Trying to understand how analysts at brokerage firms go from client requests to market surveys

Hey everyone, we are three UC Berkeley Computer Science students building AI tools for the commercial real estate world, starting with multifamily.

Over the past couple of weeks, we’ve been talking to analysts and brokers at different firms to understand how they actually work. What’s become clear is that the “client request to market survey” process still takes hours of manual work — digging through CoStar, Reonomy, phone calls, drive-bys, and scattered spreadsheets.

We’re exploring how to automate the grunt work of that research:

• aggregating and verifying property data from multiple sources,
• helping analysts judge fit and plausibility (“does this deal make sense to send to the client?”),
• and generating clean, branded market surveys in minutes instead of hours.

Right now we’re just trying to learn.
If you’ve worked on the buy-side brokerage or analyst side — especially in multifamily — we’d love to hear how your workflow actually runs day-to-day.
How are client requests captured? How do you use CoStar or other data sources? What parts feel repetitive or could be automated?

Please reach out to me at aryanv0213@berkeley . edu 

if you are willing to hop on a quick call with us. Any time or insights will be truly helpful. Thanks!