Skip to content

03Case study

Under development

Il Perfetto Srl

AI-assisted opportunity discovery and automated estimating.

Sector
Construction · architecture · building services
Where
Italia

A system under construction for a construction company: every day it analyses potential work opportunities, combines job type, area, estimated hours and location, and helps produce a first automated estimate (preventivo) for the customer.

Illustration of an opportunity-to-estimate workflow

  1. 01 · Market

    A daily stream of possible jobs: renovations, facades, interiors, technical services.

  2. 02 · Opportunity

    Each opportunity is captured with its source, timing and first known parameters.

  3. 03 · AI analysis

    AI classifies and ranks the opportunities against what the company does well and where.

  4. 04 · Job data

    Type of work, area in square meters, estimated hours, location and constraints become structured fields.

  5. 05 · Estimate

    A first preventivo is assembled automatically after the appointment — reviewed by a person before it reaches the customer.

01 · Problem

In construction the estimate is the sale. Opportunities arrive from many sources, each appointment costs hours, and pricing depends on the same handful of variables recomputed by hand every time. Slow estimates lose jobs.

02 · System

Opportunities are collected daily and ranked by AI against the company's profile. For each promising job, structured data — type of work, square meters, location, operational parameters — feeds an estimating model that produces a first preventivo. Humans review, adjust and send.

03 · Technology

Python pipelines for daily opportunity analysis and LLM-based classification, a Next.js back office, PostgreSQL for a pricing knowledge base that improves with every accepted estimate, and PDF generation for customer-ready documents.

04 · Impact

Target: a first estimate within the same day of the appointment, consistent pricing across the team, and a growing dataset of what the company actually charges and wins. Under development.

  • Daily

    Opportunity analysis

    automated

  • 5

    Variables per estimate

  • 1 day

    Target time to first estimate

    target

Capabilities

  • Daily opportunity discovery
  • AI ranking against company profile
  • Structured job data capture
  • Automated first estimate (preventivo)
  • Human review before sending
  • Pricing knowledge base — in progress

Stack

  • Python
  • LLM classification
  • Next.js
  • PostgreSQL
  • PDF generation

Want a similar system? Let's talk.

Let's talk