SaaS and AI - from trains to travel passes

When I first started in SaaS I used the analogy of a train when picking your vendor.

Because SaaS was about the regular releases and a constantly improving platform, when making your decision it was less important about where the vendor was now, and more about where they were headed.

It was important to understand the vendor's vision, buy into their roadmap, and then jump on board and enjoy the ride.

Many companies benefited from being on the Salesforce train as they rolled out the AppExchange, Chatter, Service and so on.

But this analogy breaks down in the world of AI models - because you won't just pick one - you'll use different models for different use cases, and in some scenarios multiple models for a single use case.

So now I think of a travel pass that allows you to jump from a tram, to a train, to a bus.

As leaders consider how you can set up a composable architecture that safely provides access to as many models, data sources and workflows as possible to allow your business teams to experiment, learn and deploy.