Case Study: BT spots 52% of mobile tower failures a month in advance with Kortical's AI platform

BT & 3 Network Infrastructure Provider MBNL Predictive Maintenance

BT operates some 22,000 mobile network towers and, together with infrastructure partner MBNL, faced the challenge that engineers cannot economically visit every site frequently enough to prevent intermittent equipment failures. Kortical was engaged to provide a predictive maintenance solution using its AI/ML platform to ingest diverse datasets (support tickets, manufacturer data, power meters, signal and text data) and identify equipment likely to fail before outages occur.

Kortical ran a 4‑hour workshop and built models in a 6‑week proof‑of‑value, using its platform to generate and evaluate over 50,000 machine learning solutions and produce production‑ready, explainable models. Validated by holding back six months of data, Kortical’s models predicted about 52% of failures more than a month in advance, prompting a live pilot to prove results in production and enable cost‑effective reduction of emergency site fixes.

BT