Cognitive computing uses and learns from data, enabling it to mimic the way that people use their own intelligence in making decisions by weighing multiple options and gathering evidence. It thus enables companies to hypothesise about various options and score them to determine which one has the best outcome or highest profitability, says Harper.
These systems are now able to take unstructured data such as Tweets or pdfs, make sense of them to make decisions and then use the results to evaluate options so they can learn to make better decisions in the future.
IBM’s ‘cognitive engine’’ for the front office, dubbed Watson, can use natural language processing so people can talk to machines and receive responses such as information about bond prices or advice about investment portfolios. Cognitive computing democratises knowledge, Harper believes, because companies can tap into the knowledge of traders and advisors.
Cognitive computing can also be used to understand customers better by accessing and analysing information about their sentiment, social media usage, networks, business and personality to draw conclusions about them. This knowledge can enable an advisor or salesperson to build a customer profile, understand the customer’s propensity to buy and personalise the interaction, which can increase sales productivity tremendously.
Singapore’s DBS Bank, for example, uses this information and then pumps in economic data, as well as in-house product reports, so it can provide next-best-investment actions for the advisor to offer to clients. As a result, these are more insightful than they would otherwise have been.
One step ahead
Harper also highlights the fact that in the middle office, cognitive computing is adept at making more informed decisions and outthinking the competition. In the past, companies used data and made decisions based on factors such as credit scores. Incorporating structured as well as unstructured data enables better underwriting and other higher profit decisions.
In areas such as exception handling for loan processing, Watson can quickly review thousands of cases that require exception processing and get them back into the pipeline. Banks and insurance companies have also used Watson to assess precedents, figure out what types of customers have defaulted in the past and make a decision without the huge amounts of manual checking that were required in the past.
A UK bank recently used the Watson system to aggregate data so that it could understand compliance requirements and the implications of decisions about their customers. The automated analytics enabled the firm to reduce the number of people involved from 15 to three by eliminating the manual checking. Banks in Japan, which typically can have 10,000 or more people in their call centres, have similarly freed up resources by automating and augmenting call centre processes.
Setting up cognitive computing
To help clients implement cognitive computing, one or two IBM staff offer advice in understanding and determining where the system will be most useful. They then train a small team over a period of about three months and identify the best use cases. After determining where cognitive computing can add the most value, they then design and build the system and then use pairs of decisions to train Watson over the course of about six months.
One example offered by Harper: a bank in Italy initially wanted IBM to focus on wealth management, but shifted its focus towards lending after finding that wealth added only 1% of the value they expected.
The first implementations are usually small use cases, such as queries in a call centre, which are designed as a pilot to show senior management how it works and how it can be applied across all functions. In parallel, the team expands the strategy and determines how to apply cognitive computing across the firm. The first use case is usually stand-alone and only subsequently is the system subsequently integrated into customer databases.
Watson for transaction banking
While IBM has not yet implemented cognitive computing in transaction banking, Harper regards it as “a huge opportunity.” Trade growth leads to a need to service physical supply chains, from manufacturers to warehouses. Banks can use cognitive computing in their financial supply chain for contextual management of data and in relationship platforms for the alignment of physical supply to the financial supply chain.
Whereas cognitive computing used to be just a technology and ideas for using it have tended to be cost-focused, he believes that it is now a business platform with use cases that can drive shareholder value.
Many banks around the world, large and small, continue to experience major security failures. Biometric systems such as pay-by-selfie, iris scanners and vein pattern authentication can help.
The implementation date of Europe's revised Markets in Financial Instruments Directive, aka MiFID II, is fast approaching. Yet evidence suggests that awareness about the impact of Brexit on MiFID II is, at best, only patchy and there are some alarming misconceptions.
Banks might feel justified in victim blaming when fraud occurs, but it does little for customer confidence.
Politicians have united in urging the Reserve Bank of Australia to lend its backing to the digital currency by officially recognising it.