Over 90% of hospitality companies are piloting AI. Only about 30% manage to embed it in operations. Nicolas Maynard, Senior Vice President of Data Science & AI at Accor, calls that spread the "industrialization gap," and he says closing it is the single hardest problem in scaling AI across a multi-brand, multi-market hotel group.

Maynard, who spent years building AI products at Amazon including Alexa, joins the Skift Data + AI Summit Europe on October 6, 2026, at Mastercard London. Ahead of the event, he laid out where AI deployments stall at Accor and where he expects the returns to land.

Federated model over top-down rollout

Purely top-down deployments fail across Europe's fragmented systems and localized operating rules, Maynard argues. Accor's answer is a hybrid framework. A central GenAI Center of Excellence defines core architecture, security, ethical guidelines and reusable components. Regional business teams and designated "GenAI Ambassadors" then adapt and deploy the tools locally.

"I have seen many organizations successfully launch AI pilots," Maynard said. "Far fewer manage to industrialize them."

The structure exists to solve a specific European constraint: strict data privacy and security compliance that a centralized push cannot uniformly impose across markets, balanced against the operational realities of individual hotels.

The numbers already on the board

Accor's Guest Relation AI Platform, deployed in more than 200 hotels, has cut guest query response times from 30 minutes to 1 minute. The system saves up to 100 hours per hotel per month by automating the manual drafting that previously kept front-office staff at screens rather than face-to-face with guests.

That metric drives what Accor brands "Augmented Hospitality" — AI absorbing repetitive, high-volume administrative work so staff time shifts back to the in-stay experience, which Maynard describes as fundamentally human-led.

"The most successful deployments I have seen are the ones that make people more effective, not less relevant," he said.

The labor question follows from the metric: Accor measures the strategy in hours returned to hotel teams, which raises the harder question of what each property does with the recovered time.

Three target areas for returns

Maynard identifies three areas where he expects measurable business impact, despite his own warnings about friction:

Hyper-personalized, intent-driven conversion. Moving from static marketing segments to real-time, intent-based personalization on channels like Accor's ALL.com, which he says can drive a 15% to 40% uplift in conversion and average basket size.

Shared-services efficiency. An internal "AI Butler" will act as a conversational entry point for hotel support. By late 2026, Accor targets up to 30% of internal support requests resolved end-to-end autonomously, with another 30-40% AI-assisted, leaving human experts to handle complex cases only.

Process re-engineering. Agentic platforms automating complex workflows across corporate functions — Finance, Legal, HR — to deliver measurable savings and redirect teams toward strategic work.

If he had to bet on one shift, Maynard said, it would be the move from AI assistants to AI agents executing actions autonomously.

Enterprise agents are not consumer AI

Maynard's caveat is direct. Agentic AI in corporate environments is significantly harder than consumer-facing tools suggest. Fragmented legacy systems, strict security and compliance requirements, data quality constraints and the need to orchestrate actions across multiple business processes make enterprise deployment far more complex. That gap between consumer-grade expectations and operational reality, he noted, frustrates employees who encounter enterprise-scale complexity.

For operators planning agent rollouts, the implication is budgetary: integration and data-quality work has to be funded before autonomy can be.

The open call for multi-brand groups

Maynard's federated approach keeps architecture, security and guardrails at the center while handing deployment to regional teams. For other multi-brand groups operating in Europe, the unresolved question is how much central ownership compliance requires before centralization starts slowing the hotels down.

Accor's 30% autonomous-resolution target by late 2026 gives peers a concrete benchmark for what autonomous AI resolution looks like at hotel-group scale — and a date against which to measure it.