Built An AI-Powered Platform That Turns Supplement Browsing Into Personalized Health Conversations
We designed and developed Max Nutrifit, a multi-agent AI platform for a Spanish supplement and fitness brand - replacing static e-commerce browsing with conversational recommendations, personalized guidance, and seamless Odoo integration.
Introduction
Max Nutrifit's parent brand, a supplement and fitness company based in Spain, had a problem every e-commerce store knows too well:
customers would land on the site, browse products, and leave without knowing what they actually needed. The brand didn't just want a better website - they wanted an intelligent platform that could act like a personal health coach for every visitor. One that could understand a user's body metrics, dietary preferences, allergies, and fitness goals, then recommend the right supplements and services through a genuine conversation, not a product grid.
The goals were to:
- Replace passive product browsing with a conversational, AI-driven recommendation experience
- Build a multi-agent system that handles general guidance, supplement recommendations, and service discovery
- Personalize every interaction based on user onboarding data - body metrics, allergies, diet, goals
- Integrate with the existing Odoo e-commerce system for real-time product sync and sales tracking
- Connect users with certified fitness professionals across Spanish cities
The Challenge
Building an AI that can recommend the right products, guide health decisions, and softly sell - without hallucinating, overspending tokens, or losing user trust.
- Inconsistent AI responses: guidance and product recommendations initially operated independently, leading to misaligned outputs and hallucinations
- Multi-agent complexity: a single agent couldn't handle general guidance, supplement recommendations, and service suggestions without breaking down
- Intent detection: accurately switching modes based on varied inputs - ingredient queries, general goals, allergy constraints - required dynamic rule-based logic
- Soft selling vs. hard constraints: balancing business goals (prioritizing client-owned products) with user expectations for precise, unbiased recommendations
- Cost efficiency: preventing unnecessary AI tool calls on generic inputs like greetings to control token usage
- Data synchronization: keeping product and service data in sync between Odoo and Supabase in real time
- Multilingual nuances: supporting Spanish with regional dialect variations introduced inconsistencies in tone and phrasing
Our Approach
7Span delivered Max Nutrifit as a multi-agent AI platform - three specialized agents working together, backed by Supabase, synced with Odoo, and monitored through Langfuse.
Multi-Agent AI Architecture
Split the system into three specialized agents - General, Supplement Recommendation, and Service Recommendation - each with clearly defined responsibilities, replacing the initial single-agent approach that couldn't handle the complexity.
Conversational Recommendation Flow
Designed a step-by-step chat experience where the AI asks targeted follow-up questions about goals, diet, preferences, and constraints before suggesting anything - mimicking a real consultation, not a product push.
Weighted Recommendation Logic
Built a dynamic weighting system that prioritizes user intent - whether they're searching by goal, ingredient, use case, or specific product - ensuring every recommendation is relevant and precise.
Odoo-Supabase Data Sync
Implemented cron-based synchronization that transfers and updates product data from Odoo to Supabase automatically, ensuring the AI always operates on the latest inventory.
Smart Intent Detection & Agent Routing
Built a rule-based system that dynamically identifies user intent and routes queries to the right agent, with intelligent fallback handling for ambiguous or incomplete inputs.
AI Performance Monitoring
Integrated Langfuse for structured evaluation and monitoring, tracking interactions and continuously improving recommendation quality over time.
The Impact
Max Nutrifit transformed a traditional supplement store into an AI-led personal health experience - where every user gets a recommendation that's backed by context.
3
specialized AI agents (General, Supplement, Service) working in concert
2
languages supported (English & Spanish) with dialect-aware responses
100%
automated product sync between Odoo and Supabase via cron jobs
Zero
unnecessary token spend on generic inputs through smart tool-call gating
"7Span's work led to launching the client's AI-guided marketplace platform. The team implemented intelligent tool-calling and dynamic response logic, enabling the AI system to adapt recommendations in real time. Moreover, 7Span's solution received positive feedback from the client's customers."
Engineering clarity where others add complexity. We help businesses build, modernize, and scale with the right technology. Whatever your challenge, stage, or vision, we make IT possible.
Services
For Sales
Looking For Jobs
