AI-powered competitive intelligence

Changing what it means to be a competitive intelligence professional

Changing what it means to be a competitive intelligence professional

Competitive intelligence (CI) professionals tend to rely on search engines and databases with keyword-based search criteria to find the information they need. This leads to path dependence: the results of an initial search informs the next step, the outcome of which subsequently affects the continued research. When the search is completed, the researcher may have missed important aspects that were outside the search path. A priory assumptions have a large effect on the outcome of the study and unexpected insights are unlikely.


Dcipher Analytics offers an alternative approach. Rather than relying on previous knowledge and initial assumptions, the researcher starts with a large body of relevant information. This could mean hundreds of thousands of news articles or social media posts containing certain keywords of interest. It could also take the form of financial reports, academic articles, patents, or any other structured or unstructured body of information of relevance to the research question.


Applying natural language processing and unsupervised machine learning, Dcipher Analytics identifies topics and clusters in the data. This gives the CI professional an overview of the entire landscape and the possibility to drill down into specific topics of interest for further research. By switching between overview and depth, the researcher is able to view the forest and the threes at the same time. Unexpected insights emerge bottom-up.


Deep learning-based techniques for document similarity search further augment the approach by freeing the researcher from reliance on static search criteria. Searches that match documents based on individual keywords tend to miss important documents and capture many unimportant ones; language is too complex to be described by simple rules. Similarity search, on the other hand, finds relevant documents even if they are not using the exact keywords used by the researcher.


To try out the AI-powered research paradigm, sign up for a free trial.