ENGIE: When the support queue becomes the bottleneck, an agent answers first with Dataiku
November 4, 2026/3 min read/Ann Claire Egan
ENGIE Supply & Energy Management uses Dataiku, the Platform for AI Success, to answer routine support requests the moment they're raised, so a broken job doesn't sit in a queue while the energy market keeps moving.
250+
Dataiku users
~80%
non-technical Dataiku users
~20%
productivity gain, giving a team of three an estimated day back each a week
ENGIE Supply & Energy Management (S&EM) is ENGIE’s global business unit that combines energy supply, trading, risk management, and asset optimization expertise to deliver secure, affordable, and increasingly carbon-free energy solutions to businesses, consumers, and energy asset owners worldwide.
They use Dataiku to replace manual Excel and Power BI processes, run ETL, and work with large volumes of data, including IoT data from wind and solar farms. Supporting that community fell to a central three-person team.
As adoption grew to 250+, across 80% nontechnical roles in business, back office, HR, and risk functions, so did the questions: failed jobs, connection issues, administrative requests, and questions about whether projects followed ENGIE's standards. For users, that meant waiting. For the support team, it meant more time firefighting and less time helping the community improve.
From answers to action: An agent that fixes what it finds
ENGIE built an AI support experience designed to do more than answer questions.
A Dataiku user at ENGIE can bring it a failing job and ask the agent to investigate what went wrong. It can review a project against ENGIE's internal standards, complete administrative tasks, find relevant guidance, or assemble the context needed to escalate an issue to help 250+ people keep moving without needing to be Dataiku experts.
Agent Hub coordinates four specialized agents for troubleshooting, guidance, knowledge retrieval, and human escalation, behind the scenes. Users don't have to know which agent they need. They simply bring the problem, and the agent takes it from there.
Agent orchestration across 25 servers
Dataiku coordinates support across the distributed environment of 25 servers, using roughly 24 purpose-built tools for actions like configuring connections, retrieving account information, and checking project standards. The system doesn't simply tell someone how to solve a problem. It understands the request, finds the right capability, and takes action.
"The agent lets us resolve the routine issues automatically and put a person in the loop only where judgment matters, so we spend our time helping people build better rather than firefighting."
Stéphane Raguideau, Digital & Data Accelerator, ENGIE Supply & Energy Management
Giving the support team back a day a week
ENGIE estimates the support agent can deliver about a 20% productivity gain for its three-person team, which means roughly one day every week returned to higher-value work. That time can go toward extending beyond troubleshooting—using AI to help people work more independently while making good practices easier to follow.
For a three-person team, that changes the equation. Supporting a larger community doesn't have to mean answering more questions. The agent handles the repeatable work. The team focuses on making the community better.
