There is a lot of noise in the AI market right now.
Every week, there seems to be a new tool, a new announcement, or a new claim that everything is about to change overnight. I work in this space every day, and even I think it can be hard to separate what is genuinely useful from what is just well-packaged hype.
From where I sit, the more useful question for legal and conveyancing practitioners is not “Is AI coming?” It clearly is. The better question is: where does it fit in your workflow, what can it genuinely help with today, and what still needs human judgement?
That is where I think the conversation gets more practical.
One of the mistakes people make is treating AI as if it is one category. It is not. It is an ecosystem.
Some tools are built to help people think faster, draft faster and summarise faster. Some are being tailored specifically for legal work. Some are being embedded into the systems practitioners already use to move a matter forward. And some are not really “AI tools” at all, but the trusted data sources and transaction rails that make AI useful in the first place.
That distinction is essential. If you don’t understand which layer a product sits in, it becomes very easy to expect the wrong thing from it.
This is the category most people have now seen or tried in some form. Tools like ChatGPT, Claude and Microsoft Copilot are very good at helping people get started. They can summarise information, help structure ideas, produce a first draft, and reduce a lot of the blank-page friction that slows professionals down.
That is valuable.
But on their own, these tools are not legal workflow platforms. They are not connected by default to the systems of record, the matter data, the registries, or the transaction infrastructure that legal and conveyancing work depends on.
So yes, they can save time. But no, they do not remove the need for verification, supervision or professional judgement. In my view, they are best understood as productivity layers, not end-to-end practice layers.
This is where things start becoming more interesting for practitioners.
Legal-specific AI tools are trying to solve a different problem from general-purpose assistants. They are not just trying to generate language. They are trying to create legal confidence.
That usually means they are built around legal content, citations, drafting support, document analysis, review workflows and more structured outputs. In other words, they are designed to be more useful in the context of how lawyers actually work.
I think this is an important shift. Because the gap between AI that sounds smart and AI that is professionally useful is still quite large.
If a tool is helping with internal brainstorming, a general assistant might be enough. But if it is supporting legal reasoning, drafting that will be reviewed externally, or work that affects client outcomes, the threshold is much higher. Practitioners need more than fluency. They need context, structure, traceability and something they can interrogate properly.
This, in my opinion, is where the real long-term value sits.
The legal profession does not run on isolated prompts. It runs on workflows. Matters move through stages. Documents need to be reviewed, signed, lodged, checked, shared, updated and completed. Different people touch the work at different points. Different systems hold different pieces of the truth.
The real test for AI is whether it can reduce friction inside the actual workflow. That could mean helping a practitioner progress a matter more efficiently. It could mean surfacing the next likely action. It could mean reducing copy-and-paste between systems. It could mean identifying missing information earlier. It could mean making collaboration with counterparties less fragmented.
That is a much more useful frame than asking whether AI can “replace” part of the legal profession. In most practical scenarios, the better outcome is not replacement. It is a better operating model.
This is the part that gets less attention, but I would argue it matters the most.
AI is only as good as the data it can access.
I say that often because I think it is still underappreciated. When an AI tool doesn’t have access to a trusted source of record, it fills the gap with probability. Sometimes that works fine. Sometimes, not. In a regulated environment, that is a problem.
In legal, conveyancing and property transactions, the work often depends on authoritative external data. Land titles. Company records. identity checks. Compliance systems. Government registries. Settlement networks. These are not nice-to-have inputs. They are the foundation of the workflow.
If AI cannot connect to those sources in a governed and reliable way, then its usefulness hits a ceiling very quickly.
At that point, it may still be a helpful assistant. But it is not yet a productive operator.
That is why I think the next phase of AI adoption in this space is less about bigger claims and more about better connections. Better connection to trusted data. Better connection to real workflows. Better connection to systems that can support execution, not just commentary.
Prop tech is relevant here because clients do not experience property transactions as disconnected categories.
They move across agents, brokers, lenders, lawyers, conveyancers, settlement platforms, data providers and internal business systems. From the client’s point of view, it is one journey. From a systems point of view, it is usually many.
That is why the prop tech ecosystem is foundational, even when the tools themselves are not direct competitors to legal workflow platforms. They shape the quality of information, the speed of handoffs and the amount of manual work still hidden in the process.
What I find interesting is that AI is now being used across this broader ecosystem to turn volume into clarity. Large datasets are being summarised faster. Operational activity is being made more visible. Teams are being helped to triage, review and respond more efficiently.
For practitioners, the big takeaway is this: the better those systems connect, the more AI can reduce friction across the whole transaction, not just within a single isolated task.
If I strip away the hype, I think AI is already proving useful in five practical ways.
First, it is helping people read faster. Long documents, notes, files and communications can be summarised quickly, which gives practitioners a better starting point.
Second, it is helping with first pass drafting. That includes correspondence, internal notes, matter summaries and other repeatable writing tasks. Used properly, this saves time. Used carelessly, it creates more review work. The difference is usually context and supervision.
Third, it is helping identify issues earlier. When connected to the right data, AI can surface anomalies, missing information and follow-up questions before they become bottlenecks.
Fourth, it is helping standardise onboarding and compliance-heavy processes. These are areas where consistency matters and manual repetition is high.
Fifth, and this is the one I think matters most, it is helping reduce administrative drag between systems. Not because it writes better prose, but because it can help move work along.
That is the shift I find most interesting: from AI as a chat interface to AI as part of the operating fabric of the workflow.
The market will keep moving quickly. That part is guaranteed.
But the fundamentals are not changing. Practitioners still need to know where information came from. They still need to verify outputs. They still need to understand what systems a tool can and cannot access. And they still need to make sure any AI they use fits their obligations, not just their curiosity.
My advice is simple: do not think about AI as one buying decision.
Think about it as a stack.
You may use one layer for drafting and summarising. Another for legal-specific analysis. Another inside the workflow itself. And underneath all of it, you need trusted data, governed access and systems that can support real execution.
The firms that get the most value from AI will not be the ones chasing every new tool.
They will be the ones that take a more disciplined approach. The ones that understand where AI helps, where it does not, and where trust still needs to be designed in from the start.
Because in legal and conveyancing work, that is still what matters most. Not just what the technology can say, but whether it can support work in a way that is useful, reliable and accountable.
