AI, for your data, your questions, your answers

Author

Andrew Meister
October 24, 2023

AI For Your Data, Your Questions, Your Answers

Oct 22, 2023 • 1 min read

I just started writing on Beehiiv. This is my first ever post. I can’t tell you what to expect or how often to expect it. Generally, I think my content will focus on my learnings of being an equity research analyst and my love of technical solutions to make the jobs to be done of an analyst more efficient. I will also write about my nights and weekends project, DoTadda.

Who should subscribe? Analysts, PMs, corporate strategy, and investor relations managers and of course other personas I haven’t thought of. Consultants would be a good fit here too.

When we started DoTadda, we focused on finding your businesses’ content like emails, files and notes, really, really fast. I work in asset management, FAST is a big deal.

Our hypothesis was “search” in apps is generally terrible, like searching for an email or a file. You don’t get the needle with search in these products, you get the whole haystack. I wish it was this easy, where the needle sits on top. In the real world, you’re still pawing through emails.

My favorite part of demos was to type in a stock ticker and immediately (milliseconds) get the file to the model back in our search results, because I knew how long it would take to do this simple search on the network drive. It was an easy win in a demo.

The same was true of finding broker research: type in a ticker and a broker and boom, all that broker’s research on said ticker shows up. Another easy win.

But now with LLM’s DoTadda can be even more powerful, hence the title. LLM’s can help search results, summarize an individual database entry or summarize multiple entries. It can help you dig deep into trends. But only on your data or external data you trust.

We are not abandoning blazingly fast key-word search on your data, but rather augmenting the understanding of your data with semantic seach and “chat”.

Like this (AI-generated, not investment advice or recommendation(s)) :

The difficulty we are finding is the UI, simply because blending three different types of interaction with your data is quite novel. Needless to say we’re working on an elegant, clean sheet of paper solution. Stay-tuned.

Best,

Drew

The whole post is here with demo videos:

Subscribe if you want to learn how a career equity research analyst thinks about AI.

Reply

or to participate.