The problem on the office side
Inquiries arrive through web forms, e-mail, phone notes and word of mouth. A renovation request, a commercial fit-out and a subcontractor RFQ all land in the same inbox, and whoever opens it first decides — often days later — what happens next.
In construction the first qualified response usually wins the conversation. Estimating capacity is scarce: every hour an estimator spends on an unqualified or out-of-area request is an hour not spent on a biddable project.
What the system does
A structured intake form captures the essentials: project type, location, budget range, timeline and a free-text description. The moment it is submitted, an AI step classifies the inquiry (new build, renovation, commercial, maintenance), scores it against your criteria — region, project size, trade fit — and writes a three-sentence summary with a suggested next action.
The result lands as a contact, a deal and a note in your CRM (HubSpot or similar) and as a row in a shared spreadsheet for the site managers. High-score project inquiries are routed to estimating; maintenance requests go to the service team; out-of-scope inquiries get a polite automatic reply.
Nothing about your public website needs to change — the system sits behind your existing forms and mailboxes.
What this looks like in practice
A regional builder handling around 80 inquiries a month typically sees three things change. First, response time drops from days to minutes, because qualification no longer waits for a person to read the inbox. Second, estimators only see inquiries that pass the score threshold. Third, management gets a live overview — every inquiry, its score and its status — in one sheet, instead of asking around.
We documented this pattern in detail in our construction case article, including the scoring criteria and the exact tool chain (web form, Make, OpenAI, Google Sheets, HubSpot).
Implementation
A working pipeline for one intake channel is typically live within two to three weeks: one workshop to define your qualification criteria, a build phase, and a calibration week where the AI's scores are checked against your team's judgement before anything is fully automated.
The system runs on tools you can see and own — no proprietary platform lock-in. If you already use HubSpot, Google Workspace or Make, we build on what is there.
Frequently asked questions
Do we need to replace our website or CRM?
No. The system connects to your existing website forms and writes into the CRM you already use. If you have no CRM, a structured Google Sheet is a workable starting point.
How long does setup take?
A first working pipeline — form to scored CRM entry — is typically live in two to three weeks, including a calibration phase where your team reviews the AI's scoring before it runs unattended.
What happens to inquiries the AI is unsure about?
They are flagged for human review instead of being auto-routed. The system is designed to take over the routine 80%, not to make judgement calls on edge cases.
Is this compliant with GDPR?
Yes. Inquiry data is processed under a data processing agreement, stored in your own systems (CRM, spreadsheet), and the intake form includes the consent and privacy notices your legal setup requires.
What does it cost?
Setup is a fixed project price depending on channels and integrations; running costs are the tool subscriptions (typically under €100/month at this volume) plus optional support. We give a concrete quote after a short process audit.