AI Support Control Room for Zendesk
The narration from the demo video, word for word. Timestamps mark where each part starts in the video.
0:01 Hey everyone, my name's Anwana Ntofon, and I built an AI-powered support workflow for teams who are spending more time writing replies than actually solving problems.
0:12 Support teams spend the majority of their time doing the same thing, reading a ticket, figuring out what the customer wants, and writing a reply from scratch.
0:20 I built a system that handles that analysis automatically and routes a draft to a human for approval before anything gets sent to the customer.
0:28 Let me show you how it works. I have three additional tabs open, one being the email client, which I'll use to simulate as the customer.
0:41 The second is Zendesk, where ticket management occurs. And the third is a custom dashboard that I built.
0:48 Which handles ticket management, but is also where drafted public replies are sent so a human can intervene and approve before being sent back to Zendesk.
1:00 So with that, let's go ahead and create a new ticket. This is going to be an authentication-based ticket. I'll start with the requester.
1:13 Subject: Help with authentication. Message: I need assistance authenticating my users.
1:29 I'll submit this as a new ticket. Watch what happens, in a moment these tags will auto-populate.
1:42 And we're going to see an internal note populate with an understanding of what the ticket is about, so the agent can come in already knowing what's needed.
1:52 And get that first response out as soon as possible. The Control Room refreshes every 30 seconds.
1:59 And right on time, the ticket comes in with a draft public reply. The agent can edit directly, or just approve and send.
2:09 This looks good. I'll add the agent name, confirm and send. Approved with a timestamp, and the status shows published.
2:17 This has been sent back to Zendesk. And in the email client we can see the ticket was created and a support agent has already reached back out to the customer.
2:27 This conversation will continue until the ticket is resolved. But I built this because support agents and customers both want a more streamlined process for how they collaborate. Agents want to solve problems in the best way possible.
2:45 This tool is my version of addressing those concerns. If you want to talk more about the tech stack or anything further with this project, I'd love to connect.
2:55 My email and LinkedIn are on screen. Hope you have a great day. Take care.