Natural Language Processing and Cognitive Networks Identify UK Insurers’ Trends in Investor Day Transcripts

The ability to spot key ideas, trends, and relationships between them in documents is financial services, such as banks insurers. Identifying patterns across vast amounts of domain-specific reports crucial for devising efficient targeted supervisory plans, subsequently allocating limited resources where most needed. Today, insurance planning primarily relies on quantitative metrics based numerical data (e.g., solvency returns). purpose this work assess whether Natural Language Processing (NLP) cognitive networks can highlight events relevance regulators that supervise the market, replacing human coding information with automatic text analysis. To aim, introduces a dataset NIDT=829 investor transcripts from Bloomberg explores/tunes 3 NLP techniques: (1) keyword extraction enhanced by network analysis; (2) valence/sentiment (3) topic modelling. Results analysis, enriched term frequency-inverse document frequency scores semantic framing through networks, could detect system like cyber-attacks or COVID-19 pandemic. Cognitive were found related specific transitions: frame “climate” grew size +538% 2018 2020 outlined an increased awareness agents insurers expressed towards climate change. A lexicon-based sentiment analysis achieved Pearson’s correlation ρ=0.16 (p<0.001,N=829) levels daily share prices. Although relatively weak, finding indicates jargon insightful support risk supervision. Topic modelling considered less amenable supervision, because lack results’ stability intrinsic difficulty interpret patterns. We discuss how these methods complement existing tools supporting effective oversight market.