Every Help Scout reply arrives with the history attached

What does the reply draft contain?

Three things.

The answer, the two previous conversations with this customer that bear on it, and a flag if the incoming message reads like the second time they have asked. That last one is the difference between a good reply and an infuriating one.

A companion assembles it from the conversation threads and the customer record, then puts the draft in front of you with its sources visible. The cost of a bad draft is ten seconds.

Tag lists decay and nobody looks at them

Two people invent billing and billing-issue in the same week, a third adds Billing with a capital B, and six months later reporting on billing volume means nothing.

Ask for the tag list with volumes and a companion shows you the near-duplicates and the tags used twice in a year. Retagging waits for your confirmation and happens in one bulk pass once you have read the proposal, the way bulk update records in web apps describes.

Which questions should have been a Docs article?

A companion groups a month of conversations by the underlying question, ranks by volume, and tells you which of the top ten already has a Docs article. The gaps are specific and usually embarrassing.

It then drafts the missing article from the actual conversations, in the words customers used rather than the words your product team uses. Publishing stays with you. Wider patterns: ai for support teams.

Configuration the connection never touches

Workflow builders, mailbox and folder configuration, Beacon setup, Docs theming and permissions all stay in Help Scout. The available actions follow the vendor’s public API, so verify the specific object you need before designing a routine around it.

The connected account carries exactly the access of whoever set it up. A mailbox that user cannot see in Help Scout stays invisible to the companion, which is the correct behaviour and occasionally the reason a routine returns less than expected.

Frequently asked questions