November 8, 2023
Content series about the biggest trends shaping private finance
Over the past decades, many areas in finance have increasingly embraced technology, whether that is trading moving to electronic, consumer FinTech transforming the consumer banking landscape, or enhanced data collection and processing capabilities supercharging the pricing of complex instruments and default forecasting, to name just a few. There is a clear link between the monetised benefits and technology’s contribution to these advancements.
A notable exception to this trend is the private finance industry (incl. M&A). In this content series we will introduce the new but now widely used technologies and their potential applications to private finance, as well as dig deeper into reasons of resistance and bespoke challenges in the context of this sector.
Despite it being one of the largest sub-industries of finance, private finance (and M&A) has had less of an immediate need to utilise cutting edge technology. Being largely a relationship driven business, top level service is the main selling point, and limited publicly available datasets and idiosyncratic transactions mean that statistical modelling and big data crunching often yield little benefit. However, we are now entering a period in which technology has evolved to a state where it can add real value to private finance, including M&A processes. Here, we will discuss some of the key mega trends that we see turning the industry on its head over the coming years.
Given that no discussion about technology written in 2023 would be complete without highlighting how generative AI models could benefit the industry, we start our content series with it.
Subscribe for future deep dives in:
Given the large number of unstructured text in the forms of files, emails, Q&A and notes that result from a typical private finance deal, let alone the number contracts that need to be written and reviewed, there are numerous potential use cases for generative AI. We use M&A as a specific example, although many observations apply to other types of private finance transactions as well.
Before discussing the specific use cases however, it is important to touch upon a critical aspect inherent in all generative AI (and general machine learning) models - that of trust. Any system has to gain the trust of its user, and the more complex the system, the harder it is for a human to trust it. If it is explainable, and therefore clearly understandable - for example a deterministic financial model in Excel - a user can walk through each step and make an informed decision on how much to believe. As you introduce randomness into a system (which is inherent in any machine learning algorithm), this becomes much more difficult.
Results have to be reviewed and tested continuously and only after a user has become comfortable with the consistency and performance of the model will it be able to sit within a workflow. For numeric based models (e.g. statistical derivative pricing models or other historically back testable models) it is generally possible to perform analysis in a safe environment to quantify the accuracy of a model, and thus, help a user become comfortable. With text based models that have the potential to add the most value in M&A, it is much harder, instead requiring the user to slowly get more comfortable with them over time.
The private finance world is exceptionally fast paced and often times it feels to participants that the time cost of adopting a new technology will outweigh the benefits, however with the new generation of AI models incorporating human feedback into the training process, the onboarding of a new technology system is much more akin to onboarding a new analyst. The technology can continuously evolve, learning from humans with little change in current workflows, thus reducing the overhead of manually updating models, tooling and process in order for it to add value.
As previously mentioned, private finance is primarily a relationship driven business. With this in mind, unsurprisingly, we see generative AI and natural language processing tools within the M&A use case being used to assist, but not remove the need for human oversight. Using a computer to generate an initial draft of work, or suggest how to complete a task leaves professionals more time to focus on value add tasks that are more creative and complex.
The workflow of a computer creating the initial draft, and a human reviewing (and adjusting as appropriate) has become more familiar across many industries, for example, the use of co-pilots / ChatGPT in software engineering. Private finance should be no different. The review process additionally provides the natural interface for integrating human feedback into the training of complex tools, and therefore improving the models in a seamless flow. This will enable faster adoption of gen AI tooling.
We see a number of concrete use cases for generative AI in private finance, for example:
All of these, and many more use cases, could provide material time saving benefits to professionals, allowing them to focus on executing deals at a higher quantity and quality than ever before.
We are at the very beginning of the development and integration of generative models into professional workflows, with the environment evolving everyday, and we see there being potential for an immense impact on private finance workflows.
Do you want to be the first one to read the latest monthly kicker.cloud insights? Click here to subscribe our mailing list.