After decades of labelling delays in risk aversion technology, the legal department of the company has emerged as an early adopter of artificial intelligence.
According to experts, legal services are often the fertile foundation of automation powered by AI for tasks such as drafting contracts and providing initial legal opinions.
For internal legal teams looking at legal tools for AI, the first and most important question is whether to build the tools internally or buy them from a software supplier. What are the advantages and disadvantages of each approach?
builder
Given the complexity and cost of designing, testing and deploying proprietary software, many companies building their own legal AI tools are in the technology sector.
For example, Hewlett Packard Enterprise (HPE) encourages legal teams to act as “citizen developers” by building the AI applications themselves, rather than relying solely on third-party products.
The initiative began in 2023 when HPE held a company-wide “hackathon” to propose ideas for technology projects. The Legal Department has developed a team that wins with a “proof of concept” that helps to monitor regulatory compliance using generated AI and leverages technology for a variety of tasks, including contract analysis.
HPE tried out AI software from a third-party software supplier that it uses for its legal department, and was “overwhelmed” by its performance, says Jeff Fugere, a related advisor at HPE, specializing in cybersecurity and IT.
There were also concerns about sharing confidential information such as contracts and negotiation strategies with software suppliers. “All we decided at the time was OK.
He estimates that the tool is currently being used by 50-100 people on HPE legal teams around the world.
Building the tool internally has several benefits, Fougere said, including more control over the data to HPE, minimizing the risk of leaking sensitive information.
Recommended
According to Fougere, AI was able to fix and fix issues, including generating incorrect data and “hagazes.” Building AI software, he says, has proven to be “significantly” cheaper than buying third-party tools.
Some legal departments start by building relatively basic tools and wait for them to prove their value before moving on to develop something more complex.
“AI initially deals with low complexity and massive amounts,” says Andy Cook, chief legal officer of Travelperk, an online business travel management platform.
The Barcelona-based company, which raised $200 million in January to fund the expansion, began building AI tools in 2023 to answer simple legal queries. Now, new features of the company’s platform are violating industry regulations and are being used for tasks such as monitoring whether legal information can be extracted from company emails.
To build the software, Cooke said the team starts by recreating the functionality of commercial third-party tools and connecting them together “like Lego Blocks.”
The first legal AI tool is relatively simple, with the goal being to build a “useful product that meets your business needs.” More complex tools will be built with future iterations of technology, he adds.
Buyer
Legal departments of other companies, especially those with limited and fewer IT resources, are difficult to justify building custom tools in rapidly changing technologies with a growing third-party market.
Cattaring Supply Chain Staples Canada uses software from legal AI startup brightness to manage and negotiate hundreds of thousands of contracts, says Adrian Lang, Chief Legal and Privacy Officer.
The software uses a traffic light system to identify contract terms that require closer review. Red signaling is of high concern, yellow for negotiation, green for clauses ready to be approved. She estimates that automating part of the review process saved her from hours to more than a day per contract.
Even if it was possible to justify the resources needed to build such tools internally, it could prove difficult to keep up with legal software supplier standards, says Lang, who leads a legal team of six.
“(Third-party) technology will quickly overtake (in-house) and will soon become obsolete unless you promise to actually continue investing in that build,” Lang says. “Or at least… It’s not as cutting edge as other new products on the market.”
Spanish energy group Repsol initially built its own legal AI software for drafting and negotiating contracts. The software was originally designed only for one type of contract (for exploration and production) in three countries. According to Pablo Blanco Pérez, legal advisor at Repsol, the in-house software was “working very well.” However, the company switched to Legal AI startup Harvey after testing the software and discovering it was more versatile and cost-effective.
Harvey was “a lot better” at handling different types of contracts, he says. To reach the level of versatility Harvey had, Slanco knew they had to invest more money and resources. “And I think Harvey will invest a lot of money in improving the tool,” he says.
Recommended
Despite growing interest in legal AI software, many internal legal teams lack staff, and sometimes even basic legal IT such as traditional document management systems, says Weston Wicks, a legal and compliance technology expert at research firm Gartner. It can make it difficult to promise to build your own AI tools, he says.
Legal departments with limited budgets can use “low code” and “no code” development platforms to reduce the cost of building their own tools, he says. These allow users to build applications with intuitive drag-and-drop tools that have little or no knowledge of coding.
However, internal teams are still in the minority who are willing to spend a significant amount of money on building sophisticated tools, Wix said:
