My original notes: the seed that eventually became this article.
The year is 2026 and for the last 3 years, or maybe even 4, AI has been making these waves like you've never seen before. While Business/IT leaders realize that the advent of AI brings in great new capabilities and opportunities to fore, what still remains a significant challenge is: How do you go about adopting it?
While the question itself may seem very simple for a seasoned leader - however, the truth is - in the enterprise context there are so many nuances that knowing where and how to start is actually proving difficult. Ideas from various LoBs are flowing in thick and fast - every team, every unit has compelling reasons to explore AI capabilities/hire experts and have tons of slide work ready that shows clear promise of how a whole lot of business KPIs will improve. Few have measurable data points that could explain the real improvement vs the AI spending you would to achieve it. And let me not get started on AI spending, right now - because that will send me down a rabbit hole even I am not ready for!
Sometimes, its easier to understanding a concept in relation to something else we already know well - things simply translate to what matches, what does not match and whats really just something entirely. You can then focus on the new bits or whats relevant for you first.
So, before we start thinking and planning about how to adopt AI, maybe it would help revisit another recent (at least for most of the leaders) major change that evolved the way our businesses were conducted which is - the advent of the Cloud.
Not so surprisingly - all major cloud vendors have very well thougth out Adoption frameworks, some capability-led while others are journey-led. I find capability-led as more apt for what we are doing at this stage, but before soon, we will also review the journey-led approach (another contrast, if I may).
Capabilities for AI:
Think back to when cloud adoption started en-masse, everyone saw the need and potential, but no CAF existed, and there was chaos, shadow IT, cost overruns etc. by the silver lining was, it was always possible for you to know the numbers. A private data-center move to cloud would save x amount over n years, retraining workforce would cost y, and hiring new talent would cost z, all opex now and driven by more accurate, more recent quarterly data.
CAF evolved to list key capabilities that orgs needed to building around - CCoE, Workforce training programs, EA, Governance, Risk & Compliance, Security, Observability etc. - applicable to both public and private cloud. If you think hard its not so different with AI - capabilites are still the same, but agenda is different - and it applies to Public AI as well as Soverign AI(more on this later).