How Beamy Uses Google Cloud to Drive Enterprise AI Transformation

In large organizations, digital transformation hits a major roadblock: truly understanding how work actually gets done. With the rise of AI and SaaS, technology adoption decisions are increasingly made at the department level, creating a blind spot between what exists in the official IT inventory, what teams actually use, and what drives business value.

Beamy, founded in 2017, built an AI-driven Business Transformation Platform to address this gap. The platform uses a browser extension to observe real-time employee activity—opening tickets, modifying CRM fields, validating invoices—and transforms these interactions into structured logs. AI models then reconcile events to reconstruct business process sequences across the entire application estate, revealing the living map of an organization's digital operations.

To deliver this analysis at enterprise scale, Beamy built its infrastructure on Google Cloud. Vertex AI handles model deployment and inference, managing resources dynamically without requiring Beamy to manage AI infrastructure directly. Cloud Run executes application services in a serverless architecture. BigQuery provides large-scale data storage and analysis across billions of behavioral data points. Looker delivers actionable dashboards to clients.

"The richness of Google Cloud's managed services plays a decisive role: it allows us to put our models into production and operate them at scale without the operational complexity of infrastructure," says Andréa Jacquemin, Founder and CEO of Beamy. "Vertex AI facilitates model deployment, manages inference automatically, and dynamically scales resources."

At Veolia—a major European industrial group—the impact was immediate and striking. Veolia expected a catalog of about 4,000 applications; Beamy revealed over 10,000 solutions were actually in use. Beyond the initial discovery, Beamy enables Veolia to rationalize software licenses based on actual usage patterns, achieving near-immediate ROI, and to strategically prioritize which of 1,000+ identified AI initiatives to deploy first based on where employees actually spend their time.