The bridge between cool & right AI/ML solution

Trust me when I say this - In Data Science teams today, any conversation of innovation almost instantaneously gets converted into a potential Artificial Intelligence (AI)/ Machine Learning (ML) cool technology implementation opportunity.

However, so many AI/ML endeavors run until the stage of innovative proof of concepts, and end up not being actively adopted by end users.

Is your team in this boat? Have you been facing trouble getting your AI/ ML initiatives adopted by end users? Or maybe you got a few initial users but are struggling to scale?

If yes, then read on...

Most AI/ ML courses and books focus on the tools and techniques of implementation. For a typical ML implementation, your focus would mainly be on - Gathering Data, Preparing/ Cleansing/ Transforming Data, Partitioning into Training/ Test datasets, Selecting and running ML models, and finally fine-tuning the model to generate recommendations. So, it is very easy for engineers to get caught into the excitement of cool technology innovations using AI or ML.

However, the most important activity at the start of any project often gets neglected: FRAMING THE PROBLEM

Like any other technology, AI and ML solutions don't exist in silos. They need to be relevant for users. Without spending enough time articulating the true problem at the start, several months of effort building an AI or ML solution could completely go to waste...

This is where the product managers in your teams could come in and help.

Product Managers need to play the role of a BRIDGE between cool technology and its realized business value.

Here is why:

1. They spend time establishing relationships with end customers/ users who are using existing products/ apps, understand their day in the life and pain points, and are constantly thinking of ways to reduce those pain points.
2. They spend time understanding end-to-end business processes and building the industry domain depth.
3. They spend time digesting knowledge about the various complementary products/ apps in the ecosystem, which teams owns those products and who the right points of contacts from those teams are.
4. They actively work on articulating the OKRs (Objectives and Key Results) for the problem, set customer/ stakeholder expectations about assumptions, timelines & limitations of the solution and also help demonstrate and communicate the solution in a jargon-free and business friendly language.

This is a huge responsibility on a product manager's shoulders. It is OK to take ample time at the start to do the research and crisply define the problem and pain points, even if it ends up taking several weeks or months. After all, at the starting stage, heavy investments in terms of money, time and energy have not yet gone in from engineering teams.

On the other hand, it is imperative that engineering teams involve their product managers into AI/ ML innovation projects right from day one, and work together as a team. Bringing product managers later in the game simply to land user adoption of a cool solution, carries a major risk as the problem itself might not be painful enough for the users to adopt and could lead to the entire investment going waste.

So, next time, your team embarks on an AI/ ML innovation proof of concept..