Revolutionary change tends to be a short, sharp affair. So by that criterion at least the emergence of Artificial Intelligence (AI) can best be described as ‘evolutionary’ rather than ‘revolutionary’.
If it can also be loosely characterised as technology enabling a computer to ‘think/ act’ in a more ‘human’ way, there is little doubt AI destined to gain significant traction over the coming years.
In their 2019 ‘CIO Agenda Survey’, consultants Gartner noted the proportion of organisations now deploying AI had risen from 4% to 14% vs. their previous survey.
The report gathered data from more than 3,000 CIO respondents in 89 countries – the industries they represent accounting for approximately $15 trillion in revenue/public-sector budgets and $284 billion in IT spending.
Not simply a narrative
Despite the relatively low take-up (so far) of AI a separate Gartner report – polling 400 organisations – indicated that 27% of finance departments expect to deploy some form of AI or machine learning in their department by 2020. And more than half of those questioned expect to be using predictive analytics in the same period. Predictive analytics is the practice of extracting information from existing data sets in order to determine patterns and predict future outcomes and trends.
Much like the proverbial chicken and egg, however, digital – as a general concept – has assumed pre-eminence to date – its rise has reflecting, in no small part, the changing nature of engagement companies are having with their increasingly demanding customers. The very same customers who are now more likely to want to communicate via digital channels as technology itself advances.
For companies though, AI isn’t simply a narrative about pandering to changing consumer behaviour; indeed there are equally relevant issues to be addressed closer to home.
AI and fraud detection
In the area of fraud detection, for example, machine learning-based solutions have slowly but surely been supplanting the historic-based rules approach, ostensibly because the former is more intuitive.
Rules-based approaches have traditionally focused on evident signals. Or to put it more simply, observing either large transactions, or transactions from unusual locations. While these systems will employ algorithms performing several different scenarios for fraud detection they are less intuitive than machine learning solutions because any additional scenarios need to be manually inputted.
If the rules-based approach only scratches the surface, machine learning solutions can not only employ algorithms to process larger datasets (with more variables) more quickly, as well as requiring less manual work, they can also more readily detect subtle behaviour patterns that may, or may not, be symptomatic of fraud.
Apart from being able to better detect fraudulent activity, other obvious benefits (already here) include the assessment of risk – especially relevant in the insurance sector, for example, where fraudulent claims are a very real issue.
Indeed, by using a specific Artificial Intelligence Decision Algorithm (AIDA), previous underwriting experience can be gauged to determine possible red flags now. AIDA also has practical benefits more generally when it comes to credit scoring loans, for example. Not only can loans be signed-off far more quickly they can be tailored to the individual’s needs, which in turn should reduce risk to the company.
Going forward, however, process automation is one of the key drivers of AI in financial organisations. Already, cognitive process automation – where AI systems are able to perform ever more complex automation processes, such as the rapid scanning of documents – have markedly reduced the length of time for such tasks. Especially useful for companies where there may be thousands of documents to scan through.
AI-fintech nexus
Meanwhile, fintech – aided by technological innovation – is one of the fastest growing spaces in the global ecosystem as mobile phone devices are increasingly used to perform transactions or related actions.
Hence, mobile usage is supporting the fintech revolution, which is providing its own innovative, secure, and accessible solutions for financial services and products.
Elsewhere, AI also has practical relevance when it comes to virtual assistants. Here, banks fintechs and other financial institutions can make the customer’s experience more user friendly by guiding them through their services and products – the virtual assistant answering questions the customer asks. Crucially, it allows financial institutions to improve their product targeting, as well as helping reduce costs and increasing operational efficiency.
Longer-term, the AI-fintech nexus, though still relatively small, is expected to rapidly develop – a recent report published by Industry Research estimating the size of the AI in fintech market in 2019 at $6.84 billion; but expected to grow to $26.92 billion by 2024, giving a CAGR of 31.5% over the forecast period.
Unsurprisingly, AI, along with Big Data and Predictive Analysis, is providing financial institutions with increasing reach, while at the same reducing human error in banking procedures that may prove costly in terms of not assessing risk properly.
While AI is becoming increasingly important to many organisations – AI platform as a service or AI cloud services being obvious examples – machine learning is already well established.
Machine learning is now commonplace in the global financial ecosystem with algorithmic applications being built to calibrate financial portfolios to the goals and risk tolerances of customers.
Will robots really steal our jobs?
PwC, in its report, “Will robots really steal our jobs?” An international analysis of the potential long-term impact of automation” identifies three overlapping waves we can expect to see through the 2030’s.
The first wave, known as the Algorithm Wave, which focuses on the automation of simple computational tasks and analysis of structured data in areas like finance, information and communications, is already well developed.
As is the Augmentation Wave, with it emphasis on the automation of repeatable tasks such as filling in forms, communicating, exchanging information through dynamic technological support, and statistical analysis of unstructured data in semi-controlled environments, such as aerial drones and robots in warehouses. This is likely to reach maturity in the 2020s, according to PwC.
Less clearly defined (and reflecting the slower uptake of AI itself) is the Autonomy wave, focusing on the automation of physical labour and manual dexterity, as well as problem solving in dynamic real-world situations (manufacturing? transport?) that requires responsive actions. This is forecast to hit full maturity in the 2030s.
Pessimists charge that the development of AI and automation over the longer term will be detrimental to human jobs creation as people are substituted by machines.
More likely is that AI will lead to a net jobs creation as humans (dependent upon the industrial sector) are redeployed elsewhere to be re-trained and re-skilled as industries themselves are recalibrated.
The World Economic Forum raised this point in its 2018 ‘Future of Jobs’ report when it estimated that AI globally would create a net 58 million new jobs by 2022.
If its forecast proves to be remotely correct AI is something to be embraced, not feared.