Working with Lutech, Metrovacesa launched MiA (Metrovacesa Intelligent Assistant) after a rapid eight-week implementation. MiA isn't a simple button-based bot; it is a multi-agent system that moves beyond traditional linear chatbots using Google Cloud serverless infrastructure along with Dialogflow CX and Vertex AI.

MiA's layered approach includes:

• Conversational layer: Dialogflow CX acts as the primary delivery and routing layer, managing sophisticated, multi-turn conversations and state management while allowing users to switch seamlessly between complex topics like pricing, property specifications, and financing.
• Intelligence layer: Vertex AI and Gemini models power the natural language understanding and generation components. Gemini models extract and qualify lead information from unstructured dialogue, enabling the assistant to understand intent in Spanish, English, and French.
• Knowledge-base layer: A custom website datastore built by Lutech with Vertex AI search utilizes the entire Metrovacesa website as a primary knowledge source, ensuring that property information remains accurate and consistent across all channels over time.

This architecture allows MiA to handle complex inquiries about property specifications, pricing, and availability across 100+ developments while maintaining sub-five-second response times.

The solution integrates seamlessly with Metrovacesa's Dynamics 365 CRM, ensuring that every interaction enriches the customer profile with purchase motivations and property preferences.

Lutech's role was pivotal, acting as architects and engineers to ensure the system could scale automatically during peak traffic, such as new development launches. This "always on" model transformed the top-of-funnel experience, moving Metrovacesa from a reactive call-based process to a proactive, data-rich strategy.

The implementation of MiA on Google Cloud has turned every customer interaction into a source of strategic data, providing Metrovacesa with visibility that was previously impossible to capture manually. Through custom-built Looker Studio dashboards, the team now tracks specific metrics that influence their broader commercial strategy, such as:

• Expanded market window: Real-time analytics revealed that 33% of potential customers are active when physical offices are closed.
• High-conversion micro-windows: Conversations occurring on Saturdays show a 44.0% conversion rate, while Sunday interactions reach 52.9%, significantly outperforming standard business day averages.
• Global reach: The system captures demand from international investors through seamless multilingual support in Spanish, English, and French.
• Campaign precision: Integration with Google Marketing Platform allows the marketing team to track UTM attribution, identifying exactly which portals or campaigns drive the most qualified interest.
• Product fit analysis: By capturing purchase motivations and specific property preferences during AI interactions, Metrovacesa can align its development pipeline with emerging market trends.