Case Study

50% Faster Design. 50% Faster Development. One AI-Powered FinOps Platform.

Australian Infrastructure Company

Challenge

A major Australian infrastructure and construction Tier 1 contractor sought to develop a proof of concept for an enterprise commercial control and FinOps platform. Powered by SpreadJS and Workflow.io, the target system was designed to provide operational and executive teams with real-time visibility into project financials, budgets, actual costs, and forecasting via spreadsheet-driven workflows.

The strategic objective was twofold: evaluate how much more productive an engineering team could become using modern AI-accelerated tooling (including Claude Code and Antigravity), and determine whether a bespoke custom platform could be delivered faster and more cost-effectively than heavily customised, high-licensing SaaS alternatives. Fabric Group was engaged to lead and deliver the POC using its AI-native delivery model.

The organisation faced complex technical requirements and critical commercial decisions:

  • Real-Time Financial Complexity: Delivering dynamic financial modelling, multi-layered cost tracking, and forecasting across large-scale infrastructure projects required responsive spreadsheet-like usability combined with robust back-end workflow automation.
  • SaaS Customisation Overhead: Off-the-shelf SaaS financial platforms demanded heavy tailoring, rigid workflows, and recurring licensing fees that scaled poorly across large enterprise project portfolios.
  • Aggressive Delivery Benchmarks: The enterprise needed to validate whether a bespoke solution could compete with commercial platforms on both speed-to-delivery and total cost of ownership.
  • Maintaining Engineering Rigor at Velocity: Accelerating delivery could not come at the expense of code quality, architecture integrity, or test coverage in a mission-critical financial domain.
No items found.
PRocess

Fabric deployed an AI-assisted product development lifecycle across design, engineering, and automated testing over three weekly iterations:


1. AI-Assisted Rapid Prototyping & Design

  • AI-assisted interface prototyping of key financial views and spreadsheet-driven layouts.
  • Iteratively validated interactive designs directly with business user champions to ensure operational fit.
  • Converted validated concepts directly into development-ready functional specifications, cutting design interpretation lag.

2. Agentic Engineering & Structured Implementation

  • Deployed Claude Code and Antigravity to accelerate development across 3 weekly delivery sprints.
  • Applied engineer-led structured prompts and AI-governed agentic checklists to drive each user story from backlog to working software.
  • Automated routine boilerplate and complex data-grid integration, keeping senior engineers focused on core business logic and architecture.

3. Automated Quality & Governance Gates

  • Shipped every feature with automated unit tests created and verified as part of the implementation flow.
  • Structured agentic verification across design, code, and test execution to validate each increment before review, ensuring high delivery confidence and zero quality regressions.
No items found.
Results

The three-week engagement proved that AI-accelerated engineering delivers bespoke enterprise software at SaaS-beating speed and cost:

  • 50% Reduction in Design Cycles: AI-assisted prototyping and user champion validation produced development-ready specs in half the time of traditional design processes.
  • 50% Cut in Development Time: Claude Code and Antigravity halved implementation timelines through assisted code generation, automated test creation, and structured prompt workflows.
  • Elevated Code Quality & Delivery Confidence: Agentic checklists and mandatory automated test suites across all user stories strengthened code reliability and simplified validation.
  • Custom Build Cost-Competitive with SaaS: The accelerated delivery model drove down custom development costs significantly, establishing a custom-built FinOps platform as a far more viable and attractive commercial path than high-cost SaaS alternatives with ongoing per-seat licensing.
No items found.
No items found.

Related Content

No items found.
No items found.