View all AI news articles

Navigating the Tricky Waters of Pricing Models in Generative AI: A Deep Dive

February 27, 2024

The New Frontier: Box's Innovative Approach

In the world of generative AI, pricing models are as varied as the AI features themselves. Take Box, for instance. In October, they rolled out a new, rather crafty pricing scheme for their generative AI features. It's not your grandpa's flat rate; it's a consumption-based model that's as unique as a snowflake in San Francisco. Each user starts with a treasure trove of 20 credits monthly, perfect for a mix of AI tasks totaling up to 20 events. It's like a game – each task costs a single credit. Run out? No worries! Dive into a communal pool of 2,000 extra credits. But if you're a credit-hogging machine, you might need to chat with a salesperson to buy more credits.

Box CEO Aaron Levie, the captain of this ship, states that this approach is all about charging based on usage, acknowledging that some users are more AI-hungry than others. Plus, it covers the costs of using OpenAI's API, which is like the engine room for their large language model.

Microsoft's Old-School Pricing

Meanwhile, Microsoft, playing it safe and traditional, decided to charge a straightforward $30 per user per month for its Copilot features. This is over and above the cost of a regular Office 365 subscription, which varies. It's like paying extra for the cherry on top of your Office 365 sundae.

The Panel Insights: Generative AI in SaaS

At the Web Summit panel on generative AI's impact on SaaS companies, Christine Spang, CTO at Nylas, and Manny Medina, CEO at Outreach, shared their pearls of wisdom. Spang, cutting through the hype, acknowledged generative AI as a significant leap forward. She emphasized the importance of integrating this tech into products to unlock real value in various use cases.

Medina discussed the delicate balance between offering in-demand features and pricing them to deliver value to customers while also making a profit. He pointed out the expense of frequently consulting with large language model providers, which can quickly add up.

The Price-Cutting Trend

Interestingly, model makers like OpenAI are cutting prices as they find ways to run models more efficiently. For instance, they announced in June new features that increase processing power, providing more value while lowering costs for older versions for those who don't need the latest and greatest.

Different Strokes for Different Folks

Spang's company, Nylas, already employs a consumption model based on the number of connected email or calendar applications. She plans to adopt a similar approach for generative AI features, aiming for a price point that fits the median user.

However, Medina believes that consumption models are trickier for applications than for API providers like Nylas. The verdict is still out on whether companies will be willing to pay a flat rate like Microsoft's $30 per month per user for Office 365, unless they see tangible value from the additional cost.

The Compliance Conundrum

A significant unknown is the compliance costs related to using generative AI technology. If governments require companies to disclose their AI "ingredients," and providers like OpenAI don't offer this information, it could get complicated.

The CIO Perspective

CIOs, the gatekeepers of company technology budgets, are scrutinizing this technology. Sharon Mandell, CIO at Juniper Networks, is looking at the ROI of these features. In 2024, they are testing the waters with GenAI tools to see if the benefits justify the costs. CIOs are running pilots, moving cautiously, and trying to gauge whether there's a real increase in productivity.

In summary, companies are experimenting with various pricing models, while their customers conduct pilots and proofs of concept. It's a dance of discovery, with both parties trying to figure out the real benefits and the best way forward.


  • Box introduces a unique consumption-based pricing model for generative AI features.
  • Microsoft opts for a traditional flat rate for its Copilot features.
  • Industry experts discuss the challenges of integrating generative AI into products and pricing them effectively.
  • OpenAI cuts prices on older products as newer, more efficient models are introduced.
  • Compliance and ROI are key considerations for companies and CIOs when adopting generative AI technologies.

Recent articles

View all articles