Business Post Group applies GenAI to modernize legacy datasets and unlock new digital revenue opportunities.
Business Post Group faces a common GenAI-era challenge for legacy media companies. Vast historical datasets exist, but much of the data is fragmented and outdated. A newly acquired property dataset held significant value, yet legacy Irish addressing formats limited usability. Manual matching was slow, error-prone, and impossible to scale. GenAI offered a way to transform static archives into searchable, commercial-ready assets.
The primary challenge involved aligning pre-2015 property records with Ireland’s modern Eircode system. Addresses lacked consistency and reflected outdated legal and planning information. Business Post partnered with Devoteam to solve this using AWS-native GenAI services. A sandbox environment was created using Anthropic Claude 3.5 Sonnet via Amazon Bedrock. Extensive data cleaning and normalization enriched records with Irish government open datasets.
GenAI became critical when deterministic matching methods reached their limits. Devoteam implemented a two-stage approach combining geospatial algorithms and GenAI reasoning. Latitude and longitude filtering narrowed possible matches. Where ambiguity remained, GenAI analyzed satellite imagery and contextual planning data. This allowed the system to reason spatially and contextually, surpassing traditional rule-based systems.
The proof of concept processed over 55,000 locations across County Carlow. The GenAI-enhanced pipeline achieved an 85% match rate, exceeding expectations. This success converts historic journalism data into a reusable digital resource. The benefits include faster search, new data services, and expanded monetization opportunities. Business Post now plans nationwide rollout and deeper GenAI adoption through AWS migration. The project signals a strategic shift from content-only publishing to data-driven digital growth.
