Four packaging models for 2025 and how to price them: generators, agents, co-pilots and service as software
Steven Forth is CEO of Ibbaka. Connect on LinkedIn
By the end of 2024 four typical packaging patterns had emerged for generative AI: Agents, Co-Pilots, Generators and Service as Software.
In December 2024 we conducted a simple LinkedIn Poll to ask which pattern people thought would predominate in 2025.
The poll ran from December 16 to 20 and was shared on the Design Thinking, Professional Pricing Society, Scenario Planning and Software as a Service Groups. There were a total of 118 responses.
The framing for this poll was as follows …
Generative AI is starting to change the standard patterns for B2B software. Four patterns are emerging.
Agents - that take actions and complete specific tasks
Co-Pilots - to stand beside people ready to assist
Generators - to create new content and ideas
Services as Software - as various services become scalable through software services that were provided by people can be delivered through AI
Which pattern do you think will emerge most strongly in 2025?
Co-Pilot was the most common response at 43%. This is in part because Microsoft and Google have made the use of AI co-pilots a common experience. Co-Pilots also bridge the gap between sustaining and disruptive innovation connecting the familiar with the new.
Agents and Service as Software ran about equal. The agent approach, or agentic AI, is being heavily promoted by Salesforce and attracted a lot of attention at the end of 2024. Service as Software is a more holistic approach that combines generators, agents and in some cases co-pilots to replace professional services provided by skilled humans with AIs. Accounting and legal services are often used as examples, but virtually all services are candidates for disruption.
Generative AI is based on using transformer and diffusion models to generate content. Generators are the foundation of the current AI boom and all other approaches rely on generators. They are also used directly for content creation and to support conversational AI. They are even replacing search in many use cases! This poll suggests that they may be fading into the background with other approaches to packaging coming to the forefront in 2025. At this point Ibbaka’s Customer Value Management platform relies on generators for some of its core functionality.
Each of these packaging patterns has its own approach to pricing.
Table from Ibbaka AI Monetization in 2025, to be published January 7, 2025. Sign up to get notified on publication.
These four patterns are not independent of each other. In fact all are based on Generators and both Agents and Co-Pilots can be integrated into Service as Software. The design space for B2B generative AI applications can be organized along two dimensions: time (episodic or ongoing) and autonomy (supports decisions or makes decision).
Figure from Ibbaka AI Monetization in 2025, to be published January 7, 2025. Sign up to get notified on publication.
Autonomy of Decision Making: Does the AI actively make decisions and act on the environment or does it sit beside a human who makes the decisions and takes actions?
Continuity: Is the AI only engaged under specific conditions or is it always on?
Where an application lives in this design space will have a big impact on how it creates value, how that value is captured into price and the operating cost. In other words, each niche in this space implies a different solution to the fundamental relationship that Value > Price > Cost.
