The 4 Stages Of Data And AI Product Design: From Idea To Production
The 4 Stages Of Data And AI Product Design: From Idea To Production
According to Baird market strategist Michael Antonelli, the second most common question investors are asking his firm is, "Is {the company they are evaluating} investing in AI?" But investing in data and AI isn't enough, as we've seen for over a decade. Even today, 60%-70% of data teams struggle to deliver value to the business (2023 Gartner CDO Survey).
It's not for lack of ideas. Six months ago, GPT got the idea pump flowing at every business. Prototypes that people built in public and the wave of product demos have given people lines of thought to follow. There's a greater understanding of what's possible.
C-level leaders who have reached out to me this year have a better frame of reference for AI products. They aren't looking for proposals and proofs of concept or struggle to come to terms with practical applications. They have specific use cases in mind and want to understand how to get from ideas to returns.
C-level leaders I'm working with realize there's a chasm between the two. A year ago, most approached me thinking their business was missing a few pieces and just needed a guide to point them out. C-level leaders are looking at this holistically and as an enterprise-wide transformation.
Getting from data to profit requires several frameworks. The most important among them covers how the business gets from use cases and ideas to products. In this article, I will explain the frameworks and discuss how I've used them to bridge the chasm. I introduced the 4 stages in this article and will explain them in more depth here.
I am looking at offering another live course for 1) Data and AI Strategist Certification and 2) Data and AI Product Management. If you are interested in enrolling, please email me (info@HighROIAI.com) and let me know what class, day, and time work best for your schedule. Subscribers get 15% off early enrollment pricing.
Stage 1 - Monetization
We typically have a use case that needs additional definition at this stage. I call this the problem space definition. As the name implies, the focus is on the problem to be solved. Framing it that way forces us to define what's broken about how things work now. One of the most significant barriers to monetization is spending substantial amounts of time solving already solved problems.
Businesses operate and serve customers today without additional data or AI products. For many use cases, the technology in place today is more than sufficient to meet the need. Those problems are already solved, and while data can provide an incremental improvement in functionality, that won't translate to very much value creation. Data teams waste a lot of time working under the assumption that any progress will deliver value, and that's simply not the case.