Chris Royles, EMEA Field CTO at Cloudera
Amongst all the emerging technologies of the last two decades, artificial intelligence (AI) has tipped the hype scale, causing organisations to rethink their entire digital strategy. In the financial services industry, the projected numbers on the value AI can drive are staggering. According to a recent McKinsey & Co. article, “across the global banking sector, [Generative AI] could add between $200 billion and $340 billion in value annually, or 2.8 to 4.7 per cent of total industry revenues.”
Knowing where to invest, safely
With the rapid adoption of AI and Gen AI across the industry, innovation can now be driven across various domains. Traditional machine learning (ML) models can enhance risk management, credit scoring, anti-money laundering efforts and process automation. Gen AI has the power to unlock more targeted, personalised customer experiences via virtual assistants, content creation, advanced risk and compliance analysis, and even data-driven trading strategies.
Many of the largest financial institutions in the world are already harnessing the power of AI. For example, JPMorgan Chase uses AI for personalised virtual assistants and ML models for risk management. BlackRock is using AI to automatically generate research reports and investment summaries, whilst Deloitte employs AI for risk, compliance, and analysis while also using ML models for fraud detection.
These organisations have pinpointed how high-value, high-volume tasks can benefit from automation, personalisation and rapid analysis enabled by ML, AI and Gen AI models. But it doesn’t stop there. LLMs offer a high range of essentially ‘free’ capabilities for no additional cost. For example, they can help organisations to understand complex regulations and can seamlessly interact in different languages and modalities. Plus, voice technology is a major asset. While it can be expensive to integrate, operationally it can be very cost efficient while delivering high returns on investment. Its ability to enhance customer service and support call agents is undeniable.
Like all technologies, AI is not without potential systematic risks, including biases and – particularly important in finance, care must be taken when integrating with personal and sensitive data. Understanding how these risks can spread across a financial system is vital to the safe and effective use of AI.
Balancing risk and reward
With the wealth of sensitive and personal information involved in the financial sector, data privacy and security demand rigorous protection measures. This includes robust encryption, stringent access controls and advanced anonymisation techniques to ensure financial data remains secure.
AI decision-making processes must also be transparent and explainable to ensure regulatory compliance standards are met. With transparency, organisations can foster trust with customers by explaining and verifying AI-driven decisions.
An awareness of biases and errors in training data is also essential to prevent AI delivering incorrect insights. Proactive human oversight is key to avoiding this. AI systems must be designed to detect and address biases – particularly in sensitive or protected features – so they can identify shifts in data patterns, before they impact results. Mitigating these biases will ensure that AI systems provide fair and accurate outcomes, which is critical for maintaining the integrity of financial services.
FS institutions need to create a safe and robust environment for AI to operate in, avoiding key pitfalls of the new technology that can lead to massive economic loss and customer distrust if AI offers inaccurate responses. AI must be trained on a complete set of proprietary data to avoid inaccuracies and hallucinations, which means FS organisations must tie data from disparate environments together.
These challenges can be overcome by deploying a modern data architecture underpinned by a unified data platform. This is a layer of frameworks and systems that allow organisations to store, analyse, and grant data access seamlessly. This enables organisations to be more agile and efficient through stronger data management
Without modern data architecture, agility and flexibility get lost, making scalability simply impossible, placing organisations at a competitive disadvantage as they look to deploy AI. A framework that prioritises security and governance to safely manage and analyse vast volumes of data, alongside measures such as complete ML/AI flows, model registry and methods for monitoring and profiling data is an absolute must. Only then can businesses be confident that their AI is accurate and unbiased.
Securing success with Gen AI and AI
There are significant benefits to AI, and the FS industry would be amiss to lose the chance of a competitive edge that the technology provides. The benefits are abundant, from aiding research, to enhancing customer experience and making transactions safer. But there are multiple challenges associated when deploying AI, and organisations must ensure the highest levels of data privacy and security as well as data lineage and governance that will ensure compliance with regulations.
Preparation and responsible adoption are key to driving success through AI and ensuring confidence across the financial sector. Those financial services organisations that are able to harness the power of AI will find it to be transformational, giving them a competitive advantage. With many large organisations in the sector already utilising it, the rest of the industry must follow suit.
