An AI Lab set up inside of law firm Lander & Rogers is working up “three or four” prototypes a day, testing how they interact with different data types and determining which are feasible use cases.
Jared Woodruff.
Head of AI Engineering Jared Woodruff told the iTnews Podcast that the lab, set up in February, is working through a large pipeline of ideas that have been raised in the months since launch.
“When this whole AI lab came into fruition, the number of ideas that came at us was absolutely phenomenal,” he said.
Where possible, ideas are being prototyped via a series of Microsoft Copilots.
The firm is using M365 Copilot “as the mainline” implementation vehicle for early use cases.
For ideas not conducive to being implemented through Copilot, the lab considers other AI options, including building specifically for the use case or buying in a new tool.
Other non-AI options are also taken into account: “We can run it through [Microsoft] Power Automate, or other existing workflow tools. We can even work with a team to redefine what the problem is and then go forward.”
Woodruff said that being one of the first law firms to get access to M365 Copilot, combined with access to Microsoft resources, had enabled the firm to accelerate the work of the AI Lab, despite it still being “early days”.
“Without their support, I doubt we would be where we are now with our ability to execute on that.”
Woodruff said the most valuable AI use cases were those that saved time in finding and surfacing information required by lawyers.
“That’s the key selling point for lawyers,” he said.
“You don’t need an AI to rewrite what they’re doing or anything like that. That’s not what AI is meant to do.
“The AI is there to give them all the information that they need to execute that decision and execute it with precision, saving them time.”
Through the Microsoft partnership, Lander & Rogers is evaluating more Copilots, including the GitHub Copilot as a software development aid, and the newer Copilot for Security, which aims to support incident analysis and response.
Woodruff said that more Microsoft-made Copilots are on the radar, but would not be experimented with until instances were made available on Australia-based servers.
In addition, Woodruff is excited at the promise of emerging AI agents that have computer vision capabilities.
“The vision aspect [of AI] excites me the most by far and away, because the amount of functionality and applications of vision technology is endless,” he said.
“We’ve already shown internal prototypes around the capabilities of vision and it’s doing stuff that we’d never thought was possible, but at the moment, it’s cost prohibitive to do at scale.
“Once we see advancements throughout this year, once vision becomes commoditised, that will really dominate the space and deliver real effective change for businesses.
“On the agent side I see that as closing gaps for workflow execution, complex mission planning and logical reasoning.”
One possible use case that’s being prototyped is to use AI agents with vision as a way of adding AI into legacy legal-specific applications, so that lawyers can interact with them in the same way as they would with Copilot.
“We have legacy applications in the firm that all law firms use. We now have an application that wraps itself around these legacy applications and uses vision technology to actually ‘see’ into the [legacy] application, and then gives a Copilot-like conversational experience on the right [of the screen],” Woodruff said.
“If it works, it means we don’t need to ask [legacy] vendors to add AI support into their apps. We’ll just put it into their apps ourselves.”