Article

How to export Zendesk tickets: CSV, JSON, XML

By TrustPast

Updated October 22, 2026

A company of 30 or more that runs a support desk on Zendesk has its customers’ side of the work in tickets. Every question, every complaint, every answer the team settled on, with the date. A buyer pays to license records like these to train AI models. You don’t export anything to apply. You count what you have, and an export comes later, if at all, after an agreement says what it covers. This guide settles what to count, which Zendesk plans can export, and what each format holds, in Zendesk’s own words.

Checked on 22 October 2026 against Zendesk’s help page for the Growth, Professional, Enterprise and Enterprise Plus plans, linked at each step. Menus change. If a step here differs from Zendesk’s page, Zendesk’s page wins.

You don’t export to apply

micro1 is the buyer on this page. micro1 names customer support among the operational data it looks for: “Support documentation, ticket workflows, QA processes, customer operations, and knowledge management.” A Zendesk account is most of that list.

micro1’s published steps start with the company, not with files. First “Get started”. Then “Evaluation”, where “Our team reviews your submission”. Then “Discovery”, where, if there is a potential fit, “we’ll request additional information”. Then “Partnership”. Nothing in those steps asks for an export. Its estimator asks “How many employees does your company have?” and “How many years of data do you have?”, with the note “Typically the number of years you’ve been in business”.

Files come later, and only inside an agreed scope. micro1 publishes this, word for word: “Before any participation, together we define:”

“Approved data scope and clear boundaries”
“Security and confidentiality standards”
“Anonymization and redaction requirements”
“Internal approval processes and permissions”

Then: “If approved, we’ll finalize the agreement and data requirements and begin the partnership.”

micro1 does not publish how records are delivered; that is agreed in its scope step. So this guide does not tell you to send a file anywhere. It shows what Zendesk’s export produces, so the scope you agree to is one Zendesk can actually deliver.

There is a practical reason too. On Zendesk, exports are off until the account owner asks Zendesk to switch them on. You can’t export on a whim, and you don’t need to.

Apply & License Your Data

How much you have, and where to find it

The numbers that matter are people, years and tickets. Zendesk’s export article does not describe a screen that shows the total number of tickets in an account. Zendesk does not publish one there, so here is our own method.

The years. Find the oldest ticket still in the account. Its date is where the record starts. Note whether an older help desk came before Zendesk; those years may sit in an old export somewhere.
The tickets. Every Zendesk ticket carries a number. As a rough check of our own, compare the newest ticket’s number with the oldest one’s. The gap is a ceiling, not a count, because deleted and merged tickets leave holes.
The channels. Note which ways customers reach you: email, web form, chat, phone. Each is a different kind of conversation in the same record.
The knowledge. Count the help centre articles and the macros, the saved replies the team uses. micro1 lists “knowledge management” and “ticket workflows” among the support records it looks for. A macro is a decision written down once so the team does not make it again.
The people. How many staff answer tickets, and how many work at the company in all. micro1’s published line is a company of 30 or more employees.

That gives you five lines: years, a rough ticket count, channels, knowledge articles, people. That is what we would bring to the first call; micro1’s estimator asks for two of them, employees and years. Our estimate asks the same kind of question.

Who can export, and what the plan allows

Zendesk sets three gates. (Zendesk’s export page)

The plan. Zendesk lists the Growth, Professional, Enterprise and Enterprise Plus plans, and Support Professional and Enterprise. It says: “The exporting tools described in this article are not available on Team plans.”

The switch. Exports are off by default. “The account owner must contact Zendesk Customer Support to enable data exports in your account.” Zendesk asks for the account’s Zendesk Support subdomain in that request. That one sentence is the step most owners don’t know about. Only the account owner can make the request.

The role. Zendesk admins run the export, and Zendesk adds: “For security reasons, you can restrict who can export the data.” Our advice is to use that. Limit exports to the owner and one admin, and write down who they are.

There is a fourth route Zendesk mentions. “customers on all Zendesk plans can use the Zendesk REST API to export data.” That is a developer’s job, and it pulls whatever the code asks for. If a scope ever calls for the API, the scope should say which endpoints, which dates and which fields.

