Thoughts on AI pricing

Pricing AI-powered features can be approached in different ways. We could integrate it into our plans, introduce a fixed-price add-on, or use a pay-as-you-go model. We've decided to go with the latter.

Why? There are a few unique challenges when it comes to pricing AI features:

• AI providers and models are rapidly evolving.
• Inference cost varies based on multiple factors.
• A fixed pricing model would limit how we use AI.

Several companies offer different AI models: OpenAI, Anthropic, and Google, to name a few. Each model has its strengths and weaknesses, and the cost of inference can vary depending on the model and change over time. This variability makes it difficult to predict how much we'd need to charge for AI features.

With a fixed AI budget, we'd be limited in the features we could develop. We want to experiment with different AI models and providers to find the best solution for our users. We also need to quickly adapt to changes in the AI landscape without being constrained by a fixed pricing model.

To address these challenges, we've introduced AI credits as a separate pricing component. AI credits are a pay-as-you-go option for our AI-powered features. You can purchase credits in advance and use them to access various AI capabilities. This way, you control how much you spend on AI features and which ones you enable or disable.

It's also worth noting that AI inference costs are expected to decrease over time as technology improves. With the pay-as-you-go model, we can easily pass these savings on to you.

AI is not our business model; it's a tool we use to provide better features to our users. By having a flexible pricing model, we can continue to develop AI-powered features without being limited by a fixed budget per organisation. We believe this approach is fair, transparent, and flexible, and we hope you do too.