From founder vision to production SaaS: How Business Conveyor Belt brought AI-powered cash flow planning to market

Business Conveyor Belt (BCB) presented Kiandra with far more than a conventional software development engagement. Rather than implementing a predefined specification, our role was to work alongside the BCB team to investigate, evaluate and resolve a series of significant technical uncertainties associated with building an AI-assisted financial forecasting platform for the trade and construction industry.

From the outset, the engagement required extensive solution architecture investigations, technical design reviews and iterative experimentation to determine whether disparate operational, financial and behavioural data could be transformed into a single, continuously trusted forecasting engine.

Throughout the project, Kiandra's engineering team worked collaboratively with KPI Targets to investigate competing architectural approaches across multiple technical domains, including:

• Cross-platform data integration between Xero, MYOB, Simpro and other operational systems.  
• Canonical data modelling capable of normalising structurally different financial and operational datasets.  
• AI-assisted forecasting using emerging Large Language Model (LLM) technologies.  
• Alternative orchestration frameworks, prompt engineering techniques and tool-calling architectures.  
• Forecast Trust and Dynamic Reconciliation methodologies.  
• Identity, authentication and multi-tenant SaaS architecture.  
• Performance optimisation for real-time forecasting and synchronisation.  
• Secure cloud architecture supporting enterprise-grade scalability.

Rather than progressing through a linear development lifecycle, the project involved repeated cycles of hypothesis, prototype development, evaluation and refinement. Many proposed technical approaches proved unsuitable when tested under real-world conditions, requiring the engineering team to reassess assumptions, redesign components and validate alternative solutions before progressing further.

Examples of this experimental work included:

• Comparing multiple AI providers and orchestration frameworks to determine their suitability for financial reasoning.  
• Investigating token-limit constraints and context-window behaviour when processing large financial datasets.  
• Evaluating alternative approaches to natural-language scenario modelling.  
• Redesigning architectural components following identified limitations in AI reasoning and forecasting behaviour.  
• Investigating synchronisation strategies capable of maintaining forecast integrity across multiple live business systems.  
• Experimentally validating forecasting algorithms using real trade-business datasets and customer scenarios.  

These activities were supported by a structured engineering process incorporating architecture workshops, design reviews, Jira Epics and Stories, sprint planning, prototype demonstrations, technical validation sessions, user acceptance testing, Confluence documentation and iterative product reviews.

Throughout the engagement, Kiandra worked not simply as a software delivery partner but as an experimental engineering partner, contributing to the investigation of technical uncertainty and the generation of new technical knowledge. The engineering team continually evaluated alternative technical approaches, validated hypotheses through working prototypes and refined the platform based on observed outcomes rather than predetermined solutions.

The result is a significantly more sophisticated platform than originally envisaged. The experimental engineering undertaken throughout the project has established technical capabilities that extend well beyond conventional cash flow software, including AI-assisted scenario modelling, continuously trusted forecasting, dynamic reconciliation, intelligent data integration and scalable multi-tenant architecture.

Importantly, this engineering work has positioned Business Conveyor Belt with a meaningful competitive advantage as it prepares for commercial release in late 2026. Rather than relying on conventional reporting or static forecasting, the platform has been built on a foundation of validated technical innovation that combines artificial intelligence, financial forecasting, operational data integration and cloud architecture into a differentiated SaaS offering for the Australian trade and construction sector.

From Kiandra's perspective, this engagement demonstrates the value of collaborative experimental engineering, where architecture investigations, iterative prototyping, technical validation and continuous learning are fundamental to delivering genuinely innovative software products. We are proud to have partnered with the BCB team in resolving these complex engineering challenges and establishing a platform capable of continued innovation and growth.