Jeff Cangialosi

Last week, I was faced with the tedious task of trying to sift through conference attendee lists to figure out if a handful of conferences were worth attending and determine who we should meet there. If I was at a large financial institution or consulting firm, this would be a great task for a junior analyst. Of course, I am in startup world, so there is no junior analyst. You have to do your own list sifting here.

This felt like a perfect use case to take some AI tools out for a spin and see what they could do to save me some time. I wanted to gather additional details on the attendees and flesh out the lists with a few more columns of data. There is likely a simpler way to do this, but here is how I hacked together a few free solutions to derive a workable Google Sheet that enabled me to quickly assess the value of the conferences. Here are the steps I followed:

Step 1: Most the attendee lists I had access to were PDFs. I started by converting the PDF attendee list into a CSV file that I could then dump into a Google Sheet and parse. I now had a workable Google Sheet that I could append data to about each company. I used Zamzar, but I am sure there are other solid tools out there.

Step 2: I played around with a couple of AI tools to figure out which ones would enable me to upload a CSV and then export back out a CSV file with the information it had scraped. I eventually settled on Claude. Based on my 2 minute Google search, it seemed to be the best option for this, particularly on its free version. I will note that I’ve now run this process a few times and eventually hit Claude’s question limit and had to upgrade to Pro. For the last few months, I’ve been partial to using Perplexity, but I wasn’t able to export to CSV with Perplexity on the free version.

Step 3: I uploaded my CSV file with company names to Claude and then prompted it to pull company descriptions and a few other particular details I was looking for. I asked Claude to limit company descriptions to no more than 2 sentences so that I had a concise summary I could quickly scan visually and also do a Control F on to find certain key words like “payment processor.”

Step 4: I let Claude churn for a couple minutes and then eventually it developed a CSV file that I could export. I then copied and pasted that CSV into my original Google Sheet.

Step 5: I used a simple formula =SPLIT(A1, ”,”) in Google sheets which split apart the data in each cell at the point where Sheets saw a comma. I now had a nice spreadsheet with company name and several columns of information.

Ultimately, nothing here was magic, but it was a solid use case where I derived significant time savings from AI. Ten years ago, a lowly analyst like me at JPMorgan Chase would have been tasked with doing all the work above manually. The automation tooling is now here. For someone with limited coding experience, it takes me a little tinkering time with these new tools, but the tinkering time does feel worth while and really necessarily if I’m going to keep up with all the new utility AI tools are bringing to the table.