On frequency, Zendesk’s article states no limit on how often you can export. It does not publish a figure.

The steps, in Zendesk’s words

These are Zendesk’s steps, quoted or kept to Zendesk’s menu names. Each ends with its page. Run them after an agreement, not before.

“If not done already, get Zendesk to enable data exports in your account.” This is the account owner’s request to Zendesk Customer Support. (source)
In Admin Center, click Account in the sidebar, then select Tools > Reports. If needed, click the Export tab. (source)
Select an export option. For JSON, choose a date range, select tickets, users or organizations, and click Export. For CSV, choose a date range and click Export. For XML, select “request file” next to the option you want. (source)
Wait for the email. Zendesk sends a download link when the export finishes, and the link downloads a zip file. “The download link is valid for at least three days.” (source)
If the email doesn’t arrive, use the “latest” link next to each export type. Zendesk says it downloads the most recently generated report, with the data as it was when the report was requested. (source)

Zendesk adds that export errors can sometimes be fixed by changing the date range. A narrower range is also how a scope is usually written, so the two fit.

What the file contains

The format decides what you hand over. Zendesk’s three formats hold very different things. (Zendesk’s export page)

JSON. Exports tickets, users or organizations, one at a time. Zendesk recommends it for accounts with “more than 200,000 tickets.” It carries the ticket comments, with one exception: comments are left out for any ticket over 1 MB. Organization exports include deleted organizations. It is not available in sandbox instances.

CSV. Ticket data only. Zendesk is plain about the limit: “The data does not include deleted tickets, ticket comments, or ticket descriptions.” Multi-line text, multi-select and custom date fields are also left out. A CSV is a list of tickets with their fields. Zendesk’s column list includes “Requester”, “Requester ID”, “Requester email”, “Requester domain”, “Submitter”, “Assignee”, “Subject”, “Tags”, “Status” and the dates. It is not the conversation, but it is not free of personal data either. And Zendesk says: “If a single ticket has more than 1 MB of data, the ticket is excluded from the report.”

XML. Two kinds. Full XML covers account settings, groups, organizations, tickets with comments, and users. User XML covers groups, organizations and users. Both leave out custom ticket fields, metric sets and comment metadata. XML files stop at 500 MB, about 200,000 tickets, and “Full XML and user XML exports are not available for accounts with more than 200,000 tickets.”

Here is our own reading of what that means. The value of a support record is the conversation: the customer’s words, the reply, the follow-up, and how the answer changed over the years. The CSV has none of that. If a scope ever asks for ticket content, it means JSON or full XML, and the scope should say which.

The CSV still has a use. It sizes the record: dates, tags, statuses and counts. It also names every requester with an email address, and the “Subject” column is usually what the customer typed. If an agreement asks for something to size the record first, the requester columns and the subject should be redacted or left out under the agreed rules.

The personal data in it

A support desk is full of people. The requester’s name and email address are on every ticket. The comments hold what customers wrote: order numbers, addresses, complaints, sometimes things they should never have typed into a form. The users export is a list of people. The staff who answered are named on every reply.

micro1 publishes how it treats that. “Personally identifiable information (PII) is removed when relevant, access is restricted to a limited number of authorized team members, and customer data is processed through isolated pipelines that prevent it from being commingled with other engagements.” It adds that “certain datasets may undergo synthetic rewrites that preserve workflow structure while reducing direct ties to original records.”

It also publishes the checks the company keeps. “Data is retained only for agreed periods, companies have the opportunity to review representative samples before use, and deletion requests are honored according to the terms of the engagement.” On ownership: “Companies retain ownership of their underlying data.”

What that means in practice, as our own advice:

Agree the redaction rules before any export. micro1 puts “Anonymization and redaction requirements” in the list it defines “Before any participation”.
Use the sample review. Ask to see representative samples of your own tickets, after redaction, before prepared data is used.
Read your own terms of service and privacy notice. What you told customers about their support messages is a question for your counsel before the scope is signed.
Don’t edit or delete tickets yourself. Zendesk’s CSV and JSON exports describe what is in the account; editing it changes the record. The redaction step exists for this.

