Support Stack: Inside Tettra’s Intercom Setup: How Fin Collects Error Details Automatically | Ep.1 with Jen Weaver

Every support team has experienced it — a customer sends a single-line message saying "it's broken," and six hours later your team is still chasing context that should have been captured at the start.

In this episode of Support Stack, Conor Pendergrast is joined by Jen Weaver, a customer support specialist at Supportman who manages Intercom for Tettra, to explore how Fin can be configured to ask better questions before a conversation ever reaches a human agent.

Jen demonstrates how Tettra uses Fin's Train and Guidance section in Intercom to set up custom guidance called "clarify error details." Rather than letting vague bug reports pass through to support staff unanswered, this guidance instructs Fin to proactively ask customers for the specific context it needs — the exact error message, what they were doing when it occurred, and whether it is reproducible. The result is that when a conversation does get routed to a human, the specialist already has all the information they need to diagnose the issue rather than spending the first exchange asking the same questions Fin could have handled automatically.

Conor also raises an important downstream effect that many teams overlook: without the right intake questions, support teams systematically underestimate the true impact of a bug. If only three customers push through a frustrating reporting process, the real number affected could be thirty. Configuring Fin to collect meaningful error context at the start of every conversation leads to better data, better prioritisation, and fewer surprises.

If you are working with Intercom, Fin, or exploring AI in customer support, this episode provides a practical and quick technique for getting more out of your Fin setup from day one.

Join my daily email list to get emails about future episodes.

Episode transcript

Conor (0:00)
Hi, I'm Conor Pendergrast from CustomerSuccess.cx

Jen (0:03)
And I'm Jen from Supportman

Conor (0:06)
and welcome to episode one of Support Stack.

Conor (0:17)
So one of the things, Jen, that is very frustrating, as we all know, as customer support professionals, is when you get a message from a customer and they send a message, let's say, six or seven hours ago, and you're replying to it now, and all they've said is, it's broken. And I wondered, is there anything that you've done that helps your version of Fin to ask questions and then involve the customer a little bit more before it gets to you?

Jen (0:45)
Absolutely, that's the kind of thing that AI is really great at. We set up Fin to get more information before it even hits a person.

Conor (0:56)
That's incredibly useful. So was that a workflow? Was that something else? What did you do in Intercom to get that set up?

Jen (1:03)
Yeah, it's um, I could share my screen if that works and we can walk through it.

Conor (1:10)
Yeah.

Jen (1:11)
So we go to the Fin AI agent. And then from there go to Train and Guidance. And here is where we can customise that particular getting more information thing.

Conor (1:29)
Super. So you just added it in here. Can we show which one it is so that people can see the exact wording and how it works?

Jen (1:38)
All right, so the clarify error details in context and clarification here. So this is the wording that we use for prompting Fin. This isn't what it's going to say. It's gonna use this to decide what to say to get information from the customer.

Conor (2:00)
So instead of a customer just coming along and saying, I have a bug, and then six or seven hours later, they get a message back from your team asking for the same details that Fin can just ask the questions for. Fin just goes ahead and asks those questions and gathers the context and the information.

Jen (2:19)
Exactly.

Conor (2:19)
Do you want to test it? I could just say this is broken. Do you think that's? uh

Jen (2:25)
That's great. Tettra's customers are way more clear and like a lot more communicative. But sometimes we come across other customers and other users who that's all they say. And you got to have some way of interpreting what's going on. So now Fin has come back and said, OK, there are no system wide problems, but... Fin is offering some suggestions.

Conor (2:56)
And then, yeah, Jen talked to me about what Fin says at the bottom here.

Jen (3:02)
Yeah, so it's running it through basic troubleshooting, right, and getting more information, which is the key. Because if this conversation comes to me or one of my specialists, then I don't want them to just have to ask, well, give me more information. So hopefully this person will reply, yes, that's what I was looking for, or no, here's what's going on. And this right here is key. The bullet points asking exactly what's going on.

Conor (3:35)
Yeah, that's perfect. And that saves so much back and forth later. Cause as a customer, there's nothing more frustrating. And this happened to me recently and I won't name the company, but something happened and, like, the terrible AI intake bot didn't give me any options to provide more context and to provide like screenshots or provide exact pages or anything like that. It just took a really basic description and then vanished off. And a day later I heard back quite a generic, have you tried turning it off and on again, response. And I was like, this is not what I was hoping for, not what I was aiming for. And then before I could email back, they fixed the problem that they had introduced. And so it was working again, but it still really annoying. And it's one of those things where...

Conor (4:31)
It is funny, there are downstream effects of this because what happens is if you're not getting all the correct information from customers, you're going to be underestimating the size of the impact of a bug or anything like that, because you don't have all of the information surrounding it. So you might think like, well, three people reported a bug, but that's actually three people who reported a bug and got through the whole process of reporting a bug. Maybe there are another 30 people who tried to report a bug, you weren't asking the right questions with Fin upfront to make sure that you're actually gathering all the data.

Jen (5:11)
Yeah, or like in your case, uh the support ticket just got abandoned because you got fixed. And so nobody at that team knows that your email was about that bug.

Conor (5:23)
Exactly, exactly. Yes, that's great. So this is the guidance you've set up to clarify error details. That's really helpful.

Jen (5:32)
And just to clarify, when we ask this question, this doesn't add to our tickets. This is just like a sandbox, right? You're not paying the 99 cents for these conversations to resolve them.

Conor (5:51)
Cool, good to know. Well thank you very much, Jen Weaver from Supportman. If we wanted to find out more about you and what you do, what's the best way to do that?

Jen (6:02)
LinkedIn. Connect with me on LinkedIn. I always accept connections. I love it.

Conor (6:08)
That's fantastic. Well, if you enjoyed this episode and you want to see more real Intercom setups like this, go to CustomerSuccess.cx/daily and join the email list there. I will email you every weekday and I'll also tell you about all future episodes. Otherwise, bye.