KestrelMeters RAG

TheGenieLab delivered an AI-powered knowledge layer designed to improve both product discovery and technical support through a structured, iterative implementation process. The project began with the creation of a dedicated workspace, Trello project management, and a staging environment that allowed the AI assistant to be configured, tested, and refined before its production launch.

To build a reliable knowledge base, Kestrel's technical manuals and supporting documentation were collected, consolidated, compressed, and converted into formats suitable for AI-powered retrieval. TheGenieLab then analysed this content to structure the knowledge base, enabling the assistant to provide accurate, product-specific answers using relevant technical information.

The implementation also included a review of available AI assistant platforms, including Chatbase and AskTimmy, to determine the most suitable approach for the project. Prompt configurations were carefully refined to ensure the assistant delivered consistent, accurate, and helpful responses. In parallel, existing FAQ content was optimised to improve answer quality and better address the questions customers are most likely to ask during the purchasing process.

Following configuration, testing, and review, the AI assistant was successfully launched. After go-live, TheGenieLab continued to refine the implementation by reviewing collection URL behaviour to create cleaner product discovery paths and reduce unnecessary navigation complexity.

Implementation Timeline

The project commenced in February 2026 with the initial setup phase, during which the project workspace was established, Trello project management was configured, and a staging environment was prepared for Kestrel AU.

Throughout March 2026, TheGenieLab focused on preparing the AI knowledge base by collecting, consolidating, compressing, and analysing Kestrel's technical manuals and support documentation. During this phase, the team also evaluated AI assistant platforms and configured prompt settings to establish a reliable foundation for customer support.

In late March 2026, the project moved into optimisation, with refinements made to both the FAQ content and AI prompts to improve the relevance, accuracy, and quality of customer-facing responses.

The AI assistant was finalised and deployed to production in April 2026 after successful staging review and testing.

During May 2026, post-launch refinements were carried out, including improvements to collection URL handling to support a cleaner product discovery experience and simplify customer navigation across the website.

A Smarter Way for Customers to Research, Compare, and Choose Kestrel Meters

The AI assistant transforms the way customers research and select Kestrel meters by making product discovery faster, more intuitive, and more informative. Instead of manually comparing multiple products and accessories, visitors can ask natural language questions and receive guided recommendations based on their specific needs and intended applications.

Built on a curated knowledge base, the assistant provides technical answers using information drawn from Kestrel's manuals, FAQs, comparison data, and product documentation. This enables customers to quickly access reliable, context-aware guidance while reducing the time and effort required to research specialist products. Whether customers are selecting equipment for ballistics, fire weather, agriculture, or heat stress monitoring, the assistant helps move them from uncertainty to a more confident purchasing decision.

The solution also improves support scalability by resolving many common product and technical enquiries before customers need to contact the support team directly. Carefully configured AI prompts and optimised FAQs ensure responses remain aligned with Kestrel AU's approved product information, providing consistent and trustworthy guidance. To minimise deployment risk, the implementation followed a structured staging, testing, optimisation, and production launch process before going live.

What We Delivered

The Kestrel AU AI RAG project was delivered across three core implementation pillars. The first focused on Knowledge Engineering, including manual retrieval, PDF consolidation, document compression, file conversion, and technical content analysis to create a structured AI-ready knowledge base.

The second pillar covered AI Configuration, where TheGenieLab reviewed AI assistant platforms, configured prompt behaviour, optimised FAQ content, and refined support responses to deliver accurate, customer-focused interactions.

The final pillar addressed Shopify Integration Readiness, which included staging environment setup, production launch preparation, go-live support, and post-launch reviews of collection URL behaviour to improve product discovery and overall user navigation.