Box introduced a fresh pricing approach for its theoretical AI capabilities in October, opting for a consumption-based model over a fixed rate structure.
Under this new plan, each user is allocated 20 credits monthly, which can be utilized for any combination of AI tasks totaling up to 20 events, each incurring a single credit charge. Users also have the option to delve into a supplementary pool of 2,000 credits. Should the usage surpass this limit, engaging with a sales representative to secure additional credits is recommended.
Apart from covering expenses related to leveraging the OpenAI API, which serves as the primary foundation for its expansive vocabulary model, Box’s CEO Aaron Levie elucidated that this strategy enables billing based on actual usage, acknowledging that certain users may leverage the AI functionalities more extensively than others.
In contrast, Microsoft has adhered to a more traditional pricing model, as revealed in November. The company disclosed that access to its Copilot features would entail a $30 fee per user per month, in addition to the standard monthly Office 365 subscription fee, which varies depending on the customer.
Discussions at the Web Summit in November, featuring Christine Spang, the co-founder and CTO of Nylas, a communications API provider, and Manny Medina, the CEO of the sales acceleration platform Outreach, shed light on the obstacles encountered by SaaS enterprises when implementing such features. The trend observed over the past year indicates that enterprise software firms are increasingly integrating theoretical AI functionalities.
Spang emphasized the substantial progress represented by theoretical AI, despite the surrounding buzz, stressing the necessity for software companies to embed it within their offerings. She remarked, “I’m not going to say that it’s like 10 out of 10 where the hype meets the [current] reality, but I do believe there is real value there. What will really matter is how people use the technology, connect it to other systems, other apps, and sort of drive true value in different use cases with it.”
