“Is AI necessary in investment advisory? Will it make a difference?” was one of the responses we received for this article: Retirement Planning with AutoGPT: The Informal Future of Investment Advisory.
The truth is AI is already well and truly here and here to stay. At the time of writing, in areas relevant to us, AI was being used as chatbots, custom image and video designers, data analytics, and data and text summarizers (Company reports for stock analysis are one of the earliest implementations).
Amazon has started summarising reviews with AI. AI plays a big role in deciding our social media feeds. Soon, Meta will include AI as an assistant in Facebook groups. Once Google implements it seamlessly into search, it will be a game-changer. And this is just the start! A few days ago, Google launched VideoPoet, a new large language model for text/image/sound to video generation! Grammarly now has generative AI to help with essays and email responses. Things are happening so fast that I had to update this paragraph twice in the last few days before publication!
AI is used for all sorts of content generation. At freefincal, we use AI to write article excerpts for social media posts. We use AI for headline generation. The headline is the most important part of an article. We ask ChatGpt for 7-8 alternatives (giving our headline or the entire article as input).
It is then mixed and matched to create what (hopefully) would be attractive and representative but not clickbaity (naturally, this is subjective). We have already tried the AI-generated freefincal podcast! Soon, we hope to use AI image generation for unique featured images. Incidentally, stock photo portals like Getty Images and Istock already have this option.
🔥Secure your future with our Robo-advisory tool trusted by over 3,500 investors and advisors. From effortless retirement planning to funding your children’s biggest dreams, turn your financial goals into reality. 🔥
Subscribe for money management solutions via email! (Link takes you to our email sign-up form) Join 32,000+ readers in our community.
👉 New Tool Alert! NaviPlan: A Privacy-Focused Multi-asset Tracker and Goal Planner 👈
We are not impressed with how ChatGPT performs as a chatbot when given copious amounts of freefincal information. Yes, they need meticulous training, which is quite expensive and does not make sense for us.
So, to answer the reader’s question, “Is AI necessary in investment advisory? Will it make a difference?” AI has a role in financial advisory; it will make a difference, but it must be used, monitored and regulated well.
For example, if someone trains the AI model on all the books used in the Certified Financial Planner course or the NISM Investment Advisor Exams level 1 and 2 syllabus or any other broad knowledgebase and allows it to create financial plans for clients, which are then human-verified before handover, I would not be so enthusiastic.
This is because the possibilities of creating a financial plan for a given circumstance are generally quite diverse. If we send a client to 100 SEBI Registered Investment Advisors, we will get 100 different plans. This means the AI can churn out many more combinations, most unnecessarily verbose, and the probability of some of them being wrong will be quite high.
If we send 100 clients to the same advisor, we can find patterns and create a robo-advisor with AI as an assistant to both parties.
If we can train the bot only on our instructions for different client groups (the patterns and the associated math), we stand a much better chance (with human review). Naturally, this still won’t be perfect, but it should be better than a textbook with generic instructions. The “voice” of the person or organization behind the effort will be seen to some extent, increasing the number of clients they can service.
To this extent, I see promise in AI implementation. After the financial plan creation (by human, robo or AI), clients may have many questions which could be left to the AI to summarise from the plan. This can save immense time for the advisor.
I am even more enthusiastic about using AI as the front end of a robo advisor. Take, for example, the freefincal robo advisor tool. If a large language model can appreciate the rules behind it and learn how to use it, it can take questions from a human user, feed them into the robo tool and come up with easy-to-understand action items. It can summarise and perhaps even simplify the outputs (hopefully without distortion). At the time of writing, this is more of a goal without a plan, a.k .a. a dream.
I am sure there are other ways to implement AI into financial advisors that I cannot think of, and I eagerly look forward to seeing them. In 2024, this is going to happen gradually. Regulators must quickly establish norms for entities using AI. Otherwise, it will be another case of watch, learn and react instead of pro-act. The ball is in SEBI’s court, and the clock is ticking.
Using artificial intelligence requires real intelligence. The latter has always been hard to find. So caveat emptor (buyer beware) and caveat lector (reader beware) are now more important than before.