No amount of AI algorithmic sophistication
will overcome a lack of data [architecture].
AI success starts with a simple principle: There is no AI
without an IA (information architecture). That’s why we’ve put together a prescriptive
approach composed of 5 core imperatives we call the ladder to AI. It is designed to help
our enterprise clients with a thoughtful and well-architected approach.
Step 1: Modernize all your data estates in a multicloud environment
Given the dynamic nature of AI, your data estate needs a
highly elastic and extensible multi-cloud infrastructure to unify capabilities within a
fully governed team-platform. Clients are also looking to automate their AI lifecycles
across an array of contributors through collaborative workflows. To modernize your data
means building an information architecture for AI that provides choice and flexibility
across your enterprise. As clients modernize their data estates for an AI and multicloud
world, they will find that there is less "assembly required" in expanding the impact of
AI across the organization.
Step 2: Collect data to make it simple and accessible
Enterprises need to establish a strong foundation of data,
making it simple and accessible, regardless where that data resides. Since data used in
AI is often very dynamic and fluid with ever-expanding sources, virtualizing how data is
collected is critical for clients.
Step 3: Organize data to create a business – ready analytics foundation
Just because you can access your data doesn’t mean that it’s
prepared for AI use cases. The organize rung is about creating a business-ready
analytics foundation designed to ensure your data is ready for AI. Bad data is
paralyzing to AI, so clients must integrate, cleanse, catalog, and govern the full
lifecycle of their AI data.
Step 4: Analyze - scale AI everywhere with trust and transparency
Once your data is accessible and AI-ready, then you are
better prepared to apply advanced analytics and AI models. This rung provides the
business and planning analytics capabilities that are key for success with AI. It also
provides the capabilities needed to build, deploy and manage AI models within an
integrated portfolio of technology.
Step 5: Infuse - operationalize AI throughout your business
Many businesses create highly useful AI models but then
encounter challenges in operationalizing them to attain broader business value. This
rung of the ladder infuses AI to build trust and transparency in model-recommended
decisions, decision explainability, bias detection and decision audits. For clients with
common use cases, the infuse rung operationalizes those AI use cases with pre-built
application services, speeding time to value.
Daniel G. Hernandez
VP, Offering Management, Analytics, IBM