Use Case · Energy Renovation

IDEE

idee-eco.fr

Next.js · Supabase · SEO/GEO · AI Chatbot

Next.jsSupabaseGEOSEOAgentic AIChatbot

Real calculations, not ranges. Cited by AI, converted into leads.

IDEE (Institut des Économies d'Énergie) is a lead-generation platform for French energy renovation subsidies (MaPrimeRénov', CEE, solar, heat pumps, insulation, air conditioning), with a second B2B track dedicated to the tertiary sector decree and professional renovation work. Busony designed and operates the entire site: a dozen interactive simulators with real calculations, an AI assistant trained on a dedicated knowledge base, and an ongoing technical SEO effort to maximize visibility in classic search and generative engines alike.

The context

The energy renovation subsidy market is saturated with approximate tools and largely absent from AI answers.

🧮

Simulators that guess

Most simulators in the sector display generic regional ranges, without cross-referencing official housing data or the household's real tax profile.

🤖

Search moves through AI

A growing share of project owners ask ChatGPT or Perplexity directly how much they can claim — a channel where most lead-gen platforms are absent.

📋

Two audiences, one generic door

Individuals and professionals (tertiary sector decree) have very different needs, but often share the same generic experience on competitor sites.

What Busony built

A site designed to calculate accurately, answer accurately, and be found — by humans and AI agents alike.

🧮

Interactive simulators — the solar ROI showcase

The solar simulator is the most advanced of the five: Google Solar satellite imagery to analyze the roof down to the square meter (orientation, tilt, real shading), cross-referenced with electricity consumption and official ADEME/CRE rates (self-consumption, surplus buyback rate). The result isn't a regional average but a production, self-consumption and ROI estimate specific to each home. The other simulators (energy rating, heat pump, insulation, air conditioning) follow the same logic: checking official data where it exists — the energy rating simulator queries the ADEME database first — then an indicative calculation based on the household's actual profile rather than generic ranges, subsidy amounts included, calculated per the official ANAH scale and reference tax income.

Energy rating (ADEME database)Solar (Google Solar API)Heat pumpInsulationAir conditioning
🤖

Topic-scoped AI assistant

A chatbot trained on a knowledge base dedicated to energy renovation, updated regularly — not a general-purpose LLM. It honestly declines out-of-scope questions rather than inventing an answer, limiting the risk of error on a sensitive regulatory and financial topic.

Vectorized knowledge baseDedicated Python backendReal-time Markdown renderingHonest out-of-scope refusal
📰

GEO-optimized content

Regularly published articles, FAQPage schema, cited sources, direct-answer format designed to be cited by ChatGPT, Perplexity and AI Overviews.

FAQPage schemaDirect-answer formatCited sourcesRegular publishing
🔍

Technical SEO foundations

www/non-www canonicalization, canonical tags deployed site-wide, consistent noindex hygiene across versions — the invisible work that makes everything else indexable.

www/non-www canonicalizationCanonical tagsNoindex hygieneClean indexation

Our method

01

Audit

Analysis of the user journey, existing simulators and SEO/GEO opportunities in the energy renovation sector.

02

Design

Simulator architecture (official data sources, calculation logic), design system, scope of the AI assistant's knowledge base.

03

Deployment

Progressive rollout of simulators, AI assistant and SEO/GEO work. Training included.

04

Continuous optimization

Performance monitoring, regular content publishing, technical SEO adjustments and knowledge base updates.

Stack & tools

FrameworkNext.js 16 (App Router, Server Components, Server Actions)
Backend & dataSupabase (Postgres, Row Level Security, per-simulator tables)
Design systemTailwind CSS + shadcn/ui
AI assistantVectorized knowledge base + dedicated Python backend
Chat renderingReal-time client-side Markdown
DeploymentVPS via PM2 / nginx
IntegrationGoogle Solar API (satellite imagery)

Questions about this project

Want a similar project for your industry?

We'll look at your needs in 30 minutes. Simulators, scoped AI assistant, SEO/GEO — tailored to you.