Mark III Systems Blog

If a tree falls in a forest and no one is there to hear it… does it make a sound?

Similarly, if an AI model is trained with incredible amounts of training data with a team of extremely experienced data scientists and researchers and is REALLY accurate on predicting something, YET business or institutional stakeholders can’t actually use it in their natural UX and within tools that they use everyday to be productive, does it really even exist?

So much time is spent on data science and optimized systems and software to actually gather data and train AI models that most organizations we’ve seen aren’t spending enough time thinking about how it will actually be consumed by their actual customers, partners, and/or workforce.

Will it actually be presented in a way that the stakeholders will love and be truly additive to the enterprise and what these users do everyday?  Will it be consistently updated to ensure that its always usable and timely?

Or will it be unnatural and not adopted because it isn’t within that user’s natural workflow?  Or worse, will it cause users to have to conform and be less productive than before?

When we look at the overall AI pipeline of an organization, this is something that we feel everyone needs to consider from the beginning and in tandem with any AI strategy around building and training their AI models.  The UX and how the model is actually consumed in the field after it is trained can happen anywhere and on anything (this is the “inference” layer, if you want to use AI terms), but more and more so its happening on the edge or on IoT or mobile devices – where actual people are.

And when you talk about UX, its not just via a web app or mobile app, its also via voice assistant, IoT devices, and other integrated actions and apps that may connect to “the end” of your AI pipeline.  These design skills are often not found in the same organization as data science and not accounted for, but they are absolutely needed for the future success of practical AI in the enterprise.

So when someone asks you what your AI strategy is, please don’t just think about the HOW (data science and AI models and optimized servers).  While they are important, it’s difficult to be successful without ALSO thinking about WHY you’re considering leveraging AI in the first place--- delivering amazing experiences and outcomes to HUMANS, both your customers and your incredible people.