Monetizing and Licensing AI Solutions built on LLMs

You may have optimized your LLM infrastructure and cost model.

But how do you monetize your own AI solution
with your customers?

Book a Demo

Learn more

You've Built the AI Product

You’re leveraging the best LLM in the world to underpin your product.

You’ve optimised for cost as best you can:

Model choice

You’ve chosen your preferred model and, based on its complexity and the number of parameters, you’ve got a good idea of how much it’s going to cost you to run.

Performance

You’ve chosen batch pricing or provisioned capacity to keep costs in check and decided on balance between speed, efficiency & cost that should help you provide the best service level to your customers at a reasonable cost for you internally.

Workloads

You understand that if you’re working on just text or working with multi-modals, your cost base can differ dramatically but you’ve got this under control and know what you’re doing.

But then comes the really hard question:

How do you best monetize your own AI solution with your own customers?

Different Ways to Monetize AI Solutions

For your customers, AI tooling is complex and new services are appearing pretty much every day. The pricing of them is equally complex and hard for your end customer to be able to understand and thereby commit budget to.

What are your options?

Token-based Billing

The simplest option is to add margin and pass on token-based billing to your customer.

However, token-based billing is one step removed from your customer. While you as the AI solution provider may understand it, the end customer may struggle to interpret it and budget for it.

Token-based billing may be an easy option for you, but it might not be the best model to present to your customers.

Capacity-based Licensing

The use of an LLM is restricted to a defined allocation of compute resources such as CPU consumption, I/O or memory usage.

This is similar to the AWS model of purchasing provisioned capacity. In the LLM world, however, that capacity is typically hidden from end users, and instead a certain number of tasks per user are permitted based on the account type.

The user simply understands that they can execute a defined number of tasks under the terms of their license, while you as the AI solution provider define the allocated resources and control how long tasks can run and at what cost.

Task or Success-based Licensing

In this emerging model, the end user completes a task and then accepts a result. The user pays only for the accepted outcome.

For example, a user might ask an agent to create three derivative images from a reference image. If the user accepts any of the generated images, that is considered a successful task and the user is charged.

Similarly, in agent-based solutions for customer support, a license could allow a defined number of tasks to be completed by the agent each month for a fixed cost.

Implementing AI Licensing and Usage Control

As an AI solution vendor, you’ve got a lot on your plate – designing your solution, choosing your foundation LLM, keeping your cost base under control and taking that solution to market.

You’ll also need to design and implement the licensing logic that underpins it.

That’s where 10Duke can help. As software licensing experts, you can save time by using our licensing engine rather than building your own licensing logic.

What can 10Duke do for you?

Usage Metering

Your client application generates measurement data about the capacity used when completing tasks. 10Duke records this metering information against the limits defined by the license.

License Enforcement

You define the usage limits granted to each customer. 10Duke monitors usage against those limits and helps enforce them.

Billing Integration

Metered usage data can be passed to billing engines such as Stripe to automate invoicing and payment.

End-to-End Monetization

Licensing, fulfilment, compliance monitoring, invoicing and payment working together as part of a single system.

Ready to license and monetize your AI solutions?