Will ChatGPT replace the fund selectors? Fund buyers speak out!

With over 130,000 public investment funds available globally, selecting the right fund can be a daunting task for both investors and the professionals guiding them. The question on many minds today: could generative AI (GenAI) take over this process going forward?

GenAI is adept at processing large volumes of data in a systematic and consistent manner, which is a fundamental challenge in fund selection. If machines can replace humans in this role, it could reduce the cost of implementing portfolios. These savings could then be passed on to investors in the form of lower fees, which could potentially compound to create greater retirement wealth.

But is it really that simple?

GenAI: A skilled assistant, not a senior one

GenAI is already being used by fund selectors in different parts of their process. Proper due diligence covers investment, operational, corporate and ESG aspects. This can mean analysing hundreds of pages of documents. Summarising these documents and helping to focus on what really matters and will make a difference is something at which GenAI excels. Fund selectors have also had great success using GenAI to challenge the fund managers they analyse, and to challenge themselves. Which areas of analysis have they overlooked? What should they spend more time on? What questions should they ask the fund managers to improve their analysis? Fund selectors also need to report on their findings, and GenAI has proven helpful in this area too. GenAI has therefore added significant value in terms of analysis, due diligence and reporting. The selection of which funds to analyse in depth was largely addressed already using quantitative techniques.

A need for continuous education to be a more efficient and better fund selector

Fund selectors agree that GenAI helps them to be more productive and accurate in their analysis by enabling them to focus their time and energy on what will make a difference.

However, this doesn't come for free. The quality of the answers provided by GenAI depends on the way you ask the question. Learning how to prompt a GenAI system is a science in itself, and mastering it is paramount to truly leveraging such systems.

GenAI can be subject to hallucinations, which can lead to incorrect results (sometimes providing answers that please the person who asked the question). The way to protect against this is to prompt the system correctly. This is not a topic for IT people: they lack the investment expertise to recognise hallucinations and prompting expertise to avoid them. This is a topic for the fund selectors themselves.

Everyone is equal with regards to learning how to prompt: old/young, small/big organisations, because two years ago, this tool was just not available. Therefore, training is a must, all the more so given the incredible and permanent progress being made. There will need to be an ongoing commitment to education.

Gen AI is a faithful, hard-working co-pilot, not a captain

Can Gen AI perform as well as a human fund selector? Firstly, let’s not forget that these are Large Language Models (LLMs), and language is what they deal with best. They do not perform equally well in quant analysis, for example.

Experience and intuition are also key in the selection process. A simple sentence, tone of voice, or even a look in a certain direction during a due diligence meeting can provide a fund selector with much more information than thousands of pages of reports and files. Planes have been fitted with autopilot for decades. Nevertheless, we are all glad that a human pilot can take control when the environment does not match the algorithm.

Moreover, a world where all fund selection is machine-driven would lead to uniform outcomes. If every algorithm picks the same top-quartile funds, markets become more efficient – and ironically, those same funds may no longer outperform. Human insight and differentiated views are essential to preserving market dynamism.

Above all, the human fund selector is accountable. Just as the captain of an aeroplane is responsible for bringing you safely to your destination, the fund selector is responsible for identifying the fund that best suits your needs. Neither autopilot nor GenAI can provide this accountability.

Rather than replacing fund selectors, GenAI is reshaping the role. It frees up time spent on mechanical tasks, allowing more focus on what truly matters: understanding clients, building conviction, and applying judgment. Those who embrace GenAI will thrive in this new environment. Those, who don’t, risk falling behind.

Jean-François Hirschel is the founder and CEO of H-IDEAS, a company which aims at re-establishing trust in the financial world.