Praemium's machine learning takes platform accuracy to a new level
Using machine learning across a range of data sets, Praemium has been able to identify transactions that may have been incorrectly entered or categorised by administrators.
Praemium’s Chief Technology Officer, Adam Pointon notes “When managing large volumes of data, human errors happen. Through machine learning, we have been able to identify which transactions may be incorrect and predict the correct classification. For example, a buy transaction might be incorrectly entered as a withdrawal, or income as a deposit. These errors could provide incorrect portfolio performance information or have tax implications for investors.”
“This functionality allows for errors to be detected at scale and rectified quickly and is already being used successfully with several of Praemium’s institutional clients.” Pointon adds.
The functionality expands upon Praemium’s existing machine learning capability Insights, launched in 2019, that can provide highly accurate predictive analytics that a client is demonstrating behaviours that indicate they need advice.
“Praemium’s non-custodial solution is recognised as the market leader and we continue to enhance our technology with these exciting innovations.” Pointon continues.
Recent research undertaken by Praemium with Investment Trends showed that almost 60% of advisers are managing non-custodial client assets off-platform. Typically, these assets are managed manually via spreadsheets, consuming two extra hours of adviser resource per client.
Praemium’s Chief Commercial Officer Mat Walker also commented: “Praemium’s non-custodial solution has $140bn in assets under administration and offers advisers and wealth managers the benefit of managing both custodial and non-custodial assets on a single platform. Our research indicates that advisers are feeling the burden of administering these assets and our technology not only does this efficiently but also more accurately. We also offer the option to remove the administration burden completely by outsourcing to Praemium’s Administration Service who also utilise this functionality for large volume data processing.”