Background

When we joined up with a lead trading services platform to work on one of their systems, the product was little more than an MVP. A handful of core features existed, but the platform lacked depth, had no mobile app, and was still in a fragile, experimental stage. The opportunity - and challenge - was to transform this early prototype into a robust, client-ready platform that tradespeople could rely on daily.

Challenges

The team faced the dual challenge of building at startup speed while maintaining trust through stability. On one side, new features were urgently needed to attract tradespeople; on the other, every bug threatened retention. Key challenges included:

• Expanding quoting, invoicing, scheduling, and chat functionality.​

• Building integrations with external accounting and communication tools (Xero, QuickBooks, Gmail, WhatsApp).​

• Delivering a consistent experience across web (Next.js) and mobile (React Native).​

• ​Introducing AI-powered data extraction while ensuring reliability, cost efficiency, and resilience to evolving model APIs.​

Approach

Over the span of just five months, Vertex Agility contributed across the full stack, balancing autonomy with delivery speed and quality:

• ​Backend & Integrations: Built core accounting integrations with Xero and QuickBooks, extending Fastify/TypeScript services on GCP with PostgreSQL. Implemented AI-powered pipelines to parse emails and generate leads automatically.​

• ​Frontend & Mobile: Delivered production-ready features in Next.js web and React Native mobile apps, spanning invoicing, scheduling, quoting, and user experience improvements.​

• ​AI & Prompt Engineering: Initially, the platform relied on OpenAI’s ChatGPT for data extraction, but costs quickly became unsustainable. We introduced Google Gemini alongside a test-driven AI development workflow using Vercel’s AI SDK. This enabled prompt tuning, regression-safe changes, and confidence tracking, while cutting inference costs dramatically. Later, we oversaw the seamless upgrade to a newer version of Gemini with zero regressions, proving the resilience of this approach.​

• Ecosystem Integration: We leveraged our earlier work on the client's API to integrate account linking, enabling trades to connect their platform inbox directly to this new service for a seamless cross-platform experience.​

• Quality & Stability: Resolved dozens of critical bugs across schemas, scheduling, and onboarding flows, ensuring smoother onboarding and a more reliable product.​

Results

In just five months, the platform grew from a fragile MVP into being production-ready with steady adoption: 930 active trades, including 331 new trades this month alone. The impact of our work was clear:

• Bug fixes drove retention, bringing tradespeople back with renewed confidence.​
• New features expanded adoption, from quoting and invoicing to full-scale integrations.​
• AI innovation reduced costs and future-proofed the product; switching from ChatGPT to Gemini cut inference spend dramatically while enabling advanced prompt testing and seamless model upgrades.​
• Platform integration strengthened stickiness, with tradespeople able to connect their platform account and inbox directly to the new service.​

The business has seen clear validation - user growth, expanding team investment, and increasing trust in the platform's role as being core for tradespeople.

Want to find out how Vertex Agility can help you transform from MVP to production-ready?

📧 Get in touch now or take a look at our service catalogue to find out more.