LAS VEGAS — IBM and its partners have done a good job of building the foundation for the next generation of technology advances in business, IBM CEO Gini Rometty told partners kicking off the company’s annual PartnerWorld Leadership Conference here. But the work the company and its partners today have done in fields like analytics, cognitive computing, and cloud to date is just getting things warmed up for a broader battle to come.

“We’ve done a superb job putting the foundation in place,” Rometty told partners in a keynote presentation full of examples of how partners are using the company’s Watson cognitive computing engine to drive their own innovations. “Now we’ve got to win the architectural battle, to set the table for the future.”

Rometty said the real high stakes in the near future are in a set of architectural decisions she believes customers are going to make over the next few years, decision that will be made both explicitly and implicitly. Those architectural decisions, she said, will go a long way to determining the environment into which partners sell, and the capabilities customers will have or be able to build out going forward. She presented to partners Big Blue’s “three platform” vision for these architectures.

Up first is the data platform, in IBM’s parlance the Watson Data Platform, which attempts to create a framework for ingesting, processing and storing data in all formats. While it’s much more of a “plumbing” type of area, it’s nevertheless key for partners to understand. “It doesn’t matter what type of partner you are,” Rometty told the audience, which includes VARs, MSPs, ISVs, integrators, consultants and more, “If you embrace this, you become the cognitive partner for the customer.”

The second and perhaps most important, to IBM’s way of thinking, platform is the cognitive platform, the layer which will, at its most basic level, ingest the data from the data platform, and use it to first learn, and then make decisions. She stressed that Watson has come a long way since being the PartnerWorld darling the year it took out a Jeopardy! champion.

“Chapter one of AI is over. We’re about serious business, and it’s not just consumer stuff,” Rometty said. “We’re talking about systems that are trained in an industry domain, that understand things like risk and compliance.”

The Big Blue chief told partners that all of their customers are interested in the capabilities of AI to transform their business but “they either don’t know how to start, or they’ve tried everything and ended up with all of these little islands of AI, all disconnected.” Taking a platform approach will both make it easier for customers to get started, and to get those disparate “islands” interconnected, she said.

The cognitive platform will cover a wide variety of functions and capabilities, including machine learning, vision, and learning and understanding professional domains. And around all of that, she stressed, they must be transparent, and take in mind the customers’ business model and need to retain complete sovereignty of private enterprise data, while still taking advantage of third party or public data sources.

“You need to pick a platform that’s transparent about who taught it, and using what data,” Rometty said. “The platform has to be architected so the insights stay within a company.

She concluded her pitch with her assessment that while almost all customers across all industries are looking at digitizing their business, “digital is not the destination.”

“You’re differentiated by cognitive,” she said.”Five years from now, it will impact every decision you make in some way.”

The third platform for Big Blue is the cloud platform, largely powered by its BlueMix technology, and is “how we deliver Watson and so much more.” Here, as in cognitive, she described a split amongst customers — many, she said, are unsure of which say to go, and many others, “have too many clouds right now,” again speaking to a degree of disparate platforms and islands of information.

“There’s no doubt in my mind that we’re the best cloud built for business, and that we’re at the core of what [your customers will] care about for the next decade,” Rometty said.

The wise reader will notice a common theme through those three platforms — Watson. The platforms either are themselves the essence of Watson (cognitive platform,) feed Watson (data platform,) or are the delivery method for Watson (cloud platform.)

While there are certainly a number of IBM partners doing interesting things with Watson technology, Big Blue’s challenge at this year’s PartnerWorld appears to be making it real for the rest of the “rank and file” of the channel. To that end, the company will Thursday give its partners the opportunity to get involved with Watson more directly, with a “Watson Build” event, wherein partners will get a chance to go hands on. The idea comes from IBM’s own internal Cognitive Build last year, which saw 75 per cent of the company’s employees take part in various projects that built new products, re-created internal business processes, and more. Many of the things developed in that Cognitive Build process, Rometty said, are now being developed either for internal use at Big Blue, or for products the company will eventually bring to market.

IBM seems to understand that it has to make Watson’s capabilities and impact more widely understood, especially since the technology by its very nature tends to “stand behind” the systems that customers or end users actually see or interact with. Still, she said, even if it is to many a silent partner, IBM is “on a path to touch one billion people with Watson by the end of the year.”