Online money transfer services such as PayPal and Transferwise are using machine learning for fraud prevention, by analysing millions of transactions between buyers and sellers to identify fraudulent transactions or potential cases of money laundering.
Additionally, AI and machine learning are already heavily embedded in customer service across major banks and financial institutions, helping to cut waiting times and improve customer satisfaction by assisting users with basic queries as well as managing the enquiry flow between customer support representatives. According to Accenture, AI in its current state can handle up to 80% of customer engagements.
The list of AI use cases grows daily. However, traditional sectors such as banking and finance are far behind in the automation race. The main reasons are dependency on legacy IT systems and data security concerns. Yet, the competition brought by cloud-native fintech startups has put finance and banking sectors under urgent pressure to speed up the adoption of new technologies.
On the whole, AI deployment is a complex process for businesses to adopt and IT decision makers need to consider key obstacles and opportunities for successful adoption.
Optimise the costs of AI adoption
The cost of developing AI and machine learning models in-house are often too high, and the price is usually driven by the computing power required, as well as having enough data in place to build and train an advanced model. According to UK’s Digital Catapult Centre, the average cost of a single training for a machine learning system is over £10,000.
Still, businesses are quickly adopting AI applications despite high expenses, and having the right development infrastructure is essential.
A large number of organisations are using cloud and open-source platforms to deploy AI capabilities. By operating and storing data in the cloud, businesses are optimising costs by using on-demand payment models and customisations, subsequently investing only in capabilities required.
Additionally, open-source models have limited or no costs and require less input from in-house developers. The nature of open-source software means businesses can build fit for purpose AI applications using publicly available software development kits.
Bridge the skills gap
Despite the high adoption prices, usage of computing power and data has been on a rise in the past years. As a result, developers have to be able to differentiate between a variety of automated and intelligent services to successfully deploy them in their organisations. This goes beyond finding an AI solution that solves the business problem, as developers need to identify the right DevOps partner to provide the right consultation and support during the initial deployment.
While developing the automation strategy, there are two ways for businesses to address the internal skill gap: invest in AI experts or train existing staff. According to recent reports, with millions of AI focused roles available, there are only 300,000 professionals able to fill them. Skilled developers are highly sought after and find themselves in a favourable negotiating position where they’re able to increase the price of their services. More often than not, companies find themselves investing in this area, even with increased cost, due to the business demand.
The variety of cloud and open-source resources helps businesses to battle the skills gap. Initiatives such as Fast.ai delivers free AI training models to upskill on natural language learning or image classification. Furthermore, training is delivered on an ongoing basis, ensuring developers are equipped with the latest information in-house, to enable progress and innovation within the business.
Choose the right automation strategy
Whilst AI adoption will be the main focus of IT teams, for successful transformation a whole organisation will require an attitude shift, as automation will likely affect the roles of many employees.
At first, automation will affect repetitive and strenuous tasks, driving productivity and employee satisfaction. This can be achieved by augmentation - a collaboration between AI and humans, which helps to train AI algorithms to enable automation. Businesses who implemented this technology reported dramatic results, such as 28% higher overall performance, 31% growth in financial performance and 38% growth in employee engagement.
The key question for organisations is not if they should implement AI, but how. Automation of challenging business processes is proven to deliver a positive return of investment and growth in productivity, allowing companies to grow their competitive proposition.
Finally, fast-growing open source technology offers C-level IT decision makers greater flexibility and agility for AI adoption. By providing access to development, adoption and future innovation tools, open-source technologies ensure long-term business success achieved by AI deployment. With reduced risks associated with complexities in implementation, CIOs and CTOs will now focus on applying the right mix of technologies for achieving their business goals.