The technology operating model lifecycle is a loop The first couple of years in my current role I helped early adopters move from a Center of Excellence to hub-and-spoke operating models to scale AI. Data quality became a blocker so we did the same migration for data products. Now I work with late adopters who are just starting a Centers of Excellence and I realized that operating model migration is really a loop. (See diagram below.) 1. Business units and business functions experiment independently to validate the potential to generate value. 2. A Center of Excellence is created to scale quickly. 3. The Center hits diminishing returns on scale, their backlog grows, so they decentralize development to continue scaling. McKinsey, et al. call this hub-and-spoke and Tom Davenport’s International Institute for Analytics calls it draining the Center of Excellence. 4. The Center is still a bottleneck so decentralize governance (i.e., data mesh).? Move all development and most governance to the edge of the business. The Center still manages a shared platform to gain tech (not people or process) economies of scale and facilitate interdisciplinary collaboration. This may seem like a small incremental change, but to many in IT it’s the biggest cultural change since personal computers a generation ago and has made its inventor, Zhamak Dehghani, world famous. 5. The technology (AI or data products) matures and central governance is no longer needed; e.g., most companies don’t have a central Excel governance team. Some of our pharma customers are at this stage with AI/ML. 6. The technology becomes obsolete, is replaced by a new one, and we start the cycle all over. For more see https://lnkd.in/dpqBwEZp and https://lnkd.in/d6Y4eimd #ai #dataproducts #datamesh
Doug I love how you bring clarity where most people see chaos. Every business leader looking at how to generate growth should read this article to understand: the investment required, the maturity of their customer and their own maturity and capacity. Keep the insights and wisdom coming.
That’s very insightful! What do you think about ChatGPT being an Enhanced Keyboard? Handwriting -> Typewriter -> Keyboard -> ? It allows to do summarization, translation, and ideating Here’s a blog post where I share my thoughts https://open.substack.com/pub/chrisbora/p/chatgpt-is-just-a-keyboard
Vice President, Global Strategic Sales at Dataiku
1 年This is really good, and a powerful concept. It shows a continuum that aligns well with most maturity models--from pure experimentation to identifying the needs for rules and then finding practical ways to apply rules without just locking everything down. I think a lot of companies get stuck around steps 2 and 3 because they build their ecosystems for specific project purposes, and fail to build with and eye toward enabling future flexibility and value realization.