Challenge

1. An over-reliance on a handful of data scientists shared across business units makes it costly to those business units that utilize the data scientists and they aren’t typically available on demand.

2. The people who were trying to solve difficult problems with big data had no way of knowing if a solution already existed or if others were working on the same thing.

3. The organization promoted many internal applications that had redundant functionality, each with a unique UI and interaction model, leading to confusion and frustration among the users who were supposed to adopt and benefit from them.

Solution

We interviewed stakeholders, product owners, data scientists, data analysts, and the many people who would be affected by and potentially benefit from our work.

After our first round of interviews and multiple Q&As at monthly Town Hall meetings hosted by the Chief Data Office, I led a series of focused workshops, including a GV-style Design Sprint with key stakeholders and would-be users. Coming out of the workshop we decided to take a multi-pronged approach.

1. We audited internal applications in the AI/ML ecosystem that people expressed frustration with during our interviews, and areas of functional redundancy and conflict were identified.

2. Working with the product owners we created a common UI and interaction model that would be used across this suite of tools and immediately started to publish these new patterns, decreasing user confusion and increasing literacy across tools.

3. We designed, tested, and engineered a functional prototype, named “Data Stories”, to serve as the data and AI collaboration hub for the suite of tools we audited, equipping employees from all business units, regardless of experience, with user-friendly yet advanced AI and machine learning tools.

4. We designed, prototyped, and tested a web app that takes queries and data sets and recommends existing ML models or guides non-data scientists through their creation. The prototype was extremely successful with our audience.

This initiative is empowering the client’s users to more easily discover meaningful and trustworthy AI-driven insights, enhancing their ability to serve their customers effectively.

By democratizing data science, organizations drive innovation and stay ahead of the competition, driving consistency, reuse, and collaboration.