Instead, it is the context that qualifies how useful and purposeful such intelligence is. Artificial Intelligence (AI) is no different and, if anything, is more prone to such criticism.
Gary might be an expert at 80’s pop (strictly 1984 to 1989), but he also has the ability to apply himself (albeit badly) towards other subjects. AI doesn’t have this capability, it will always be restricted by a multitude of constraints ranging from its source inputs, bias and its programmatic decision making.
So, what does this mean to financial services and the intermediary sector?
AI is being touted as the predominant power behind robo-advisers and other digitally enhanced advice tools. It presents a shining solution to the difficult and complex nature of collecting vast amounts of data and consistently creating a set of decisions.
But it is inherently difficult to create, maintain and reliably generate consistent results.
Add the evolving and fragmented nature of financial services and we may have the automated advice tools that work for today but these can be horrendously misaligned in a month’s time.
For me, this poses the question – are the emerging capabilities of AI best employed in simply replacing advice, regardless of whether this is cheaper, more efficient or more accessible?
At 360 Dot Net, our investigations have brought us to the conclusion that AI is powerful, it is capable of automating complex tasks, but these tasks must be extremely narrow and well defined.
With a marketing hat and using anonymised sale data, we can begin to accurately match products to specific consumer demographics. This can then be used to determine the effectiveness of marketing those products to other consumers with a higher degree of targeting.
An example is that we can process a client bank and highlight product opportunities depending on where in the accumulation or decumulation journey that client may be, but be far more nuanced as we consider the individual’s circumstances.
Supporting good advice
Another area is to reverse the mechanism of automated product selection and use this as a compliance function to determine whether the products advised upon and actions performed match up against an automated selection and historical provenance.
Under the established banner that compliance is a function to provide consistent quality outcomes, an organisation may only be able to sample a specific percentage of cases, through AI it is possible to process every case in minutes, flagging up recommended file checks.
AI might get the sensationalist headlines of taking over jobs and even the world. But the reality is that narrow, routine and often complex tasks can be excellent candidates for automation, especially in the legislative or administrative domains.
This means that advisers can concentrate on advising, administrators can more effectively fulfil cases, and principals enjoy the benefits of technology that works for their business rather than the other way.