Covid-19 – What is The New Normal behaviour?

Original article was published on Artificial Intelligence on Medium

Covid-19 – What is The New Normal behaviour?

How Alpha Reply leverages Data & Machine Learning to help Organisations understand new behaviours

source: agility — PRsolutions

Covid-19 has changed life as we know it. There seems to be a New Normal for every situation. Interactions and behaviours have shifted to adapt to the new circumstances, be it how we work,exercise, parent, shop, socialise and spend our money.

Whilst we still grapple with the new reality, our spending patterns changed overnight, driven by self-isolation and the uncertainty about when –
if ever — things will revert back to normal.

Astound Commerce effectively summed up the shifting consumer behaviours in the chart below.

astound commerce

The change in spending has proven to be an unpredictable challenge for any kind of business: from supermarkets and food suppliers unable to meet demand for certain goods at first, to Financial Institutions struggling to predict “New Normal” customer behaviour in their fight against Financial Crime.

Changing Behaviour and Financial Crime

When it comes to illicit activities, given the lack of adequate regulation to effectively face this unprecedented pandemic, the current environment has made the life of money mules and fraudsters easier: with less cash being used (to avoid spreading the virus) and a shift to online transactions, there has been a proliferation of phishing cases and cyber-frauds at the expense of less-seasoned internet surfers.

Like everyone else, we at Alpha Reply were caught off guard by the pandemic; yet luckily we have been developing a fully adaptive and data-driven behaviour detection system that, by construction, is 100% Covid-proof.

Fusion — Alpha Reply’s AI Solution to changing Behaviours

Our solution uses AI to cluster customers into homogeneous segments, based on their (relative and absolute) financial behaviour.

  • By monitoring customers and comparing their activity to ‘peers’ within their cluster, we can identify what the “normal” behaviour of each segment is.
  • Via Unsupervised Learning techniques, the customer base will be clustered into homogeneous segments in a fully data-driven fashion.
  • Customers receive a ‘normality score’ based on their behaviour within their segment.
  • Customers whose behaviour is consistent across their segment can be cleared; the ones behaving abnormally can be flagged for further investigation (i.e. Requests for Information of Customer Reviews).
  • By scanning customers’ behaviour on a continuous basis, our approach helps Knowing-Your-Customers by letting the data speak for it-self and drastically improves the AML detection processes.

Stay tuned to find more about our AI solution in the coming weeks and get in touch to book a demo.