January 11, 2023 | Tim Ward
Estimated reading time: 4 minutes
The Art of Search
We are all used to using Google for scouring the Internet. So much so that “to Google” has become a verb that we use in our everyday language. A good search engine should be fast, accurate and easy to use. It should return relevant results in response to a user’s query, and it should be able to handle a wide variety of search terms and types of information. In addition, a good search engine can cope with misspellings, synonyms and typing errors to provide the best possible results for the user.
When a user types in a common word or phrase, there may be thousands, millions or even billions of results. Almost useless, unless these results are sorted in a way that puts the most likely or relevant results at the top of that list. Google has spent years perfecting and tweaking this secret algorithm to do just that.
How CUBE masters the art of search
Here at CUBE, we also think a lot about how to ensure our users get the information they require when they need it with minimum effort. Searching across regulatory information presents some unique challenges. First of all, there is the sheer volume of information. Terabytes are stored across thousands of different sources presented in many different languages. Each source arranges the information differently and of course, the information is always changing as new rules become available and existing rules are modified to adapt to the latest environment and trends.
There is so much work to do before even the first search can take place. CUBE’s unique platform constantly monitors the world’s regulatory internet, and captures relevant information, normalizing and enriching it on the way. Only when all data sits in a central location, has useful metadata and has been classified and indexed is there a chance of a performant and effective search.
Users can search across the regulatory internet quickly and easily by typing a word or phrase. Results are displayed and can be sorted to suit. But there is still a potentially vast number of articles and regulatory rules that may be returned. This is when RegPlatform comes into its own. Because CUBE has been able to use machine learning and natural language processing to accurately classify every article and rule section – we can introduce the concepts of profiling and filtering.
Profiling and filtering
Profiling allows a firm to specify its unique characteristics including the countries they operate within, the products and services they offer and the types of regulatory content they are most interested in. Now when a user searches, results can be pre-filtered to match their pre-configured profile – vastly increasing the chances of a relevant hit.
Filtering takes things a step further. Users can set up pre-defined filters to only search within a specific jurisdiction or on a particular topic. CUBE has developed a unique financial ontology that conveniently describes the topics, products, and compliance concepts that a user may be most interested in.
Using profiling, filtering and searching in an intelligent way can be the difference between spending hours and days reading versus hitting the right content the first time and covering what you need to know in a few minutes. Some compliance completionists strive to read every issuance about the topics or products that their firm is engaged in – a task only made possible through accurate profiling.
There are other interesting search-related use cases as well. Sometimes you don’t quite know what you are looking for. More exploration than search. CUBE can help here as well. We automatically pick out obligations – the things you have to or should be doing, and these are available cross-jurisdiction and cross-topic. We provide citations to link to other regulatory issuances, and a timeline to show what has changed and colleagues may have also linked articles and documents together in ways that you might find interesting.
The future of search
We are also very excited about the future of Search. Using artificial intelligence and machine learning will allow us to make searching even more efficient and intuitive, associating articles together based on semantic similarity or recommending articles based on similar usage patterns. We are all about using technology to make the day-to-day lives of regulatory analysts easier. And there is no better way than focusing on something that we all do tens or even hundreds of times a day.
Make no mistake, searching is a skill. Finding the right term, setting up an accurate profile, and configuring the correct pre-filters is an art that will save days’ worth of effort over a year if done correctly. But at the end of the day, the search results will only be as good as the underlying data. This is why CUBE is obsessed with the coverage, quality, and classification of every piece of regulatory content we touch.
Keep ahead of emerging regulations by speaking to CUBE.