Making Data Science and AI Approachable in your Organization

Scaling AI Tony Olson

This blog post is part of a series of guest publications by Excelion Partners.

When you’re getting ready to pitch a new AI initiative, it can be difficult to get buy-in from the rest of your team. One way to make it easier to get the rest of your team on board is to make data science more approachable within your organization.

Get rid of the noise

There is a lot of noise surrounding data science and AI. As a result, people have preconceived notions that could be detrimental to your cause. By educating around the basics of data science and promoting a data literate culture, you can clear up that confusion and ensure your team is on the same page.

One good way to start educating your team on AI and its business impact is by walking through the levels of data science capabilities and exploring the benefits of progressing to more advanced levels.

In order, the four levels of data analytics can be defined as:

  • Descriptive - What is my customer doing?
  • Diagnostic - Why is my customer doing that?
  • Predictive - What will my customers do?
  • Prescriptive - How can I make my customers do what I want?

Each different level provides a more advanced degree of data science capabilities, helping your organization answer different and more advanced questions about their business. In essence, transforming your business from being reactive to proactive.

Gartner 4 Analytic Capabilities

Source: Gartner (October 2016)

Be transparent

It’s easy for data science and AI to just be this “secret magic” that most of your team doesn’t understand, however that isn’t beneficial in the long run. Instead, communicate with your team transparently about your efforts.

Demonstrate what you’re going to do to get the answer they’re looking for. Combined with the right tool, this makes it very approachable for your team to see what’s really happening. Do this by:

  • Walking your team through the data sources you’re using.
  • Explore example actions your team will take to clean and prep the data.
  • Demonstrate how to build a feature.
  • Work through an example of an interpretable model to show that this is not magic.
  • Share the different mediums (graphs, charts, and APIs) that can be used to consume insights.

Pro tip: When educating beginners, Decision Tree models are very easy to explain and comprehend how predictions are made.

The benefit of making this approachable for your team doesn’t only lie in getting buy-in. When your team understands your efforts, they can contribute their individual experience and knowledge (even if they’re not a data scientist) to improve your initiative.

Use tooling to help visualize and explain

A tool that can walk people visually through the process of data science goes a long way. It’s important, however, that this tool goes end-to-end (clickers to coders) in the data science process in an approachable and valuable way.

As you approach your next data science initiative, it’s imperative that you make it approachable to other team members in your organization. The most impactful way of doing this is by educating your team. Teach your team about what you’re doing and how it works so you can be on the same page. By doing this, your team then has an opportunity to contribute as well.

 

This blog post is part of a series of guest publications by Excelion Partners, a data science and IoT consulting organization focused on building solutions and helping you discover evidence that creates better business decisions. Check out their blog and read about their projects here.

You May Also Like

AI Platforms in 2020 and Beyond: What to Look For

Read More

Is It AI or Not? A Score Card With 4 Dimensions

Read More

Infographic: How Can Data Quality Be Improved?

Read More

Diversity and Inclusion in Data Science

Read More