arrow-left arrow-right brightness-2 chevron-left chevron-right circle-half-full facebook-box facebook loader magnify menu-down rss-box star twitter-box twitter white-balance-sunny window-close
Automatic support guidelines through intercom conversation notes
2 min read

Automatic support guidelines through intercom conversation notes

In 2020 we handled about 3300 intercom conversations. These are support questions, replies to automatic alerts or feature suggestions.

Prezly Client-Day

At Prezly we have a client-day rotation system where everyone from CEO to server person takes one day of support per week. We do this because

1) we need everyone to understand the pains & struggles of our customers.
2) as a product-first company, we need our people to understand the platform the best

This is working out pretty well with a high satisfaction rate on support. Also, we don't have dedicated support roles.

Quality Standards for Customer Support

The challenge with this approach is that you need to invest in good onboarding for everyone to understand the product, do a lot of information sharing (what do i tell the customer?) and have solid procedures in place how to handle common problems or questions:

  • This email address didn't get my pitch. Why?
  • Can I add an HTML block to the editor?
  • I can't get my IT team to verify DKIM/SPF on Any tips?

We document everything in notion making the support section look like this:

Support section in Notion

And this is just for internal use. For our customers we have (which we want to move to a Prezly newsroom).

Meet Intercom Bot

To assist everyone on client day I thought it was useful to add a contextual note to a conversation once it is tagged.

Every conversation is tagged in a few categories:

  • type_bug
  • type_feature (syncs to ProductBoard)
  • type_subscription (upgrades, downgrades, ...)
  • type_question (how do I ?)

Right after the message is tagged a note is automatically created with the most important information about that specific type.

Here is a quick GIF showing how it works:

Example note to intercom conversation

It's a super simple piece of code (100 LOC) but it accepts an intercom webhook, looks for an HTML template with that tag name and if it finds that reply as an intercom note within the same conversation.

I have open-sourced the code for this:

Contribute to prezly/intercom-bot-nextjs development by creating an account on GitHub.

It should take about 5 minutes to set it up. We have enabled automatic deployment using Vercel so anyone can edit the html templates and see the bot updated in almost in realtime.

Let me know what you think. You can find me @digitalbase