Legora's CEO predicts software firms will move beyond seat-based pricing models

For years, software companies have relied on a straightforward system of charging customers by the number of user licenses, or “seats,” purchased. But as artificial intelligence makes significant inroads into workplace tools, the economics underlying this model are starting to fray. Legora, a legal tech startup founded by Max Junestrand, is among the companies leading a shift away from subscription pricing in the wake of AI’s rapid advances and unpredictable costs.

Junestrand launched Legora at the tail end of the software subscription boom, selling legal industry software with the familiar approach of charging by seat. That model, which worked well during the rise of cloud software, charged customers the same whether their employees used the product constantly or rarely. However, as AI capabilities have been integrated, the old rules have begun to falter.

The turning point came in June with the rollout of Legora Agent, a new AI-powered system capable of automating tasks traditionally handled by junior lawyers from beginning to end. This innovation introduced a new complexity: every query submitted to the AI expends computing resources, and some tasks require much more computing power than others. Companies like Legora found that while two customers might each have one license, their usage - and therefore the cost to Legora - could vary dramatically.

According to a company spokesperson, Legora saw daily usage rates increase by 13% among users of the AI Agent feature compared to those not using it. Junestrand pointed out that it no longer made sense for a light user to pay the same as someone extracting maximum value from the software. In response, Legora overhauled its pricing approach. Existing customers could retain their current contracts for the core software but were offered the chance to access the pro-tier AI agents on a pay-per-use model. New customers shifted to a fully consumption-based, pay-as-you-go structure starting in June.

The response from law firms has been “good” but “mixed,” Junestrand reported. Some customers appreciated being able to correlate software expenses with specific legal matters. For example, an in-house lawyer could now see the cost of using Legora to review a set of documents and directly compare it with the cost of outsourcing the work to external counsel. To support this transition, Legora introduced a dashboard for usage tracking and a spending forecast calculator.

Other clients, particularly those still exploring the role of AI in their operations, were more hesitant. These firms are continuing to test Legora’s standard AI agents under their existing seat-based agreements before committing to the new pricing model.

The stakes are significant for startups using third-party AI models such as those provided by OpenAI and Anthropic, which charge developers based on usage from the outset. Heavy customer use can mean rising costs for startups without corresponding revenue under a fixed seat model, making usage-based pricing as much a necessity as an innovation.

Legora is not alone in this transition. Other startups, such as Cursor and Lovable in the coding software space, have also adopted usage-based pricing, charging heavy users more. However, seat-based pricing has not disappeared completely. Harvey, a major competitor to Legora, intends to retain the traditional seat model to offer clients “transparency and predictability,” though it is also considering other options for customers seeking additional value.

As AI continues to reshape software business models, firms are weighing the tradeoff between paying for unused capacity and facing uncertain costs when usage surges. The eventual dominance of either pay-as-you-go or seat-based pricing will likely hinge on which form of uncertainty customers - especially in the cautious legal industry - find more palatable.