Corporate Website for a Construction Company with a Cost Calculator and AI Lead Qualification
We designed and launched a managed corporate website that systematically captures leads. The calculator helps visitors determine a realistic cost range, while AI generates a short summary and highlights gaps, so managers can process leads from ads and organic traffic faster.
About the Project
A full-featured, sales-focused website for a construction company. It includes service pages for different types of work and properties, a portfolio with case studies, a process page, trust blocks explaining guarantees, approach, and frequently asked questions, a content section for organic traffic, a cost calculator, and a request form. A user can arrive from an ad on a specific service, quickly understand the scope, estimate a budget in the calculator, and submit a request with parameters. The sales team then immediately receives a structured brief with an AI summary for quick qualification and the first call.
About the Customer
Max Gallardo from the USA is an investor and analyst in high-risk tech with experience in dev outsourcing and a systematic approach to UX and processes. He was looking for a team that could deliver a product quickly and clearly, and handle many clarifications and exacting requirements for details, copy, and form logic. The domain is high-ticket construction services, where the site has to build trust for cold traffic while also giving hot traffic a short path to a request. At the same time, it had to filter out "raw inquiries" as much as possible, so the sales team isn't overloaded.
Sales from ads and organic traffic
We needed a structure where paid traffic lands directly in the right context by service, property type, city, and segment, while organic traffic has a clear page hierarchy to support growth.
Trust in high-ticket sales
Construction is a risk for the client, so the site needed proof and clear boundaries: scope of work, process, guarantees, case studies, common questions, and expectations for timelines and budget. All of this had to be built into the UX without turning the site into a wall of text.
Cost calculator
The preliminary estimate had to give a reasonable range and explain the cost drivers without false promises, since those lead to conflicts.
Data collection forms
We had to find the right balance: a minimum of essential fields, with soft hints and validation, so the team doesn't receive meaningless requests.
AI qualification
The AI had to help the manager by creating a summary, identifying risk tags, and suggesting clarifying questions, but never inventing a budget, area, or type of work. It needed strict limits: work only with what the user entered.
Content scalability
A full website means constant change: new services, case studies, FAQs, articles, and updated offers. We needed a manageable content model so the site could keep growing.
Proposed Solution
We built a full website on headless architecture. Next.js delivers a fast frontend and landing pages for different types of work and properties. Directus manages content and page structure (services, case studies, process, FAQ, trust) without developer involvement. A Node.js API layer handles the lead business logic.
Information architecture built around real traffic
We structured the pages so that every type of query has a matching landing page: services (with work subtypes), pages for segments and property types, case studies with filters, the work process, FAQ, contacts, and a content section for organic traffic.
Custom UI/UX
We designed the site as a high-ticket product: readable, with clear accents, proper form states, and clear CTAs. Key information goes on the first screens: services, work sequence, client outcome, and next steps.
Preliminary estimate calculator
We added a calculator that works on multiple pages, not just the home page, and delivers a result with detailed explanations.
Requests as briefs
We separated the handling of estimate requests, consultation requests, and requests that already include a project or drawings. Each type generates its own set of fields, so the manager doesn't have to ask the obvious again.
AI qualification for sales
We designed an AI assistant that creates a 5–7 line summary, assigns risk tags, and suggests follow-up questions.
Next.js site focused on speed and managed growth
We introduced a component structure and optimized the first screen, media, and critical resources. We also set up event analytics for service views, calculator starts and completions, form submissions, and contact clicks.
We started by breaking down the client's request into specific sales tasks. We identified the types of work to feature on the site, the most common requests, and the manager's steps leading up to the first call. We agreed on the minimum set of brief fields: property, area, type of work, location, timeline, budget range, and additional conditions. We also determined what the site should filter out or flag immediately: a missing budget, missing timeline, or an overly broad request. In parallel, we agreed on how the calculator would work: it would give only a range and an explanation of the cost drivers, with no guaranteed figures, to avoid creating false expectations.
Results
The site stopped being just a business card and became a system that consistently brings in quality leads and saves time for the sales team.
77%
of requests arrive with key parameters filled in
72 seconds
time users take to complete the calculator in a typical scenario
5 lines
length of the AI summary for each inquiry
2–4 risk tags
per lead, simplifying prioritization of inbound ad leads
3–5 clarifying questions
suggested by AI for "raw" requests
90+
mobile PageSpeed score at release for key pages