Why I changed my mind about IBM’s AI Ladder

Rob Thomas, General Manager of IBM’s Data and Artificial Intelligence (AI) business has talked extensively about a concept called the “AI Ladder”, and how “AI is not magic: it’s Computer Science!”.

Put simply, the AI Ladder provides a practical guide for understanding the process by which an organization can reap the benefits of AI:

• First, COLLECT data (because AI feeds on data)
• Second, ORGANIZE the data (so you can know and trust what you have)
• Third, ANALYZE the data using business intelligence, machine learning (and other models)
• Finally, INFUSE the data into your organization and its business processes

As Rob explains in this conversation with Tim O’Reilly, it’s hard work but there is a clear path best tackled through small experiments.

Now for a confession: when I first heard about the AI Ladder almost 2 years ago, I was not impressed! Why? Because:

• Looking at it through my computer scientist’s lens, the AI Ladder simply states the obvious. OF COURSE you need to have data that is organized, before you can analyze it or train a machine learning model! OF COURSE your models must demonstrate a meaningful improvement on the status quo before the people in your organization will embrace them. And OF COURSE the more detailed, organized and voluminous the data, the better your models will perform. And so on.
• My firm QueBIT (an IBM Data and AI Business Partner) had posited our own version of the AI Ladder two years before Rob Thomas (see side by side comparison below), and have moved away from it to a more successful (for us) solution selling approach. For example, today we talk about our Predictive Demand Planning solution when talking to supply chain planners, and about Embedding AI in FP&A made easy when talking to financial planners.

This brings me to the point of this blog post: I have changed my mind, and now I wholeheartedly agree that the AI Ladder framework is relevant, insightful and valuable - even though my original reasons for rejecting it remain sound!

Let me explain the apparent contradiction!

When I first learned of the AI Ladder, I evaluated it from my own narrow perspective as a computer scientist, and an IBM business partner whose mission it is to guide real companies through digital transformation to solve real business problems with real data and build real models used by real people to do their real jobs better.

Now I see (duh!) that Rob Thomas’ audience is much broader than mine, and more senior. He is reaching out to, and likely being heard by, CEOs, CFOs and CIOs who have been primed to listen by analyst firms like Gartner, IDC, Forrester and others (not to mention IBM’s own Watson ads on TV!)

Rob’s timing is also better than ours: when QueBIT presented the Layers of Analytics to Vice Presidents of Finance or Operations in 2016, we elicited a combination of these disengaged reactions:

1. This is interesting but abstract
2. I just need better planning/reporting/dashboards: this is too broad
3. This makes sense but I would never be able to sell the idea to my boss

If we were to revisit the Layers of Analytics (or, more likely, adopt IBM’s AI Ladder) in our sales presentations today, I believe we would get more engagement. Today, any boss worth their salt is paying attention to the potential of AI and should be receptive to having a conversation about it. That same boss - unless they have a background in technology (and most don’t!) - likely has a lot of questions about practical aspects of getting it done, and the AI Ladder IS an excellent framework for telling the story. They may variously be excited by the opportunity to gain business advantages through AI-driven efficiencies or be scared that competitors will beat them to it and leave them in the dust.

The clincher comes from understanding the very special role that IBM Business Partners like QueBIT play, which is different from IBM’s. While Rob Thomas and IBM are selling the IDEA of the AI infused enterprise, consultants like QueBIT need to take that idea and turn it into a vision, then a roadmap, and finally a working solution that is tailored to each customer’s specific needs and capabilities. IBM is like the manufacturer that supplies bricks, wood or concrete to the construction industry. QueBIT is the architect, builder and interior designer who works with you to build the house of your dreams, that fits with your budget and your current reality.

Even though we share customers with IBM, each business partner’s target market is but a sliver of everyone IBM is talking to. Business partners specialize in specific industries or specific functions, or combinations of the two. It’s the best of all worlds: IBM invests in technology platforms and broad visions for how to use them, and business partners invest in being experts on and developing solutions for one unique business at a time. They go broad, and we can go deep.

IBM and QueBIT participate in a symbiotic relationship, with a common goal of customer success. QueBIT needs IBM to educate executive leaders on what it takes to embrace AI in their businesses, and to set their expectations that this is just as much a cultural and business process challenge as a technology one. IBM then needs partners like QueBIT to help companies map out a path to digital transformation, that balances their aspirations with resource constraints. The AI Ladder is a valuable tool in pursuit of this end.