Corlytics develops FCA’s intelligent handbook

Posted: 28 September 2017 | | No comments yet

Corlytics has helped develop the world’s first intelligent regulatory handbook for the UK’s Financial Conduct Authority.

Corlytics develops FCA’s intelligent handbook

The FCA handbook contains regulatory obligations which are binding, as well as manuals which contain provisions. The structure was previously an online legal text book. Corlytics has worked with the team at the FCA to develop a central, common taxonomy to categorise all regulations. Having put this in place, the existing material in the handbook can be tagged and machine read. This allows for a fully searchable handbook to be developed.

Democratising the handbook

Previously the FCA handbook didn’t have built-in search capability by theme. It had grown overtime as an indexed document. Now it has been totally restructured and is searchable in multiple directions; by teams, by business line, by activity and by theme.

The FCA wanted to democratise the handbook, to ensure it is accessible to all.

John Byrne, CEO of Corlytics explains, “We have put a metadata structure, much like that used by Google, in place, thereby transforming the handbook from a comprehensive legal index to a highly accessible tool for all users.”

Metadata delivers accurate and consistent search

By using a metadata structure, it has made adding an intelligent regulatory advisor as a front end to the regulatory handbook possible. The addition of front-end artificial intelligence needs consistent and accurate data. It is just not possible to develop this from indexing technology, which is neither consistent or accurate. Automation needs metadata so that the text is fully machine readable.

What we have learned from medical science

Like law, medical research has lots of opinionated papers and lots of non-aligned studies. By analysing the text using machine learning analytics, oncology research has made great strides.

Byrne continues, “At Corlytics we have moved into the same building as a lot of specialist medical data scientists to better understand what they do. Using ‘trained models’ we are able to teach the models how to understand and interpret the data.

“To best do this you need subject experts who can programme and understand analytics, working alongside data scientists. Lawyers in our case who can code. We have swapped the oncologists with regulatory lawyers. Their training makes for consistent and accurate analytics.”

Tagging the entire FCA handbook

3,000 metadata tags have been added to the original core infrastructure to make all technology projects around the democratisation of the handbook possible.

The teams have gone to different sections of the handbook and machine learnt them. Then using a combination of regulatory lawyers and data scientists, they have auto tagged the handbook.

Send this to a friend