How we’re thinking about pricing for enterprise AI
With a new category like enterprise AI, how do you get to the right pricing? Here are our learnings so far.
Glenn Jaume
Product Manager at Coda
How we’re thinking about pricing for enterprise AI
By Glenn Jaume
Share this article
AI · 6 min read
Consumer AI ≠ enterprise AI.
When thinking about pricing for our AI products, we first looked to consumer AIs, as they were the primary AI tools on the market at the time. Typically, consumer AIs are either free, like Meta AI, or work on a freemium model with paid plans for higher limits or additional features, like Claude from Anthropic or ChatGPT from OpenAI. But the reality is that, although they are both built on LLM (large language model) technology, consumer AIs and enterprise AIs are very different. The biggest difference is that consumer AIs don’t have access to your company data, meaning they lack context for the tasks you’re asking them to do. Conversely, enterprise AIs are connected directly to the tools you use, so they can provide much more specific, personalized responses—and, therefore, deliver more value. Enterprise AIs like Coda Brain are also built into the tools you’re already using for your work. That means they go beyond answering questions and enable you to actually take action, like providing a table of your highest-value deals that you can insert straight into a sales weekly meeting dashboard. Despite this, we chose to include Coda AI—our AI assistant built directly into your Coda docs—for free for doc makers. That’s because we knew Coda AI would be incredibly useful and wanted anyone making docs to benefit from that power as part of their core experience with Coda. We plan to do similar for our enterprise customers with Coda Brain, our turnkey AI platform specifically designed to meet the data volume challenges typically felt by larger businesses.
Flexibility and transparency are must-haves.
One thing we hear repeatedly from customers is that they don’t want to be surprised by higher bills than they were expecting. That’s true for any product but is especially true of something new like AI—businesses are often still “testing the waters” and want proof of value before investing more. One way to solve for this is “pay-for-what-you-use” pricing, where your bill is directly aligned to how much you use the product. But the challenge with pure usage-based pricing is that customers often struggle to know what their consumption will be—especially if it’s a type of product they haven’t tried before—which makes predicting spend extremely difficult. Some ways to solve for this are allowing customers to set spend limits or offering tiers that customers can move freely between. Providing a free trial period so they can get a sense of their usage before committing can also be really helpful. With Coda AI, we’ve chosen a model that combines usage (in the form of AI credits, which roughly equate to the amount of content in your AI prompts and responses) with a seat-based approach. That means that doc makers get Coda AI for free, with a generous allowance of credits that can be pooled and used across your team. Admins can see how many credits have been used, and by whom. They can also set limits and buy more credits if needed, giving them complete control and flexibility over how much they spend.