CognitiveScale Product VP on The Maturation of AI Tech (PART 3)

Consider the complexity of managing the DevOps lifecycle with real production environment feedback. How do you retrain these models? Furthermore, what’s the ideal user experience you want to provide? Data scientists and engineers might understand the concept. But remember – they’re data-minded and might not see a clear connection. You need an intermediary user interface to collect the data. As a result, the basics of AI tech and product management must go together. 

Who’s Your User?

Do you know your users and what drives them? How does the value of your application manifest itself to them? For certain users, having them trust a black box prediction will negatively impact adoption. For example, a wealth advisor application may suggest client investment options, but a good advisor won’t take those predictions at face value. They want the ‘why’ behind the recommendation. This is called explainability. In industries such as healthcare, defense, intelligence and wealth advisory, the higher the subject matter expertise, the more they will ask why. So, supporting evidence is critical. Without explainability, users don’t trust the system.

In other AI and product management scenarios, explainability is less important. Large e-commerce sites provide recommendations based on your known preferences. People don’t ask why in these scenarios. It’s more about impulse and emotion. 

Trust Your Engineering Team

For nearly any project, you’ll have assumptions that need validation and risks to be mitigated. In general, it’s best not to build up to giant waterfall releases. You can make sound judgments with a good team of data engineers looking at goals and data volume/characteristics. From there, determine if you can produce an accurate model to create a starting point. Ultimately, it’s critical to partner with and trust the insights of engineers and data scientists. 

Where’s The Value?

When combining AI and product management, try to measure value as much as you can beforehand. Most importantly, can you trace it up to tangible business value? Sometimes it’s better to start with a business outcome and drill down. 

For example, think about scalability and end-user adoption. In other words, this is how you drive transparency into AI applications to understand how they perform from a business, machine learning and technical perspective. Plus, it must be delivered in a responsible way. Ultimately, these are the big issues that lead to customer struggles. Specifically, the reason why is great AI products must focus on real business outcomes.

In the end, every AI tech or product management strategy should address a specific, desired outcome. For example, are my wealth clients doing better? Furthermore, is advice paying off? Most importantly, am I retaining clients? Tweak the algorithm to drive towards concrete outcomes. 

 

Click here for Part 1

Click here for Part 2

About the Speaker
Jon Richter
CognitiveScale VP Product
Jon Richter is the VP of Product at CognitiveScale - leading the product team in developing AI solutions for businesses who are looking to streamline operations with augmented reality solutions. Prior to joining CognitiveScale, Jon spent over a decade at IBM - where we worked as a program director in developing Waston (one of the first AI products at scale). Prior to working on Watson, Jon was the lead software developer for IBM's software services division. Jon holds an MBA from The University of Texas at Austin and currently lives in Austin.

Recent Posts