AI Pulse by Inblix

Mirakl's new mandate: Every employee must now build AI agents

OpenAI Blog · Jul 11, 2026 · 2 min read · Read original article →

Curated by the Inblix editorial team


Featured image for article: Mirakl's new mandate: Every employee must now build AI agents

Mirakl isn’t just dabbling in AI. The commerce platform company has drawn a hard cultural line: every employee, from technical writers to support staff, is expected to be a builder of AI agents, not just a user of them. Co-founder Adrien Nussenbaum put it bluntly—the initial vision of 100% adoption gave way to 100% active development, tasking every worker with redesigning their own workflows.

I’ve seen plenty of companies claim AI fluency, but Mirakl’s internal numbers back up the talk. Technical writers using ChatGPT Enterprise slashed documentation time by 70%. A custom-built support agent, drawing on Mirakl’s own knowledge base, drove a 37% efficiency boost while keeping satisfaction at 96%. Those aren’t marginal gains. Anne-Claire Baschet, the company’s Chief Data & AI Officer, framed the next step as scaling value by identifying cross-team workflows and applying leadership focus to make them happen.

The product side mirrors this evolution from passive assistance to active agency. The standout is the Catalog Transformer, which cut the grueling seller onboarding process down by a staggering 91% while halving categorization errors. For marketplaces, that’s a direct line to faster revenue and cleaner data. It moves AI from a helpful co-pilot to a system acting on behalf of the user within structured limits, with humans still holding the reins for nuanced judgment.

Looking ahead, Mirakl is planting a flag in agent-driven shopping with an infrastructure play called Mirakl Nexus. The idea is to handle multi-merchant baskets and complex transaction flows so retailers don’t have to rebuild their stacks from scratch. Nussenbaum’s advice to leaders is to set system-level ambitions, not just optimize features. It’s a smart bet that the next era of commerce won’t be browsed—it will be negotiated by machines on our behalf. Whether shoppers are ready to trust agents with their carts is the multi-billion-dollar question.

💡 Key Takeaways

  1. Mirakl shifted from requiring 100% of employees to use AI to demanding that 100% actively build agents, fundamentally changing workforce expectations.
  2. Internal deployment of AI agents led to a 37% improvement in support efficiency and a 70% reduction in documentation time, proving value before productization.
  3. The AI-native Catalog Transformer slashes seller onboarding time by 91% and cuts categorization errors in half, directly improving marketplace revenue velocity.
  4. Mirakl Nexus represents an infrastructure bet on agent-native commerce, preparing for a future where AI acts on behalf of shoppers and merchants across entire transactions.

Keep reading: See related articles below for more coverage on this topic.

Get smarter about AI

The sharpest AI news, curated daily. Delivered free to your inbox.

Learn more

Glossary terms

← Back to all articles