No-Code AI for Demand Prediction

Intro

With 97GW production capacity, 30% of which is renewables, ENGIE is one of the largest utilities in the world and a leader of the Zero Carbon Transition. ENGIE uses C3 AI technology to deliver energy management services to public, enterprise, and consumer customers and to increase the productivity of its energy operations.

Challenge

When COVID-19 struck France in March 2020, ENGIE needed to rapidly determine how stay-at-home measures would affect customer energy bills. They faced two obstacles—

  1. ENGIE product owners tasked with monitoring energy usage and ensuring customer satisfaction did not have the data science depth required to build sophisticated predictive models.
  2. Completing the work meant stitching together diverse smart meter and customer systems, building models in Python, and training them on local machines—slow, painful, manual work with no promise of a high-performing model at the end.

“COVID-19 presented a serious challenge for the business. With conditions changing every day, we needed to quickly determine how our customers were going to be impacted so that we could prepare them for the effects of the nation-wide shutdown,” said Erwan Conq, CIO of ENGIE’s Electricity & Gas French retail business unit.

Approach

ENGIE turned to C3 AI Ex Machina to overcome these challenges. Using C3 AI Ex Machina’s no-code interface and native cloud scalability, product owners were able to rapidly configure a predictive consumption model.

The ENGIE project ingests daily energy consumption for residential customers, third-party weather data, and customer profile data, runs it on a model trained on the first few weeks of quarantine data, and compares the results with traditional pre-containment models to determine variances.

As a final step, model results are made available to the ENGIE User Experience team responsible for customer alerting in order to ensure that residential customers remain informed and satisfied throughout sustained quarantines.

“In just a few hours in C3 AI Ex Machina, our product owners built an accurate simulation, trained on all the data available, that let us project how COVID-19 was going to impact our residential customers,” said Conq. “Delivering these insights before events had even unfolded, without demanding time from our data scientists, was unheard of before we had this solution.”

Next Steps

Given the success of its initial C3 AI Ex Machina project, ENGIE has decided to address additional use cases with the product:

    • ENGIE will use C3 AI Ex Machina to profile electricity load curves and identify savings opportunities using high-frequency data from smart meters. Ex Machina’s ability to scale up to handle these data volumes and its large library of out-of-the-box ML algorithms makes it an ideal prototyping solution for product owners.
    • ENGIE will use C3 AI Ex Machina to accelerate its customer energy claims process. C3 AI Ex Machina will automatically identify when bills appear anomalous, eliminating the current manual investigative burden. The result will be fewer anomalous bills sent out, faster complaint resolution, and increased customer satisfaction.

“C3 AI Ex Machina is a new tool in ENGIE’s future digital efforts,” said Conq. “For the first time, our product owners and analysts can build predictive solutions with a no-code tool, thus saving our data scientists time in data exploration. Finally, it will let us provide better services for our customers and improve the safety, reliability, and efficiency of our grid.”

Want more information?

Contact us
Free Trial