Can an agent reply to customers through this?

No, and we would push back on wiring one up if you asked. Reading is where an agent adds something no one has time to do; replying is where a wrong answer is sent in your name before anyone sees it. Use it to prepare the answer, and let a person send it.

Does it see personal data in the conversations?

It sees what the conversations contain, which in a support inbox routinely means names, email addresses and whatever a customer pasted while frustrated. Treat it as the sensitive read it is: it is worth knowing which of your agents have this connector enabled before you turn it on.

Is this better than Intercom’s own reporting?

It answers a different question. Intercom counts and tags well; an agent reads. "How many tickets about billing" is reporting’s job, and "what specifically is confusing people about billing" is this one’s.

Customer support

Intercom + Burrak AI

Read what your customers actually said, not what the dashboard counted.

Connects with OAuth — revoke any time

Support inboxes hold the most honest product feedback a company gets, and almost nobody has time to read them in bulk. An agent can: work through the conversations, find the question five people asked differently this week, and tell you whether the help centre already answers it. That is a reading job at a volume no one does by hand, and it is where the value is — not in replying faster.

In the connector catalogue: Search conversations, look up contacts, find articles.

Things you could ask for

Written the way you would actually say them. An agent plans the steps itself.

“Read this week’s conversations and tell me the five things people are actually confused about.”
“Find every conversation where somebody asked for a feature we do not have, and group them.”
“Which questions come up repeatedly that our help centre has no article for?”
“Look up this customer and summarise every issue they have raised, oldest first.”
“Compare complaint volume before and after last month’s release.”

This connector reads — conversations, contacts and help-centre articles. It is not a way to have an agent answer your customers, and that is the right shape: the reply your customer gets should be a decision somebody made.

Connecting it

1Create a Burrak account — the free plan is enough to try this.
2Open Connectors and choose Intercom.
3Authorise Burrak in Intercom's own window. Your password never reaches us; we receive a token you can revoke whenever you like.
4Give an agent a task. It will use the connector when the work calls for it.

Start free How connectors work

Intercom and Burrak, in practice

Can an agent reply to customers through this?: No, and we would push back on wiring one up if you asked. Reading is where an agent adds something no one has time to do; replying is where a wrong answer is sent in your name before anyone sees it. Use it to prepare the answer, and let a person send it.
Does it see personal data in the conversations?: It sees what the conversations contain, which in a support inbox routinely means names, email addresses and whatever a customer pasted while frustrated. Treat it as the sensitive read it is: it is worth knowing which of your agents have this connector enabled before you turn it on.
Is this better than Intercom’s own reporting?: It answers a different question. Intercom counts and tags well; an agent reads. "How many tickets about billing" is reporting’s job, and "what specifically is confusing people about billing" is this one’s.