Free 30-minute AI & automation audit with a senior engineer. You keep the notes. Book yours

Use case · Construction & retail development

Managing multi-site store build-outs: projects, RFIs, change orders and punch lists

A construction project management platform for brands rolling out stores, with Gantt schedules, RFIs, change orders, punch lists, approvals and fine-grained access.

Client details are kept confidential. This use case describes the problem and what we built.

The brief

A construction-technology company serves brands that build and refit physical stores across many locations. Its customers needed one system to plan and track those projects, from early scheduling through RFIs and change orders to the final punch list.

We engineered the full platform with them: the Spring Boot backend, the React web application and the automated pipelines that ship each release to every customer’s own environment.

The challenge

What needed solving

Many locations, many projects

Brands opening and refitting stores run dozens of projects at once across markets and sites. Status, budgets and approvals were scattered across email and spreadsheets.

Construction workflows are formal

RFIs, change orders with revisions and approvals, punch lists and deliverables each follow their own rules and need a full audit trail.

Different people see different things

Owners, project managers, contractors and consultants needed access to specific locations and projects only, with single sign-on for corporate users.

Each customer wanted its own instance

Customers expected their data and configuration to stay separate, with their own branding and terminology.

The solution

What we built

  1. Audited Spring Boot backend

    A JHipster-based Spring Boot service on PostgreSQL with Liquibase models brands, markets, locations, sites, projects, phases, deliverables, RFIs, change orders, cost categories and custom fields. Hibernate Envers audits every change.

  2. Entity-level permissions and SSO

    A custom ACL layer with Spring Security method annotations controls access per entity. Users sign in with Microsoft Entra ID (Azure AD) single sign-on, JWT or API keys for integrations.

  3. Change orders with history

    Change orders carry items, revisions, reasons, status transitions, rejection reasons and a timeline of events, so every cost change can be traced.

  4. Search, documents and background jobs

    Elasticsearch through Hibernate Search powers global full-text search. Attachments are stored on Amazon S3 with derivative files produced through a document conversion API, and JobRunr runs scheduled jobs.

  5. React application with Gantt scheduling

    A React 18 and TypeScript front end with Redux Toolkit, Ant Design and Bryntum Gantt covers dashboards, nested location, site and project views, executive reports, rich-text editing and maps.

  6. Per-customer deployments

    GitHub Actions workflows build each version once and deploy it to separate staging, demo and production instances per customer, with database-backed feature flags and per-customer branding and translations.

The outcome

What changed

One place for every store project

Teams see all locations, sites and projects together, from schedule to change orders to closeout.

Every decision is traceable

Audited entities and change-order timelines show who changed what and when.

Faster customer onboarding

Shared code with per-customer deployments, branding and feature flags lets new customers go live without forking the product.

Technology used

  • Java
  • Spring Boot
  • JHipster
  • PostgreSQL
  • Liquibase
  • Hibernate Envers
  • Elasticsearch
  • Hibernate Search
  • JobRunr
  • Amazon S3
  • Microsoft Entra ID
  • React
  • TypeScript
  • Redux Toolkit
  • Ant Design
  • Bryntum Gantt
  • Lexical
  • Google Maps
  • Jest
  • Cypress
  • Docker
  • GitHub Actions

Have a similar problem to solve?

Share a few lines about your project. A senior engineer — not a salesperson — replies within one business day.

Free AI audit Let’s talk