AI and the drive to digitise

I wrote this article just before Easter when, across the country, parents were arranging egg hunts and calculating how much of their children’s chocolate haul they could pilfer without detection. At the same time, conveyancers were basking in a different kind of seasonal relief.

The stamp duty rate changes announced in September 2022 finally took effect on 31 March 2025, bringing a merciful end to the chaos that preceded it. The final week of March, as one weary colleague put it when I asked how work was going, felt like something straight out of a Dickensian novel – frantic, desperate, and punctuated by the occasional existential crisis. But now that the deadline has passed, relative calm has been restored … at least for now.

Building targets

The respite, however, is likely to be short lived. The government have set themselves a steep target of building 5 million new build properties over the next five years. The Office for Budget Responsibility (OBR) predicts that planning reforms may lead to the highest levels of house building in four decades, potentially culminating in 300,000 new builds hitting the market per annum by 2029. Policy aside, broader macro-economic trends and a growing appetite for lending within the financial sector (meaning easier to obtain mortgages) have led many to predict that property transaction volumes could soon reach and surpass the record levels last seen in 2007.

This should sound like an unqualified win for conveyancers: more transactions, more fees, and more reason to keep that celebratory bottle of fizz permanently on standby. However, the number of fee-earning conveyancers in the UK has shrunk by somewhere between 9% and 15% since the COVID-19 pandemic in 2020, and while transaction volumes may be heading skyward, the workforce responsible for handling them is looking decidedly lean.

Digitising conveyancing

Fortunately, some progress has been made. HM Land Registry (HMLR) is expanding the digitisation of their services and much of this improved functionality is already being integrated into the digital tools used across the profession. From accurate digital identity verification to the digitisation of 95% of traditionally paper-based transactions, such as applications to HMLR, the capacity to conduct business digitally has expanded significantly. This has reduced both administrative burdens and human error (goodbye, never-ending requisitions). Soon, artificial intelligence (AI) will) play a major part in the way that property transactions work. For example, know your client (KYC) tools across the industry have already introduced AI-powered technology into their workflows. This has led to a substantial reduction both in the volume of errors and in successful malicious activity – following the judgment in Dreamvar v Mischon de Reya the technology industry listened.

Technology will continue to deliver advancements in improving processes and while incremental improvements are impressive in their own right, upcoming changes will be nothing short of transformative.

The latest generation of AI, already in beta testing for select users, is proving to be very impressive. In comparison, ChatGPT and others like it are little more than glorified search engines – adept at sifting through information, sure, but ultimately just better ways to google things. That iteration, known as single-modal AI, was limited to receiving and responding in one format: text in, text out. AI agents, however, are an entirely different beast. These are multi-modal powerhouses that can digest a voice recording of a poetic citation, analyse a series of Baroque paintings, and then craft an essay offering a nuanced critique of 17th-century artistic melancholia – all without breaking a (virtual) sweat.

The real game-changing potential lies in their ability to interact directly with your computer. With control over your mouse and keyboard, AI agents can take on the drudgery of administrative work. An agent can, for example, be instructed to log into Outlook every evening, scan through client emails for update requests, cross-check case progress in your practice management system, and then draft tailored responses – ready for a human’s final approval before sending. These small efficiencies – five minutes saved per email, per client, per day quickly stack up. This allows overburdened conveyancers to focus on higher-value (and billable) tasks while clients benefit from a faster, more efficient service.

Usurper or aid?

When discussing AI’s role in the legal profession, many solicitors express concerns about job security, but such apprehensions are premature. In chess, for example, Grandmaster Magnus Carlsen boasts an Elo rating surpassing 2800, while GPT-4’s estimated rating hovers around 1250 – placing it closer to a keen amateur than a world champion. AI currently serves better as an assistant than as a substitute for nuanced human judgment. It‘s smart, but it’s not yet even close to being smarter than us.

Moreover, the environmental footprint of AI is substantial. Single training sessions for sophisticated models like GPT-4 reportedly consume approximately 700,000 litres of water, comparable to the annual water usage of several hundred households. Additionally, generating a mere 100 words with GPT-4 can consume up to 1.4 litres of water, depending on the data centre’s location. The significant resources required to power and cool AI systems, suggest that the pace of its advancement may be tempered by the need to develop more sustainable infrastructure.

Conveyancers can, for now, set aside their existential dread of AI as a usurper. The future of the profession is not one of obsolescence but rather evolution. With initiatives to streamline and digitise the homebuying process and a buoyant property market on the horizon, the outlook is decidedly rosy.