Feedback source
Zendesk tickets become ranked product feedback
Feed new Zendesk tickets into FeedbackFlow and the product signal buried in your support queue becomes a prioritized, deduped list.
Why connect Zendesk
- Mine the support queue. The feature requests inside support tickets stop getting lost.
- One trigger, done. A single Zendesk trigger sends tickets in as they are created.
- Ranked with the rest. Zendesk feedback is scored and grouped alongside every other source.
"Export keeps timing out on large projects, this is blocking our team."
Deduped and ranked alongside every other source.
How it works
- 1
Create a webhook
Point a Zendesk webhook at your FeedbackFlow ingest URL.
- 2
Add a trigger
Run the webhook whenever a ticket is created.
- 3
See it triaged
Tickets arrive as feedback items, deduped and ranked.
Prefer the full walkthrough? Read the Zendesk setup guide.
Before you start
- Zendesk Admin Center access, since both the webhook and the trigger live in admin-only areas.
- A FeedbackFlow project, which supplies the Request URL and the JSON body you will paste in.
- A decision about scope: every new ticket, or only the ones matching a condition you set on the trigger.
How to connect Zendesk to FeedbackFlow
- 1
Create the webhook
In Admin Center, go to Apps and integrations, then Webhooks, and create a webhook with method POST pointing at the Request URL shown on the Zendesk card in FeedbackFlow. This is the endpoint that receives your tickets, and the URL is unique to your project.
- 2
Paste the JSON body template
Set the webhook JSON body to the template FeedbackFlow gives you. It maps four Zendesk placeholders onto the fields FeedbackFlow expects: ticket ID, description, requester email, and requester name. Copy it exactly, because the field names are what the ingest endpoint reads.
- 3
Copy the signing secret back
Zendesk generates a signing secret for the webhook. Copy it into the Zendesk card in FeedbackFlow. Every delivery is signed with it, and FeedbackFlow rejects anything whose signature does not verify, so a leaked URL on its own is not enough to inject feedback.
- 4
Add the trigger that fires it
In Admin Center, go to Objects and rules, then Triggers, and add a trigger that runs your webhook when a ticket is created. This is where you choose scope: fire on every new ticket, or add conditions so only tickets in a particular group, form, or tag are sent. The webhook does nothing until a trigger calls it.
The exact Request URL and any values you need to paste are shown on the Zendesk card inside your project settings, because they are specific to your project. Full Zendesk setup guide.
What FeedbackFlow reads from Zendesk
Exactly the four fields in the JSON body template, and nothing else. Zendesk decides what to send, which means you control the payload rather than handing over blanket access to your queue.
| Field | What FeedbackFlow does with it |
|---|---|
| ticket.description | Becomes the feedback item and the text the AI triages |
| ticket.id | Stored as the external ID, so a retried delivery is not ingested twice |
| ticket.requester.email | Attached as the requester, for closing the loop when work ships |
| ticket.requester.name | Shown alongside the email so topics carry real names |
What this looks like in practice
A ticket arrives that is really a feature request
Your agent answers it as usual. In the background the description becomes a feedback item with the requester attached, so the product signal survives even though the ticket gets closed.
The same request shows up in thirty tickets over a quarter
They collapse into one topic with a count of thirty and thirty named requesters, turning a vague sense that "people keep asking for this" into a number.
Product picks it up
One click drafts a Jira or Linear issue summarizing all thirty tickets, with acceptance criteria written from what customers actually described.
Limits worth knowing
What this connector does not do, so you can find out here rather than after you have wired it up.
- Only the ticket description is sent, which is the first message. Later comments on the ticket, including the ones where the customer clarifies what they meant, are not captured.
- The trigger fires on creation, so tickets that turn into feature requests later in their life are not picked up unless you write a trigger that handles that case.
- Every ticket matching your trigger becomes a feedback item, support noise included. Narrowing the trigger conditions is the way to control that, and it is worth doing.
- This is a one-way source. FeedbackFlow does not write back to Zendesk: no tags, no comments, no ticket updates.
Frequently asked questions
- Will this send my entire Zendesk queue into FeedbackFlow?
- Only what your trigger matches. If you fire on every ticket created then yes, everything arrives, which is fine for a small queue and noisy for a large one. Most teams add conditions so only tickets in a particular group, with a particular form, or carrying a feedback tag are sent. You can start broad and tighten the trigger once you see what is coming through.
- Does my support team have to change how they work?
- No. The integration is one-way and invisible from the agent side: nothing is written back to Zendesk, no tags are applied, and no ticket state changes. Agents answer tickets exactly as they do now. The only thing that changes is that the product signal in those tickets stops dying when the ticket is closed.
- Which Zendesk fields does FeedbackFlow actually receive?
- Four, and only four: the ticket ID, the ticket description, the requester email, and the requester name. They are set by the JSON body template you paste into the webhook, so the payload is defined in your Zendesk admin rather than by a broad API scope. If you do not want requester identity leaving Zendesk, remove those two lines from the template and the rest still works.
- What happens to duplicate or retried deliveries?
- Each ticket is stored against its Zendesk ticket ID. If the webhook is retried, or a trigger misfires and sends the same ticket twice, the second delivery is recognised and discarded rather than creating a second feedback item. AI dedupe then handles the different problem of many distinct tickets describing the same underlying request.
- Can I connect Zendesk and Intercom at the same time?
- Yes, and it is where the dedupe earns its keep. The same feature request arriving through Zendesk tickets, Intercom chats, and your in-app widget collapses into a single ranked topic with the combined count and every requester attached. Ranking across all your channels at once is a materially better prioritization signal than ranking each queue separately.
Connect Zendesk today
Free plan. No credit card. Live in 2 minutes.