Domain Analysis
One of the key methods bridge_ci use to extract insights with OSINT, Alternative or internal data (when available) is what we call a “domain analysis.” The approach uses natural language processing and auto-machine learning to extract trends on any topic, company, stock, person, or organization from millions of structured and unstructured data points - before making assumptions about the person, company, topic, etc. This enables firms to produce analytical, qualitative, and quantitative insights and establish a strategy and resources that are more likely to be successful and self-aware of both opportunities and risks.
Domain Analysis Process
Terms and datasets associated with a given domain, trend, or topic are chosen. This could be a global issue, such as COVID, a trend like “working from home,” a regulatory issue, such as the EU’s Digital Single Market, a currency like the Euro, a social topic such as Black Lives Matter, a brand, such as Goldman Sachs or stock such as Apple. It is also possible to combine multiple terms and find connections between them.
A machine crawls OSINT and alternative datasets to retrieve all of the associated documents, articles, and social media mentions focused on the domain.
Machine Intelligence platforms then process the unstructured and structured data to find connections and patterns within each domain.
These connections are explored, which allows firms to extract insights that are most relevant to their firm.
Based on the insights gathered, strategies and asset allocations can be set and then implemented.
Then, outcomes are measured, and more refined strategies can be executed.
The process can be repeated indefinitely.
The technique above can be applied to most business functions, ranging from PE, risk management, M&A, VC diligence deals, product development, procurement and supply chain management, FP&A, geopolitical and security analysis, HR, communications and marketing, public and government affairs, and in nearly any industry context.
Benefits
People see all the information and data on a given topic or issue, not just a few news articles or data points.
Personal and biased information feedback loops are filtered out at the beginning.
Systematic machine processing detects and flags biases and misinformation.
“Soft” trends, such as political, regulatory, or social risk, are measured analytically rather than according to “gut feelings,” and then these are compared with structured datasets (e.g., stock prices, sales, profits, etc.) to present a holistic picture of that domain. This is particularly valuable in politically and emotionally charged environments.
Stress-tests strategic hypotheses in minutes or hours, not days or weeks, before making massive investments in time, assets, and resources.