Epicor Kinetic and Jira integration

What moves between them

The flow runs from Epicor Kinetic into Jira. ml-connector polls Epicor Kinetic at a schedule tied to your weekly or biweekly PO cycle, detecting new and updated purchase orders and supplier invoices. For each record, ml-connector creates a Jira issue in the target project with the PO number or invoice number as the issue key, the supplier name and amount as description and custom fields, and a link back to Epicor Kinetic. If the PO or invoice is updated in Epicor Kinetic, ml-connector updates the corresponding Jira issue. Jira issues are read-write, so your team can add notes, change status, and attach documents without affecting the source data in Epicor Kinetic.

How ml-connector handles it

ml-connector stores the Epicor Kinetic OAuth2 client credentials and Jira OAuth2 tokens encrypted, refreshing them on every use. Polling queries Epicor Kinetic's OData endpoint with a $filter on UpdatedOn to fetch only records changed since the last run, and it respects the Company segment requirement in the Epicor Kinetic API URL. For each new PO or invoice, ml-connector maps the supplier name to the issue summary, the PO total or invoice amount to custom fields, and the Epicor Kinetic record ID to a link so Jira users can trace back to the source. Jira webhook registration is automatic, and ml-connector tracks webhook refresh dates so expiry does not cause outages. If Epicor Kinetic returns no new records on a poll, the run completes cleanly without creating duplicate issues. If a Jira custom field is missing, ml-connector logs the error and surfaces it in the audit trail so you can add the field and replay the job.

A real-world example

A medium-sized distributor manages inventory and supplier orders in Epicor Kinetic and runs operations projects in Jira. Before the integration, the procurement team created a Jira issue for each major PO by hand, often days late, and the operations lead chased email and Slack to keep project status in sync with procurement status. With Epicor Kinetic and Jira connected, new purchase orders appear as Jira issues within minutes of creation in Epicor Kinetic, linked to the operations project and labeled by supplier. The operations team now sees inbound orders, shipment dates, and invoice status in Jira without context-switching, and finance can comment on issues with notes that never get lost in email.

What you can do

• •Poll Epicor Kinetic on a schedule to detect new and updated purchase orders and supplier invoices.
• •Create Jira issues in the target project for each new PO or invoice, with supplier name, order total, and link back to Epicor Kinetic.
• •Update Jira issues when purchase orders or invoices change status or amount in Epicor Kinetic.
• •Refresh Jira webhooks and OAuth tokens automatically before expiry, with no manual intervention.
• •Track every poll, issue create, and update in a full audit trail, and replay failed records if Jira custom fields are added later.

Questions

Does data flow both ways between Epicor Kinetic and Jira?

No. Data flows one way, from Epicor Kinetic into Jira. Jira issues created by ml-connector can be edited by your team (adding notes, changing status, attaching documents), but those changes do not flow back into Epicor Kinetic. Epicor Kinetic is the source of truth for PO and invoice amounts, supplier names, and dates.

How does ml-connector handle Epicor Kinetic's lack of native webhooks?

ml-connector polls Epicor Kinetic on a schedule you set (e.g., daily or every 4 hours) using OData $filter queries to fetch only records updated since the last poll. This avoids constant API calls and works with both cloud and on-premises Epicor Kinetic instances. The Company segment requirement in the Epicor Kinetic URL is handled per customer.

What happens if a Jira custom field is not present or has the wrong type?

ml-connector logs the error and includes it in the audit trail with the field name and issue ID. You can then add or fix the custom field in Jira and use the audit trail to replay the failed records without re-running the entire poll.