AddThisFeature

Freshdesk Ticket Creation

Send support issues to Freshdesk with the context an agent needs to resolve them.

moderate Sales & Marketing Tools

What it adds

A Freshdesk delivery path that maps in-app support requests onto the helpdesk's requester, group, product, and field structure.

What your agent is told to do

5
  1. 1

    Build an operator-editable mapping from the app's request types onto requester, group, product, source, status, priority, and any custom fields. Field identifiers change whenever the helpdesk is reconfigured, so read them from configuration rather than embedding them in code.

  2. 2

    Look up the contact by the session's address before creating one, and store the returned contact identifier on the app's user record so every later ticket reuses it instead of spawning a fresh contact each time.

  3. 3

    Validate attachments against the provider's size and type limits in the browser and again on the server, and tell the user exactly what was rejected before the ticket is submitted.

  4. 4

    Carry a request identifier generated before the first attempt and check for an existing ticket bearing it before creating one, so a timeout that actually succeeded does not file a duplicate.

  5. 5

    If Zendesk or Help Scout briefs are also applied to this app, one shared submission path owns the form, context capture, attachments, and deduplication. This brief owns only the Freshdesk mapping and delivery.

Edge cases it handles

8
  • Creating a new contact on every submission is the default failure mode here. Resolve and store the contact identifier, and reconcile it when the provider merges two contacts behind the app's back.
  • A group, product, or custom field that has been renamed or removed must fail into an operator alert with the ticket still queued, rather than being dropped or filed into whatever queue happens to accept it.
  • Files over the size limit or of a rejected type must be reported before submission, with the option to send the ticket without them rather than losing the whole request.
  • A submission retried after a timeout must be deduplicated on the request identifier, because the first attempt may have already filed the ticket.
  • Raw provider errors stay in the app's logs. The user sees a plain result and a reference they can quote, and agents do not receive the app's stack traces either.
  • When the provider is unavailable the request must be accepted and queued with the user told it was received. Failing a help request when the user is already stuck is the worst possible moment to fail.
  • Automatically attached context must exclude credentials, tokens, and full request bodies, which would otherwise be readable by every agent on the account.
  • The ticket identifier must be stored against the app record so status can be shown and replies threaded correctly.

Definition of done

9
  • Request types map onto requester, group, product, source, status, and custom fields through configuration rather than code.
  • Repeat submissions from the same user reuse a single provider contact.
  • Attachment limits are enforced before submission and every rejection is explained to the user.
  • A submission retried after a timeout yields exactly one ticket.
  • Provider errors are logged internally while the user sees a plain outcome and a reference.
  • Provider downtime queues the request instead of failing the submission.
  • The ticket reference is stored against the app record.
  • The feature matches the existing design system.
  • No existing functionality is broken.

Related features

How it works

  1. 1

    Copy the link

    Grab the Markdown instruction URL for this feature.

  2. 2

    Give it to your AI

    Paste it into Claude Code, Cursor, v0, Lovable — whatever you build with.

  3. 3

    It inspects, then implements

    Your agent reads your existing app first, then adds the feature to fit it.

Works with your stack

These instructions are written to adapt. They tell the agent to detect your framework, match your existing design system, and reuse what you already have — rather than assuming a particular stack.

Need it tighter than that? Customize the feature and tell it exactly what you're running.