AI and the person behind the curtain
I’m sure most of you have seen the Wizard of Oz. Dorothy and her pals go to the emerald city to see the Wizard of OZ. The wizard looks like an immense disembodied head with a short temper. Given it was filmed in 1939, It wont come as a spoiler that in reality the wizard was a man behind a curtain running a bunch of machines.
So what has this got to do with AI?
In large companies there is a real possibility that AI apps and services end up being much like the wizard of oz. Spectacular and seemingly automated on the outside. With an army of people pulling and pushing levers in the back-office.
There are plenty of factors that might limit our adoption of AI.
Skills, Operating Model and Accessibility of Data are my top 3. This week what worried me is accessibility of data.
AI (much like analytics) apps and services rely on data, a lot of it. There will be AI apps and services that are in themselves the source of new and innovative data. But I wager that in large companies the first use cases will be exploiting existing data.
Beyond the proofs of concept. Beyond the experimentation with massive PR claims that “we are doing AI”. I’m referring to industrialisation of use cases.
Accessibility of data is key. otherwise the use cases or journeys we chose might be limited to what we “can” access and have limited value.
Accessibility of Data for me is; Discoverability, Access and Exploitation. These are practical things. As such they are not binary. They are a journey.
In simple terms;
I’m sure you all see how obvious these things sound. But in large companies with a legacy of technologies, you have to work towards these goals.
In some cases the operating model, incentives, goals, etc of the company are not aligned to make data succeed. Thus the “journey”.
Otherwise what happens is all applications and services, which hold your wealth of data, tend to work for themselves and not the greater good. They become insular and incompatible with others discovering, accessing and exploiting their data.
We need strategies and processes that move us TOWARDS the goals of discovering, accessing and exploiting data across all data.
I like to think of these as Data Foundations. They are as relevant to AI as they are to analytics and other information dependent apps or services.
Whilst most of what we see around AI are the use cases and the sizzle.
Get thinking about your data foundations!!
This article was originally posted here and was reposted with permission.
