AI Automation to Parse Lead Information from Aggregator Emails

The client is a wedding industry professional who is on a listed on a number of "aggregator"/directory services, including WeddingWire, The Knot, and Zola. If you are unfamiliar with these services, think of them as the cars.com of the weddings industry.

The client tracked their leads in Airtable. The problem was that the aggregators provided no API or webhook functionality that would enable the client to automatically import the data into Airtable, and as about half of the leads from the aggregators are of a lower quality, it was fast becoming not worth the time to manually enter the information for each and every lead into the system.

The solution: parse the email notifications from the aggregators and load the data into Airtable automatically.

Initially, the client automated this with increasingly complicated regular expressions, but each aggregators' emails were formatted slightly differently and the styling of the emails changed frequently, so this required near constant maintenance.

Examples of the differently formatted messages that would come in from different aggregators

I solved this problem by using OpenAI's GPT3.5 API to parse the emails, and then pull them into Airtable automatically.



The final automation

A lead parsed into Airtable