Extraction to Coherence
Project description
This project aims to extract and measure the coherence between company statements such as mission, vision, values, goals, and strategy. Using NLP, statements are converted into sentence embeddings and compared with cosine similarity to assess semantic alignment. The results are presented in excel coherence matrices to support expert evaluation of strategic consistency.
Context
The project presented in this proposal focuses on developing a GEA (General
Enterprise Architecture) framework with the support of AI . GEA is an enterprise
architecture methodology that analyzes the coherence within an organization by
examining core elements such as mission, vision, core values, strategy, and goals.
A GEA framework makes this coherence explicit, measurable, and manageable,
thereby providing managers with a powerful tool to improve decision-making
and organizational performance.
Results
The project successfully developed a proof of concept pipeline for extracting, classifying and comparing corporate strategy statements (mission, vision, core values, goals and strategies) from websites and PDF documents. The accuracy remains limited due to the relatively small and imbalanced labeled dataset and the high variability in how companies formulate their statements.
About the project group
The group contains 3 dedicated members coming from the business, software and media profile. 16 hours have been dedicated to the project per person per week and split the task
