$150M in new annual revenue.
Speed-to-market up 30%.

AI stayed stuck in pilot mode while teams fought for the same customers.

Marketing and Product were targeting the same customers with volume-based tactics, producing inconsistent experiences, offer cannibalisation and low return on marketing spend. Previous AI implementations were siloed and built on over-complicated academic practices or inaccurate methods, so models never made it past the experimental phase.

Without translatable evidence of AI’s commercial value, the executive team couldn’t justify further investment in data and AI capability. Technical delivery was fragmented across consumer divisions, with specialists working in isolation and no central “ways of working” that could scale across Mobile, Fixed and Loyalty at the same time.

Three constraints were non-negotiable:

• AI models stuck in proof-of-concept, with no path to production
• Marketing and Product cannibalising the same offers and customers
• No shared ways of working across Mobile, Fixed and Loyalty

One mission to take AI from data to execution.

Our Fractional Head of AI built the business case for a flagship “Lighthouse AI Mission” and took it to the executive team — laying out the structural changes needed to move AI from data to execution, the shifts required in data management, and the revenue uplift on offer. That secured leadership buy-in across the functions needed to give the mission a clear structure and remit.

ML models began ranking customers daily across multiple dimensions: likelihood to convert, churn risk, and expected lifetime value across products. Once the models proved their value in proof-of-concept, ML-based targeting replaced broad demographic and volume-based segments as the default approach, with auto-segments closely monitored to protect the customer experience.

A live revenue dashboard tracked AI-selected strategies against existing rules-based campaigns on conversion, revenue per customer and cost per acquisition — reviewed daily for day-to-day optimisation and fortnightly with executives for bigger calls. Campaigns where AI won were scaled automatically; the rest were descaled or paused. Underneath it all, a cross-functional team of 20 data science specialists was unified under one standard way of building and monitoring AI, cutting time spent on data engineering and accelerating model production by up to 30%.

A mission that turned pilots into permanent revenue.

The Lighthouse AI Mission converted experimental models into dependable, production-grade capability. Tracking AI-selected strategies against legacy rules-based campaigns on conversion, revenue, lifetime value and cost-per-acquisition proved AI’s superior ROI, cut wasted spend on low-intent customers, and gave Telstra a repeatable, industrialised path for its next AI initiatives.