Adam Bede

    M3: Data Strategy & Gov.

    Professor: Krishna Cheriath

    What has enabled AI

    • Growth of processing power (general processing unit) and cloud

    Ratios to watch:

    • # of AI papers across countries

    Klarna use-case of firing 60% of workforce

    “Before any headcount comes to me, here’s what I’ll ask:”

    1. Human-only
    2. AI-only
    3. Human-augmented

    AI-first organizations

    • Strategy: If your AI strategy doesn’t create dissent, meaning everyone agrees, then its meaningless
      • Simplicity (parsimony + Occam’s), don’t need competing digital strategies, should paint a vision of the future.
        • To do that, reimagine how to design for adoption and usability
      • “Does it describe the value imperatives in a way that a 10-year-old can understand?”
      • Need a data strategy, need integrated w/ a value focus for the business that is translated into the functions that will use it… Identify where we seek to create value in the business.
      • 1. How do you create value? 2. How do you improve the bottom line? 3. How do you improve the customer experience? 4. How do you improve the employee experience?
      • “none of that matters if you aren’t willing to adopt or change (re-imagine how you work) and how do you ensure the people in your org are digitally fluent”
      • He hates “enterprise data strategy”; start w/ the most important decisions and actions.
      • Army Vantage as an example of how an enterprise play obscures value
    • The Basics
      • Quantitative intuition + isolating key actions and the outcomes we want them to drive + I Wish I knew and I wish I could frameworks to frame biz problems
      • Minimize data entry; explain to the users the value of the data
      • Reject the premise: Who is using the tech (don’t count those servicing it!); then what value/impact is it driving (has to be a decision or action). What are you trying to achieve (impact, decision, etc.)
      • Golf example: AI caddy requires 0 data entry, gives you information when you need to make a decision
      • Data as a broad definition (see slide)
    • Examples
      • Amazon shopping cart prediction >90% and they start the supply chain process
        • The idea that we have free will
    • Framework
      • Tech foundation - set your system’s of record; tech blueprint before any acquisition
      • Am I maximizing the value of my existing system
      • Examples
        • Data decision registry (decision log that creates a network as well)
        • Analytics and insight marketplace
        • Ask about his vantage workflow
        • Ferrari vs. Cargo Ship
      • Roles: Break them into tiers and define a foundation and then bucket the technical roles and the nontechnical roles into categories that you can generalize what kind of behaviors and abilities you need