Email Intent Parsing with Bot Flows

We are migrating our email routing infrastructure. I need to determine if Dialog Engine Bot Flows support native parsing of inbound HTML emails to extract regular expression entities, such as invoice numbers. Furthermore, if the bot successfully classifies the intent Inbound Email flow?

Greetings! This is a fascinating architectural question that we frequently discuss in our local user group meetings! While Dialog Engine Bot Flows are incredibly powerful for synchronous conversational channels like Web Messaging, they are not designed to natively ingest and parse complex HTML email payloads. The optimal design pattern involves utilizing a standard Architect Inbound Email Flow as your primary entry point. Within that email flow, you can utilize the FindRegex string expression to extract the invoice number from the Email.Message.body variable.

Subsequently, you can evaluate the text and dispatch an automated reply directly from the email flow utilizing the Canned Responses integration, completely bypassing the need for a Dialog Engine bot!

Hey! Yeah, do not try to use a Bot Flow for emails. I inherited an org where the previous admin tried to do exactly that, and it was a complete disaster. Bot flows are built for turn-based conversations.

An email is just one giant block of text. As the other reply mentioned, you should just use the standard inbound email flow.

You can use the ‘Send Auto Reply’ action and select a canned response that has your HTML template in it. It is way easier to maintain.

I spent three weeks trying to force an email body into a Dialog Engine bot slot and it is completely impossible! The bot engine strips out all the HTML formatting anyway before it even tries to process the intent! It is incredibly frustrating that the documentation does not make this clear immediately. You absolutely must use the Architect Inbound Email flow for this. Do not waste your time trying to build a chatbot to read emails.

Just use regular expressions in the Architect flow and call it a day!