Retaining customers using AI with Virginie Marelli

How do you use AI to increase customer retention while lowering marketing costs? The answer is churn prediction.

In this episode of the AI and Digital Transformation podcast with our guest Virginie Marelli, Head of research at dataroots, we will discuss how to use churn prediction to identify the customers who most likely to leave your business and how to make them stay.

Tune in to find out how to add AI and churn prediction to your marketing strategy!

Who is Virginie Marelli?

She started out her career at ING when AI was still called data science and big data. Thanks to self-learning and her inquisitive nature to solve problems with data, she jumped at the opportunity to join the Leuven-based AI startup dataroots in 2017.as a machine learning engineer.

When dataroots grew from start up to scale up, her role in the company evolved, as she became its tech lead in AI.

As of 2022, Virginie heads dataroots Research, the umbrella for all R&D projects and (open source) knowledge sharing initiatives. Also in 2022, she was one of the inspiring 50 in Deep Tech Benelux.

Check out our show notes to know more about Virginie.

Chapters

[00:00:00] Introduction of our Guest Virginie Marelli

[00:03:01] The importance of a Mathematical education in AI

[00:07:01:] What is churn prediction?

[00:10:13:15] The target of churn prediction

[00:13:10:15] How is success estimated for a churn prediction model

[00:14:31] How much data so you need?

[00:17:32] How do you monitor the performance of the model in production?

[00:22:17:15] The importance of ML operations for reliable model performance

[00:34:20] Churn prediction and A/B testing

[00:36:50:15] The importance of explainability for a successful ML solution

[00:42:50] Data privacy

[00:46:52:15] How to choose a good A.I. solution provider

[00:51:41] Virginie’s book recommendation

---

Follow us on our socials:

• ⁠⁠LinkedIn⁠⁠
• ⁠⁠Twitter⁠⁠
• ⁠⁠YouTube⁠⁠

⁠⁠Book an appointment⁠⁠ with us.

---

Music credits: storyblocks.com

Logo credits: ⁠⁠Joshua Coleman, Unsplash