Case Studies
AI-Powered Internal Tools for Small Businesses
We set up Anthropic's Claude as an internal tool for businesses, connecting it to their existing data, documentation, and other sources. This empowers employees to automate workflows well beyond basic chatbot use.
A typical example is a data detective bot that produces visualisations and insights on demand, turning hours of manual analysis into a simple conversation.
This has been delivered across multiple client engagements.
Executive Performance Review & Anomaly Detection System
Designed and delivered a recurring executive-level performance review for a client's platform, incorporating all funnel metrics. Built an AI-powered anomaly detection system that not only flags anomalies but automatically explains their likely causes.
This replaced static dashboards with a self-driven, exception-based approach: surfacing the biggest surprises in metric changes before leadership was aware of them.
Impact: Empowered decision-makers to act on metric changes quickly. Reduced time-to-react on metric anomalies by days to weeks.
North Star Metric Strategy
Led the development of a client's 2026 growth north star metric, replacing their existing metric with one that was more measurable, more relevant to the business, and less prone to noise.
Impact: Gave the company a clearer, more actionable measure of growth to align their teams around.
Customer Lifetime Value Prediction Collapse Investigation
Investigated a sudden and alarming drop in customer lifetime value predictions that was top-of-mind for the senior leadership team. Traced the root cause to a hidden segment concentration issue, then produced a series of follow-up analyses and a summary document for leadership.
Impact: Resolved a metrics crisis that had been blocking executive decision-making for weeks. Restored confidence in the forecasting pipeline and surfaced a concentration risk affecting ~20% of projected revenue.
Fraud Signal Discovery & Data Integrity Overhaul
Uncovered a long-standing fraud data issue that had been distorting a client's understanding of their own performance. Ramped up to full fluency on the company's fraud systems in just two days through rapid interviews with senior engineers and management, then produced example models and documentation to serve as the basis for new fraud systems.
Impact: Resolved a systemic data integrity issue that had been distorting key business metrics for over a year, leading to more accurate reporting across the entire funnel.
Macroeconomic Sensitivity Analysis
Led a highly cross-departmental initiative to model and communicate the impact of macroeconomic softening on the business. Built a dedicated reporting tool that synthesised signals from across the organisation into a single view for leadership.
Impact: Gave the executive team a clear, data-driven picture of their exposure to macroeconomic shifts, enabling proactive planning.
Measuring AI Chatbot Impact
A client had recently implemented an AI chatbot and needed robust measurement to determine whether it was helping or hindering their business. After implementing a measurement system over a few weeks, we found the chatbot was actually hurting performance due to poor responses and insufficient training data.
Impact: Identified the problem and guided subsequent improvements. Customer satisfaction scores have since increased by an average of 10%.