Over the past few years, there has been much debate about whether increasing AI use to automate tasks will pressure law firms into dropping their billable-hour pricing.
Most law firms are still clinging to their long-standing pricing model, but software companies supplying law firms and company legal departments are starting to change how they charge.
Legora, one of the largest suppliers of legal AI, has announced “consumption based” pricing for its most advanced artificial intelligence product.
Others are expected to follow. By 2028, more than 35% of company legal departments' new spending on legaltech will be tied to their consumption of AI, research company Gartner has predicted. This is in contrast to the traditional software pricing model that charges companies a fixed fee based on how many people can use the software.
“Per-seat pricing made sense to get the market started, because it was easy to buy, the product was simpler and usage was more predictable,” Alex Fortescue-Webb, global head of legal engineering at Legora, told Global Legal Post. “It was never the right long-term model for AI, because usage varies too much between people for a flat fee to hold up.”
For customers, it means the price of a task reflects what that task actually needs, he adds.
“A simple job is priced like a simple job and a heavy one like a heavy one,” he said. “A team that puts 10-times more work through the agent will pay more for that work, and the team next door that doesn’t, won’t.”
Customer reaction to the pricing change has been positive, he says, in part because they get a better understanding of how much each instance of AI usage costs depending on what it is used for.
“It enables users to assess ROI at a much more granular level than ever before,” he said.
Global Legal Post spoke to law firm partners, heads of company legal departments and an analyst about what pay-per-use AI may mean for lawyers.
Many legal AI users are trying to figure out if consumption-based pricing – where, for example, suppliers may provide customers with a finite amount of AI ‘tokens’ – will make it harder for them to budget legal AI costs and get value for money from it.
As with any variable cost, consumption-based pricing will create challenges for predicting the cost of delivery in advance, whether for an in-house team or a law firm using the technology, says Ben Allgrove, partner and chief innovation officer at Baker McKenzie. “The days of throw the latest and newest AI at every problem are over,” he said.
Baker McKenzie is planning for consumption-based AI pricing to be a “significant, new marginal [cost]” in the delivery of AI-enabled legal services, he added.
Others, however, say that there is no consensus in legaltech about how suppliers will charge for AI in the near future.
Mark Brennan, global managing partner for digitalisation at law firm Hogan Lovells Cadwalader (HLC), says suppliers are exploring several alternatives to traditional per-user pricing, including combining a fixed subscription with usage allowances for AI, credits or additional charges linked to the volume or complexity of AI consumption.
“The market has not settled on a single approach [to pricing],” he said, in part because legal AI suppliers are “being incredibly competitive to win law firm business”.
For HLC and other law firms, when picking legaltech suppliers, law firms want “transparent pricing and predictable tool costs”, he said.
Company legal departments are also contemplating the implications of consumption-based AI, especially any impact on IT costs.
“The biggest headache is predictability,” said Patricija Corey, legal operations manager at investment company Franklin Templeton. “Legal teams are used to software being a fairly fixed annual expense. Token pricing turns AI into a variable cost, and the more successful you are at driving adoption, the more you may spend.”
Legal departments may struggle to keep track of who is using AI tools in their team and how much each tool is costing, she says.
“Asking a simple question, reviewing a 100-page contract and running an agent across thousands of documents are very different activities, but the end user may have no idea what each one costs,” she said. “That also makes ROI harder.”
The answer, according to Corey, is not necessarily telling lawyers, “don’t use too many [AI] tokens”, but instead building guardrails on AI use – for example, using dashboards to monitor its usage and agreeing a spending cap.
Yet what if consumption-based AI could help in-house legal departments get a clearer picture of their activities and the cost of external legal advisers?
Andrew Cooke, chief legal officer at Perk, an online platform for managing business travel, says that tracking law firm billing is not straightforward.
“There is no minute-by-minute visibility provided by law firms on their fees,” he said. “Generally, I am reliant on periodic updates to assess what might be on the clock on various matters. This makes sense for the law firms, just as consumption-based pricing makes sense for vendors of AI products.”
If legal AI suppliers incorporate cost-tracking features into their software, it may “ease the pain” of tracking legal costs by providing more detail on the “dollar cost” of a job or legal task, he added.
Still, for many law firms and in-house legal teams, any switch from per-user to consumption-based software pricing may be a challenging and expensive transition, says Gartner legaltech analyst Shannon Nakamoto.
She estimated that consumption-based pricing for legal AI may increase IT costs for law firms and company legal departments by as much as 25%, especially for information-intensive or reasoning-based tasks such as litigation and M&A-related due diligence.
“A lot of company legal departments [and] law firms have got used to using these [legal AI] tools or have experimented with them and use them regularly now,” she said.
Consumption-based pricing may create a “layer of additional cost that may be hard to predict or hard to budget for”, she said. “There’s a lot of concern around that and how to manage that.”
Email your news and story ideas to: [email protected]