What is never included

micro1 says every partnership has “clear controls over what is shared, how it is handled, and what is never included.” micro1 settles the list with each company in its scope step. It does not publish one fixed list for support data.

Here is what we tell owners to keep out, as our own advice, not micro1’s:

payment-card numbers, bank details and Social Security numbers that customers typed into tickets;
health information, and tickets from a health or benefits line;
tickets about minors, and tickets from minors;
attorney-client threads and anything escalated to your lawyer;
tickets for a client whose contract says its support data stays its own;
internal HR tickets, if the same Zendesk account handles staff requests.

Two Zendesk points sit beside that list. If your team writes internal notes on tickets, those notes are often the frankest part of the record; decide with micro1 whether they are in scope. And a JSON users export, or a user XML export, is a list of people with no work in it. A scope about how the team works rarely needs it.

Check the export

If an agreement asks for an export, check it before it goes anywhere. Five checks, as our own advice:

The format. Confirm it is the one the agreement names. A CSV where the scope says JSON is a different record.
The dates. Open the file and find the oldest and newest ticket. They should match the agreed range.
The counts. Compare the number of tickets in the file with your own count for the same range.
The big tickets. JSON leaves out the comments of tickets over 1 MB, and CSV leaves those tickets out altogether. Note how many there are, so the gap is known.
The link. Download within three days, keep the zip somewhere access-controlled, and note who downloaded it.

Other office systems

Support desks rarely stand alone. The mail behind the tickets sits in Google Workspace or Microsoft 365. The Google Workspace export guide covers Gmail, Drive and Chat. The Microsoft 365 export guide covers Outlook, SharePoint, OneDrive and Teams.

The records themselves are described elsewhere on our site. The email archives page goes through what an inbox shows about how the work was done; email tickets are the same conversations. The CRM records page covers the customer history a support desk sits beside. Both are written for trade companies; the shape of the record is the same in a support team.

Where we fit

We, TrustPast, explain the process and give you a range with our estimate. The estimate is ours, not an offer. You apply to micro1 yourself, through the button on this page, with counts, not files. micro1 pays us a referral fee. You pay us nothing. We never see your tickets or your exports. Why buyers want records like these at all, and what they leave out, is on what data buyers actually want from a trade business.

What a company of 30 or more holds, and why that headcount is micro1’s published line, is on the page for companies with 30 or more staff. What happens after the application is on what happens after you apply. How the records are protected is on how it’s safe.

Worth two minutes to see your range? Find the oldest ticket first.

TrustPast

Questions and answers

Do we have to export our tickets to apply to micro1?

No. micro1’s estimator asks how many employees and how many years of data; we would bring a few more counts to the first call: years of tickets, roughly how many tickets. micro1 says the approved data scope is defined together “Before any participation”. micro1 does not publish how records are delivered; that is agreed in its scope step. Export only after an agreement names it.

Why can’t I see the export option in Admin Center?

Zendesk switches exports off by default. Zendesk says “The account owner must contact Zendesk Customer Support to enable data exports in your account.” The exporting tools are also “not available on Team plans.”

Does the CSV export include what customers wrote?

No. Zendesk says the CSV data “does not include deleted tickets, ticket comments, or ticket descriptions.” The conversation is in the JSON export and the full XML export, and JSON leaves out the comments of any ticket over 1 MB.

Which format holds the most?

For tickets, JSON, which Zendesk recommends for accounts with more than 200,000 tickets. Full XML also carries tickets with comments, but its files stop at 500 MB, about 200,000 tickets, and it is not available above that.

How long is the download link valid?

Zendesk says “The download link is valid for at least three days.” If the email does not arrive, the “latest” link next to each export type downloads the most recent file.

Does the export include our help centre articles?

Zendesk’s export article lists tickets, users, organizations, account settings and groups. It does not list help centre articles. Count them separately; micro1 lists knowledge management among the customer-support records it looks for.