Skip to content

Real client

Real estate investors: incoming deals valued, 11,927 runs in one month

Deals arrive from email, feeds, JV partners and Facebook. Each address now gets value, condition, legal description and sale status from about 15 data sources, and the spread is worked out.

Real screen, names boxed
Real client system in Airtable. Addresses and prices on screen are changed.
Read the transcript

More than like 15 APIs have been running on their own, and they're getting the information, and that's all real-time. Some emails come, there's an AI bot running, the process, those emails, and it's getting information and send it to the data table. And then we need to check some data for other resources, gathers the information based on the addresses and gets the real-time value, its condition, and its attributes, and its legal description, whether it's for sale or not for sale. The Airtable is very powerful here in the calculations. Recording the spread, and there are other formulas. Like, if I want to process one record manually, I would just press this Process button and it's going to all the APIs that are being used here. And all these are going to come back here in the real time. I have created like many automations here, and all of them are like very critical, doing some very business critical processes. And then like there are some interfaces, and you can just create some dashboards the way you like. And even if I don't press it, still it just works.

Trade
Real estate investors
Tools
  • Airtable
  • Python
  • Noloco
  • about 15 property data APIs

Before and after

Deals came in from email, feeds, partners and Facebook. Each address was checked by hand.

Before and after
BeforeAfter
Same house, two sourcesChecked twice by handFlagged as a duplicate
Address checksSomeone checked every address by handValue, condition and sale status from about 15 sources

What we built

  • Seven ways in. Parsed emails, RSS feeds, JV submissions, Facebook, direct-to-seller, portfolio lists and manual entry land in one table.
  • About 15 data sources. Value, condition, legal description and sale status for each address.
  • Duplicates flagged. The same house from two sources gets flagged, so nobody checks it twice.
  • One morning list. Daily and weekly digests show the deals waiting for underwriting.
  • Runs without the button. A Process button runs one record by hand. Nobody has to press it.

How it works

  1. Deals come in from email, feeds, partners and FacebookAirtable automations
  2. AI reads each deal email into one recordAI email parsing
  3. Every address gets value, condition and sale status from about 15 data sourcesPython, property data APIs
  4. Duplicates are flagged and the spread is worked out per sourceAirtable
  5. Each morning the team reads one list of deals waiting for underwritingAirtable digest

What broke and how we fixed it

What broke

The data sources did not always agree on what a house was worth.

How we fixed it

The spread is worked out per source and as an average, so the team sees when they disagree.

The result

11,927 automation runs on new property records in one month

Source: Seen on screen in the build recording: Airtable automation counters, Apr 2025

Checking every deal address by hand?

Tell us where your deals come from and what you check on each one.

Show us your deal flow

We reply within one business day. No card, no contract.

More builds like this

More on AI Workflow AutomationSee all case studies