Adopting AI in Industry Leading Organizations
This is a review of episode 3 of the MMC original podcast “Beyond The Hype: Artificial Intelligence”. A chat with Barclays’ Steven Roberts (Strategic Transformation Director) and Rob Otter (Head of Machine Learning), discussing how today’s leading firms survive, and thrive, in the age of AI.
Host, David Kelnar, asks the pressing questions surrounding the world of AI in a business as large as Barclays, a bank serving half the UK.
From how executives can better understand AI, to the benefits, challenges and future of A.I. Check out the key learnings and insights!
Who fuels the adoption of A.I within organizations?
The change for AI is being driven from the top-down and bottom-up within management in large businesses. Corporate executives can’t help but be aware of AI, as the noise surrounding the topic continues to grow. Whilst they read about it everywhere, whether or not they have a clue about it is an entirely different question. The growth of AI has been likened to the early adoption of the internet within organizations. The corporate executives know AI exists, know it has great potential, but don’t yet understand the process of unlocking this potential.
Within the finance industry, a big catalyst for the growth of AI & ML in banking was the FX LIBOR manipulation and unauthorised trading events. The adoption of AI by industry players in the area of investment banking then accelerated the growth of AI & ML initiatives. This artificial intelligence adoption allowed these banks to look at unusual trades, trading volumes and patterns to predict erratic behaviour.
How Adopting AI Unfolds in a Company as Large as Barclays
In a company as data-rich as Barclays, AI & ML are a top priority amongst corporate executives. In particular, adoption starts from the standardisation of data. Machine learning comes in to provide actionable business insights. The (successful) adoption of AI & ML here facilitates fraud prevention and detection, revenue generation, credit card impairment analysis, and prediction of when people may default on credit card payments.
Adopting AI in a large bank like Barclays commonly means cloud AI is not an option. Multidimensional sets of data require highly optimized AI which can take hours to compute — meaning time is more than precious when expensive individuals such as data scientists, ML experts and graph experts. Expensive hardware AI dealing with multidimensional datasets is one of many barriers to AI adoption.
Adopting AI is no walk in the park: AI adoption challenges
AI models are not immune to error. This means an incorrect model in a live system related to credit cards, for example, could be incorrectly declining transactions. Tracing evidence through a neural network on why transactions are declining presents major challenges. This challenge of evidencing AI has no easy solution as of yet and needs to be cracked soon. Some firms currently get past this challenge by adding human analysis to triage AI recommendations.
It is fair to say that in many ways AI is mankind first, for both risk and opportunity. Company AI adoption can even be compared to an entire firm quickly learning to drive – it could drive great change or it could crash and burn. That’s why understanding the risks and benefits of AI is of paramount importance.
How to implement AI in businesses is one of the key challenges in adopting AI. correctly and avoiding the crash and burn scenario But much like the ignorant attitudes towards X-Rays in centuries past, the benefit of AI must not blind sight the need to weigh up the risk.
AI unlocks new levels of personalisation for customers
Firms in the banking industry that understand the potential of AI towards customisation have brought banking back to its roots — a personalised banking experience. Think hyper-personalisation in banking!
In pre-industrial revolution medical practices, the affluent members of society would consume medicine crafted on a case by case basis depending on the ailments. Fast forward to the era of standardisation and economies of scale and the desire from companies to supply as near to a one size fits all solution has become the norm. This is where the power of AI reinvigorates a personalised approach. Instead of receiving banking products/services categorised into say 5 different risk models, AI gives customers bespoke fit for purpose solutions. Like stepping back in time and meeting with a branch manager without the hassle!
AI’s future in the banking industry shows huge potential, but not for a lack of risk. As for the future impact of adopting AI from a broader perspective, the experts from Barclays share some of their predictions.
- AI will have the most profound impact on the transport industry
- More jobs will be destroyed than created with AI
- Artificial intelligence will never have rights