realised when volunteering for a pro bono initiative on sexual offences that existing material designed to help victims wasn’t as user friendly as it could be. Six weeks later, in late 2016, and with just £300 of investment, it was up and running.
At that point, the concept was similar to DoNotPay, the bot developed by US student Joshua Browder to help appeal parking fines. By spring this year, it had grown to cover 26 offences and started handling divorce claims. Now rebooted as LawBotX, the project is being rolled out on Facebook Messenger in seven countries.
LawBotX’s UK version has an additional feature specifically developed with law firms in mind. The process that powers the automated conversations with users helps identify the problem and narrow down the issue. Combined with LawBot’s own algorithm, which calculates the user’s chances of winning, it acts as a funnel to identify the strongest leads.
With 1.7 billion people and 65 million businesses on Facebook, the social media platform is the place where firms should look for new business, says Rebecca Agliolo, LawBot’s marketing director (pictured). ‘A chatbot can engage users in their natural habitat, the familiar medium of their daily messaging platform.’
Firms will have a choice between letting Facebook Messenger drive traffic to their site or installing a widget on theirs – or both. But it is the triage component of the process, which grades the seriousness of a lead based on the bot’s prediction algorithm, which could be the greatest attraction for law firms if it proves to be reliably accurate. ‘Unlike prediction systems developed by other legal tech start-ups, LawBot doesn’t use decision-tree reasoning’, Agliolo told Solicitors Journal, ‘and it has a 71 per cent accuracy rate’.
Law firms have become wary of collective marketing initiatives thrown at them promising quality leads that don’t live up to expectations. But Agliolo is confident the process will deliver the efficiencies and the leads quality law firms demand. The prediction algorithm, she says, acts as a disincentive for users with weak claims or who are unwilling to pay for advice.
At the other end, law firms can build in specific questions to filter potential clients further based on their own requirements or field of expertise. What’s more, machine-learning algorithms will ensure that the system ‘learns’ from the law firm’s choices.
Initially, law firms will be presented with all potential clients based on their original preferences. These will be gradually – and continually – refined and adjusted. ‘They can then choose to delete, keep or archive users,’ Agliolo says. ‘Whether deleted or retained, the algorithm that matches clients to law firms develops an understanding of the firm’s preferences. The quality of the leads improves over time as the system learns the preferences of law firms that actively manage their accounts.’
At the moment LawBotX is being proposed as what is effectively a beta version to a select number of firms – most of them on the private client side. But Agliolo and her business partners are looking at bringing 50 firms on board in the next six months.
As to costs, the pricing structure hasn’t been finalised yet. Agliolo suggested the price could be per lead and range from ‘£5 to £100 per lead generation’. A clearer commercial picture should emerge in the autumn, where rates will be ‘based on the value created’.
Jean-Yves Gilg, editor-in-chief
jean-yves.gilg@solicitorsjournal.co.uk
