Keeping it simple: taking a pragmatic approach to deploying AI in financial services

The transformative nature of AI has been well-documented in the last two years but realising this vision still seems like a distant dream for many financial services providers. The potential appears so vast, so far-reaching, that some firms are getting caught in an inertia trap where they simply don’t know where to start. Add to this the pressure from stakeholders to deliver a rapid return from early pilots and the risk of failure becomes almost overwhelming.

The solution lies in focusing on a simpler, more practical approach to deliver small gains from initial applications that can be used to demonstrate the value of AI. From here, chosen projects can embrace more ambitious goals and build greater momentum for further investment.

Last week, we joined banks and wealth management firms at the 2024 Private Banking DACH Conference in Zürich, Switzerland where we were a Gold Sponsor. We also joined a fireside chat on “Alleviating C-Suite Executives’ Pressure” where our speaker was Puneet Saraswat, SVP Global Sales & Solutions for Linguistic AI & Regulated Industries. 

Puneet touched on some examples of how AI is currently being applied to specific use cases in financial services. These focused on three areas: 

Global growth: the challenge for many of our clients across industries is how to scale globally – launch new products, expand into new markets and maximise the potential buying audience – this is especially true in financial services. We work with clients in asset management for instance who want to accelerate their fund launches internationally. Automation, and the increased prevalence of AI to support that with increased scale and accuracy, is becoming fundamental to achieving and maintaining a competitive edge in a crowded market. 

Fund management: AI is supporting the shift from actively-managed to passive funds through the use of sophisticated algorithms. But active fund managers are increasingly levelling the playing field by leveraging AI for enhanced research, particularly in analysing large volumes of unstructured data from company commentaries and narratives. 

Sales and research: AI is being used in low-risk, high-gain areas like improving sales productivity, enabling self-service with multilingual virtual assistants and improving the client experience with personalisation.

The session moved onto how creativity and human expertise are being impacted by the development of AI. Puneet disputed the common misapprehension that AI is a threat to human expertise as being too simplistic. He referred to the finite human resources in areas like translation and how AI can actually help humans to increase their productivity. Also, that the creative element in areas like product development, marketing and new investment themes, rely heavily on human ingenuity. 

Indeed, the rapid progress of AI is being supported by a massive human effort in training AI models to ensure that the data on which they are based is optimised to avoid pitfalls like bias, ambiguity and inaccuracy.

So the big picture remains as large as ever but the journey to achieving it depends on a more restrained methodology. Puneet concluded by summarising some critical steps to get started: 

1. Take a practical approach to moving from experimentation to productisation. Of course, firms want their projects to be monetised as quickly as possible to demonstrate ROI but perhaps the key is to focus on modest marginal gains that prove the concept first. 
2. Look for ways to increase automation, particularly of routine tasks and high-volume workloads, rather than completely replacing existing processes that may upset the human and cultural equilibrium in your organisation. 
3. Optimise your use of human resources to balance costs but also to realise the productivity gains offered by AI. 

You can get started on this more practical approach by talking to us about our linguistic AI capabilities as well as optimising your AI models with our TrainAI data services.