The client runs a residential construction business (Honka-style timber/concrete homes) that generates leads through a website inquiry form. Every form submission needs to be captured, logged, routed to the sales team, and entered into the CRM — and then nurtured through email marketing depending on how ready the lead actually is to buy.
Previously, every new lead was added to the same email marketing list and received identical campaigns — regardless of whether they were ready to start building immediately, just browsing for information, or had submitted the form by mistake with no real interest. This one-size-fits-all approach meant hot leads weren't getting the urgency-driven follow-up they needed, while cold and unqualified leads were receiving the same sales-heavy messaging as serious buyers — diluting engagement and wasting marketing budget on uninterested contacts.
I built an n8n automation that captures every lead form submission, normalizes the data (name, phone, email, comment — handling inconsistent field labels from the multilingual Wix form), and distributes it across multiple systems in parallel: it logs the lead in Google Sheets, sends an instant Telegram notification to the sales team, and creates a new lead record in Zoho CRM.
The core of the solution is an AI-powered lead qualifier: an LLM agent (via OpenRouter) reads the client's comment and classifies the lead into one of three segments — hot (ready to build now), warm (interested but not immediate), or cold (no real interest / accidental submission). Based on this classification, the workflow automatically adds the contact to the matching Klaviyo email list, so each segment receives messaging tailored to their actual buying intent — urgent, consultative, or re-engagement campaigns respectively.
1. Discovery & process analysis
Reviewed how leads flowed from the website form to the sales team and the email platform, and where intent was being lost. Found the core gap: every lead — hot, warm, or accidental — landed in the same campaign. Defined three actionable segments with the client and what messaging each should get.
2. Solution design
Designed an n8n flow that captures each submission, normalizes the messy multilingual Wix data, fans it out to four systems in parallel, and adds an AI qualification step that decides which email track each lead enters.
3. Implementation
Built the pipeline and integrated the systems: