Prets helps homeowners make informed and sustainable decisions when upgrading their homes. Their platform connects consumers to budget-neutral energy-efficient upgrades based on individual circumstances like house type, budget, and regional subsidies. As the complexity of this guidance grew, Prets needed more than a static recommendation engine. They needed an AI system capable of reasoning, personalizing, and proactively guiding users the way a human expert would. That's where our collaboration began.

The Challenge

Home upgrade decisions are among the most complex consumer choices — shaped by house type, budget, local subsidies, and expected energy savings. Prets recognized that guiding homeowners through this process required far more than a static recommendation engine. They needed an AI system that could reason, personalize, and proactively navigate users through a highly individualized, multi-step decision journey. Three core challenges stood in the way:

  • Reasoning across complex, diverse data: The system needed to connect and reason across structured user data, energy models, regulatory content, and financial information simultaneously — something traditional tools were not equipped to handle.
  • Replicating human advisory expertise: Generic automation couldn’t capture the nuance and intuition of a human advisor. The system needed to adapt dynamically to each homeowner’s unique situation and guide them proactively, not just respond reactively.
  • Navigating a multi-step decision journey: A single interaction wasn’t enough. The AI needed to accompany users across multiple touchpoints, maintaining context and continuity throughout a complex, personalized decision process.

Our Approach

Following a design sprint to define the architecture and identify the right intervention points across the customer journey, Enjins collaborated with Prets to design and build a production-ready multi-agent AI system that acts like a team of proactive experts.

  • Multi-Agent Orchestration: Agents were embedded throughout the customer journey — not as simple chatbots, but as autonomous collaborators that trigger the right actions and guidance at the right moment, adapting dynamically to each homeowner’s context.
  • Deep Platform Integration: We built the underlying AI and data infrastructure, connecting structured user data, energy consumption models, regulatory content, and analytics. LLMOps and retrieval-augmented generation enabled agents to reason dynamically across all inputs in real time.
  • Team Enablement: We onboarded and trained Prets’ internal data and AI team, equipping them with the tools and expertise needed to scale and evolve the system independently after delivery.

 

Share:

More Cases

SnappCar
New ML Ops setup delivers 50% faster queries and significant cost reduction for peer-to-peer car-sharing platform.
Crisp
Implement a new forecasting system for Crisp that improves prediction accuracy (WAPE) by >20%.
CircularPSP
Turning fragmented municipal data into actionable circular economy intelligence for policymakers across Europe.