Speaking to Fintech Business, Advice RegTech co-founder Fraser Jack said that RegTech’s ability to identify possible compliance breaches through natural word processing and other methods relies on those methods’ ability to interpret the text it is analysing.
Mr Jack explained that data exists in two formats; structured and unstructured, and that the information included in SOAs and other significant documents was routinely unstructured, due to the use of freeform text.
Unstructured data, he said, must be structured before it can be properly interpreted and used, and this requires RegTech solutions to identify what is being said in the documents.
“When it comes to data you have structured data and unstructured data formats, and most of the time when you’re writing an email or a statement of advice its unstructured data because it’s free text,” he said.
“That’s where the big problem lies – a word that could be a positive could also be a negative if it has the word ‘not’ in front of it, so you have to try to create a dictionary or a road map of all that information.”
Mr Jack said dealer groups and industry associations should be looking to work together to develop this taxonomy, ideally in an “open-source format like Wikipedia” so relevant companies are able to develop usable services for advisers.
“There’s certainly a bit of a gap in that area when it comes to the RegTech space – that’s the area that needs the most work and I think from a regulation point of view, we need to create that piece of work as an industry and not as individuals,” he said.
“Then all the companies that are involved can work together using this one ontology – this is something that if we want to be professional we need to regulate ourselves and not wait for that regulation to be handed down to us.”