top of page
  • Writer's pictureMiel Paek and Uliana Ankudinova

Leveraging AI and Machine Learning in Finance: The Future of Sustainable Investment Strategies

Updated: Nov 29, 2023

The integration of artificial intelligence into the realm of fintech heralds a transformative era, most notably in the context of pioneering more sustainable investment strategies. Empirical evidence clearly points to this trend for the near future. A McKinsey analysis, corroborated by data from the Bank of England, indicates that 72% of financial service firms have integrated some form of artificial intelligence into their operations, a percentage significantly eclipsing the 55% adoption rate across other business sectors​​. This widespread adoption of AI in fintech is not merely a testament to the budding technology's versatility, but also to its profound capacity to reshape traditional financial paradigms.

When considering the implications of AI’s potential in the world of fintech, the example of Vistra Corp offers a compelling narrative. Its adoption of AI to enhance the efficiency of its power plants, in partnership with QuantumBlack, a McKinsey company, resulted in a 2% increase in efficiency in only three months, yielding $4.5 million in annual savings and a significant reduction of 340,000 tons of carbon emissions. Given such an example of success, it is safe to conclude that potential applications of AI in fintech - especially in regards to developing sustainable investment strategies - are promising.

AI’s analytical strength can be harnessed towards identifying and predicting the success of companies that choose to implement certain sustainable practices, thereby guiding investment decisions towards environmentally responsible entities. These applications are particularly pertinent in our current financial climate, where ESG criteria are increasingly pivotal in investment considerations. This kind of approach wherein complex, nonlinear relationships in data are meticulously analysed to optimise operational efficiency are best exemplified in the AI-driven ‘Heart Rate Optimiser’ (HRO) model. To shed light on Vistra’s success, we must first investigate how it was able to increase its operational efficiency.

A key advancement in the field of AI in optimising operational efficiency is the artificial neural network (ANN). Operational efficiency, considered among the leading strategies for climate solutions, is a sphere in which the application of ANN is becoming increasingly relevant because of the complexity of information and tasks that are performed on a second-by-second basis. As a result of its fascinating nature and its ability to mimic the human neural structure, ANN, as described below, offers an unparalleled insight into this new world of data processing and demonstrates just how far these new capabilities may extend.

But how exactly does ANN work? ANN is an information-processing paradigm that imitates the operation of the human brain. Its most widely applied model is the MultiLayer Perceptrons MLP network (Szoplik, 2015). To understand the operational process of ANN, however, we must consider how the perceptron works. A single-layer perceptron consists of five parts: the input values, the weights and bias, the sum, the Activation Function (also called the ‘Transfer Function’), and the output.

Given a specific task, a signal is sent to the input layer and to initial, randomly assigned weights. These are combined with the corresponding weights to produce a weighted sum. Subsequently, if the sum exceeds the threshold, the SLP is activated, and output is received. If the output matches the desired output, then weights remain unchanged. Alternatively, the weights are adjusted. A MultiLayer Perceptrons (MLP) network is an ANN consisting of more than a single layer of perceptrons: the ‘Input Layer’, the ‘Hidden Layers’, and the ‘Output Layer’. Still, it replicates the logic of the operation of a single-layer perceptron.

To demonstrate this branch of AI’s promise in this field, the case of Vistra Corp must again be considered. Vistra is a company that specialises in power generation and the provision of electricity, and provides its services across 20 U.S. states. Its segments of operation range from Retail and Corporate to MISO (Midcontinental Independent System). Vistra has pledged to reduce its greenhouse gas (GHG) emissions and has already achieved 72% of its 2030 target. It plans to reduce its emissions by 60% (relative to the 2010 baseline) by 2030, and to reach net-zero emissions by 2050.

One of the steps the company has taken to achieve an emissions reduction of 72% was to increase its operational efficiency. This was facilitated by the AI-powered multilayered neural-network model developed by Vistra in collaboration with QuantumBlack AI by McKinsey, dubbed the Heart Rate Optimiser (HRO). In the energy industry, a power plant’s heart rate measures its thermal efficiency. This value shows how much fuel is required to produce one unit of electricity. Companies therefore aim to optimise their heart rate as a means to attain optimal energy production efficiency levels. But optimising heart rate requires monitoring an immense volume of information from different variables, called set points.

The model developed by Vistra has undergone training on two years’ worth of data. Based on this data, the HRO was able to learn combinations of external and internal decisions that optimise the algorithm and, by consequence, the heart rate. This model was successfully applied to increase the efficiency of electricity production by 2% in the space of three months. This spike in operational efficiency yielded 1.6 million tons of carbon emission reduction per year, the equivalent of eliminating 66,000 cars from use. Emboldened by this success, Vistra is expanding its use of HRO - it has already introduced HRO across 26 plants and achieved an average efficiency enhancement of 1%.

Vistra’s success story is but one of many promising uses of AI in sustainable investment strategies. As AI evolves, its applications within the world of fintech are anticipated to transcend traditional boundaries, enhancing operational efficiencies and propelling the financial industry towards more ethical and sustainable practices. This new combination of AI, energy efficiency, and fintech not only epitomises fast-paced, well-adapted innovation, but also parallels global efforts to achieve climate goals and sustainable development. This in turn reinforces AI’s indispensable role in the forefront of financial evolution.

bottom of page