How do you price AI software?

How you price AI software has a direct effect on whether your customers adopt it and keep using it. We have been thinking about this for our own model at 1KE, and a story from another company made the point clearly.

What happened at Lightfield

Lightfield is an AI based CRM. It plays in a traditional space with an AI product, and it started where most software starts: a fixed price per head licensing model.

That ran into trouble quickly. People were using the tooling more than the company could afford at that price, which is obviously not a sustainable business model. We are all fairly used to token based or credit based consumption at this point, so the company did the logical thing. It shifted entirely to consumption. No more per user licensing. Everything consumption based.

The impact was that they possibly went too far the other way. Customer usage dropped away significantly when they made the change.

Why usage dropped

The reason was subtle. In a CRM, a lot of what customers do is business as usual work: normal things like adding a record. They did not see that work in a consumption frame, and they could not understand how adding a record was something they were consuming. So there was a fear of suddenly using a lot of credits to do normal work.

The lesson is that there are different types of work in AI software. Some functionality lends itself to running AI models. Some of it is simple automation. You have to be clear with users about which is which, and what each one means for what they pay.

Predictability first

This is not a new problem. We have seen it with cloud consumption over the last ten years. Enterprises have learned that if they do not manage and govern their consumption they end up with big bills, and some of that caution comes with them into the AI era.

Fundamentally, core functions in software need to have predictable costs. Customers will understand that things that are more variable in nature, or where they are getting a lot of value from AI, carry some variability, and they will accept it.

We are seeing more of this, and we have been working through it for our own product. Things like meetings are fairly standard and not particularly expensive, so we can cover those. Generating a really complex proposal takes a lot more effort, and that is where customers see most of the value, so they are comfortable with that being priced differently.

Three things to get right

It is an evolving commercial and operational challenge, and I think there are three elements to it.

First, help customers get comfortable during the initial engagement, so they feel safe buying your product. They need some feeling of control and certainty at the outset.

Second, give them visibility while they are using it, so they can see what is actually happening. AI coding tools are getting good at showing you your incurred cost as you go.

Third, differentiate the high value activities from the commodity activities, so customers can understand where the value sits and feel comfortable paying for it.

Where channel partners fit

Channel partners usually sit in an advisory role, between the customer and the vendor, helping the customer understand the choices they should be making. Quite often they are also responsible for driving adoption, for the customer or for the vendor.

That gives them an opportunity: to help end users understand what the consumption models look like, make choices that are commercially sensible for them, and perhaps put some governance across the top. It will be up to the channel to be educated and understand the implications well, and for anyone who does, it is a service line in its own right.

Watch the full conversation with Sidney on YouTube: https://youtu.be/d0M0R3Bvwt0

Subscribe to The AI Take for one AI lesson a week, and what it changes for your business.

About 1KE:

1KE is the presales autopilot for technical sales teams. It turns discovery calls, emails, and technical documents into structured requirements and customer-ready solutions in minutes, so teams can respond while the deal is still warm.

Learn more → https://1ke.ai

Michael → https://linkedin.com/in/michaelchanter/ · https://x.com/1keceo

Sidney → https://linkedin.com/in/sidneyminassian/

Follow 1KE → https://linkedin.com/company/1ke-ai/

‍