Beyond SaaS: why the next wave of AI value is co-created, not licensed
The SaaS model was a remarkable simplification. Instead of buying software, configuring it, and maintaining it yourself, you paid a monthly fee and got access to something that worked. Predictable costs, automatic updates, no infrastructure headaches. For a generation of business software, it was the right model.
But SaaS has a structural limitation that becomes visible the moment AI enters the picture. You are licensing access to a fixed capability. The product evolves on the vendor's roadmap, not yours. And when the value you need is not a feature but a process — a specific combination of capabilities applied to your context, your customers, your data — a licence does not get you there.
The next wave of AI value cannot be packaged and sold. It has to be built together.
The shift from features to outcomes
Traditional SaaS sells features. You pay for the enrichment engine, the notification layer, the analytics dashboard. The features exist. You activate them. What you do with them is your problem.
The emerging model inverts this. The starting point is not "here are the features" but "what outcomes are you trying to create?" From there, the relevant modules are selected, combined, and configured to serve that specific objective. The platform provides the building blocks. The service provides the architecture.
This is not consultancy in the traditional sense — a project with a defined end date and a final deliverable. It is ongoing co-creation: a continuous collaboration between the platform team and the client organisation to extract value, evolve the approach, and respond to what the data reveals over time.
Composable modules, designed for your context
The technical foundation of this model is modularity. A platform built from composable pieces — enrichment engine, push and RCS notification layer, mindset insights collector, behavioural context generator for AI agents — can be assembled differently for different clients and different objectives.
A wealth management firm might combine enrichment with a mindset collector focused on financial anxiety signals and a gamification layer designed to surface genuine risk tolerance over time. A utility company might prioritise energy consumption pattern detection with contextual prompts triggered at billing moments. A streaming platform might focus on engagement signals and regret indicators to reduce passive churn.
The modules are the same. The combination, the configuration, and the process of extracting value from them is unique to each client.
What long-term co-creation looks like in practice
This is where the model most resembles what Palantir built at enterprise scale — but applied to organisations of any size that are serious about AI as a competitive layer rather than a feature addition.
In practice, co-creation means several things running in parallel. Designing the insight extraction processes: which behavioural signals to prioritise, how to measure the quality of enrichment, how to track NPS not just for the current product but for the future AI agents that will act on this data. Suggesting and iterating gamification mechanics that generate genuine 0.5 party data rather than superficial engagement. Designing targeted campaigns to capture specific mindset data when the client needs to understand something concrete — how customers feel about a new product, a pricing change, a service disruption.
None of this is a one-time configuration. It evolves as the client's understanding deepens, as the data compounds, and as the agentic AI layer moves from pilot to production.
The compounding advantage
The long-term nature of the model creates a compounding dynamic that point-in-time consulting cannot replicate. Every month of co-creation produces richer data, better-calibrated processes, and a deeper shared understanding of what works for that specific client and their specific customers.
By the time agentic AI is fully deployed, the organisations that have been building this foundation through continuous co-creation will have a structural advantage that cannot be acquired quickly. The mindset data, the behavioural context, the tuned enrichment processes — these are not things a competitor can replicate by purchasing a licence.
SaaS democratised access to software. Co-created AI platforms democratise access to something more valuable: the ongoing intelligence work that makes AI genuinely useful, at scale, for the long term.
The licence gets you the tool. The co-creation builds the advantage.