
Panasar: Empowering lawyers
A technology-only law firm built around a “fully deterministic legal model” has been authorised by the Solicitors Regulation Authority (SRA), the first of its kind.
LawFairy’s system encodes the law into structured decision pathways to reach an auditable output, in contrast to generative AI, which uses generates responses by predicting statistically likely text.
It is just the second technology-only firm to be authorised, following Garfield last year – but Garfield uses an AI model.
LawFairy is set to go live next month initially in immigration law, an area where founder Raj Panasar told Legal Futures there was “a real gap between the need and access”.
It will help consumers, sponsors and law firms check eligibility visas, nationality and settlement, and create some of the documents needed to apply for them.
If consumers and sponsors want legal advice on the next stages of the process – having received an eligibility report at a cost of £149 – they will be passed to other law firms under partnership arrangements LawFairy is developing.
For law firm users, who will be able to take out subscriptions, the aim is to cut out the process element of the work.
Mr Panasar stressed that the system “isn’t intended to get rid of the lawyers, it’s not intended to disintermediate them. It’s very much intended to empower them… I think the margin improvement for law firms will be very significant.”
Indeed, if Lawfairy detects that a user is vulnerable, it will advise them to seek advice earlier in the process.
Mr Panasar was a partner in the London office of US firm Cleary Gottlieb and then a corporate finance partner at Hogan Lovells, before quitting in 2021 to build the system, having learnt how to code.
“It was a very big decision… but I wanted to create a company dedicated to making legal work more accessible, quicker, and making lawyers more productive.”
He has coded the system to follow the intricacies of UK immigration law, essentially putting guardrails in place so that the same facts produce the same result every time.
Each outcome is produced through the verified, auditable application of rules. Mr Panasar said it had been tested thousands of times, including by immigration lawyers, to ensure it provided the correct answers. As a result, it does not require ongoing supervision.
In the event the law changed, whether through legislation or a court ruling, Mr Panasar would simply adjust the ‘guardrails’ accordingly. He said AI models, which were “heavily influenced by what’s happened in the past”, would struggle with this.
Though some called the type of technology LawFairy used ‘symbolic AI’, he said he preferred not to refer to it as AI because the term was now associated with probablistic large language models, like ChatGPT and Microsoft Copilot.
The SRA authorisation process was “rigorous”, Mr Panasar said, but the key “was when they understood that the rules are embedded into the system” and decisions were auditable. It was the same with professional indemnity insurers.
He would not say which areas of law he hoped to move onto but the focus would be on those with high levels of regulation and defined criteria.
Mr Panesar added: “Most legal AI produces probabilistic outputs – statistically likely answers generated from patterns in data. That is fundamentally unsuitable for regulated legal work, where an outcome is either right or wrong.
“The law contains vast areas governed by precise rules – statutory tests, defined thresholds and fixed eligibility criteria. These do not require discretion. They require disciplined, consistent application. Deterministic technology is designed precisely for that task.
“Authorisation demonstrates that our model, when structured around verifiable rules and robust governance, can meet the same regulatory standards as any other authorised firm.
“We believe it also opens regulated legal services to people who currently find them too costly or uncertain to pursue.”
LawFairy will next week also announce a pro bono initiative with a leading charity and a global law firm.












Leave a Comment