With 3rd quarter earnings coming out now every day, I know that I should be listening or reading releases and call transcripts for all the companies I own and especially any new ones and any that may be on my watch list.
That can be a lot of work, especially for me since I have quite a few stocks I own and want to follow.
Well, I recently found out that AI does this for you. I have perplexity on my phone (it’s free) but you can use ChatGPT or any other AI app. You simply ask it to summarize the latest earnings release and conference call for your favorite stock symbol and ask it to include both positive and negative points. It quickly comes back with a great summary and lets you ask follow up questions like what were the major questions during the follow up, etc.
I recognize that it may have some mistakes but it’s a great starting point and very simple and quick to read.
Anyway, maybe everyone already knew this already but if you didn’t, try it out, you may be pleasantly surprised! I was, it’s a standard process for me now!
There are also services like Fiscal.ai that provide the call in audio format, transcribe to text, and provide an optional summary automatically. I do agree that the overview component of these can be helpful. I would, however, urge caution and encourage prudent investors to still double-check the numbers and run through at least the Q&A once in completeness for positions you are holding/thinking of purchasing as these tools are known to be error-prone. It would be a shame to lose one’s hard-earned money on a technicality.
I have a standard question I ask AI, immediately prior to entry exit actions to confirm if any recent changes since my last confirmation from guidance. I ask for quarter over quarter and year over year revenue growth % reported in a standard format. While this needs to be validated, it does provide a quick way to validate. When significant changes from guidance occur, further investigation is justified.
I would challenge this part that it takes too much time to review the earnings transcripts for companies we own. Realistically it may take 30-45 minutes to re-read the prior transcript. I’ll usually re-read the transcript 1-2 days before the company reports, or even the same day so the information is top of mind.
The earnings season happens only once every three months, so we are looking at maybe a 5-10 hour time commitment per 3 months. I don’t think asking AI to summarize a report is nearly as effective as actually reading them. There’s been a number of recent threads about the huge limitations to the approach mentioned here, especially the positive spin that AI layers on top.
If we own so many companies that it’s impossible to review all the transcripts for the companies we own, it is a sign we own too many companies. It’s also not really Saul style investing if owning so many companies is overwhelming us.
Yes, fully agree. You can only be truly certain of the correct metrics if you get the numbers directly from the IR of the company themselves.
Also, AI will miss so much of the detail and tone on the ER that you will always be wondering what it chose to leave out. Easier to listen and read (I do both) yourself.
It is time consuming yes, especially for those of us who are still working full time in a different job. That’s another reason I keep a very condensed portfolio of only 6 to 7 stocks.
I knew some should not like it, if you own 8-12 stock as Saul does then maybe not as critical. For me it is a fantastic find. Especially if you hear about a stock and you want a concise 5 minute read as to whether there is something there to dig deeper on.
There is certainly value there. As a primer, these AI summaries are quite helpful. I don’t think you intended them to be literally taken as a stand alone research method. (I don’t think anyone else stated so, either).
This seems like a different use case then originally described which included using the AI summaries to review for companies you own. I view these as two separate use cases where shortcuts to ramp up on research are generally beneficial but using those same short cuts for companies we own hurts us.
What I mean is with companies we own, we already have money on the line. Taking shortcuts for understanding our companies better, is really only shortcutting ourselves and results. On the other hand, if we are ramping up to learn about a new company quickly using AI with the intention to read the actual report this seems reasonable.
I also see it as a difference if we do not have time to dedicate to investing, or if we view these AI assistance techniques as a superior way to research new companies.
Well, I can tell you a story on how AI would have helped me. I read the earnings call, and they mentioned they had capped calls, and revenue growth of 20% per year, and an ebitda margin of 20%. They did not explicitly mention dilution or eps growth.
I missed the fact that capped calls combined with convertible bond issues means 0 share dilution unless the stock price soars, which in that case, it does not matter.
I also missed the fact that if you combine revenue growth with a stable ebitda margin of 20%, that means EPS growth of 20%.
I used free cash flow for my valuation, but not knowing there would be zero dilution made my valuation slightly off.
AI caught it.
I say, read the earnings call, and run it through AI.