Drafts before posting
Spark prepares work for review. It does not post, pay, or authorise on its own.
Security
Not a black box. A glass box.
Spark is built for finance work where accuracy, review, and accountability matter. It prepares the work, shows the working, and waits for sign-off.
Named approvers
Review decisions are attached to people, not a vague system action.
Workpapers included
Journals and schedules include supporting calculations so reviewers can follow the working.
OAuth connections
Xero access uses OAuth. Access can be revoked from the connected platform.
Audit trail
Requests, drafts, approvals, and downstream changes are logged for later review.
SOC 2 in progress
We are building toward SOC 2 while keeping the product approval-first from the start.
Approval model
Finance decisions stay with finance people.
Spark can extract invoice data, suggest coding, assemble schedules, and draft Xero entries. The reviewer still sees what changed, why it was suggested, and what will happen next.
That keeps the system useful during month-end without weakening segregation of duties or creating mystery postings in the ledger.
Read privacy policyFAQ
Security questions, answered.
Straight answers on data handling, OAuth, model training, and review controls.
Does Spark store my financial data?+
Spark is designed to keep Xero as the system of record. It processes data needed to prepare drafts, workpapers, and audit logs without becoming a second ledger.
Can access be revoked?+
Yes. Spark connects to Xero through OAuth, and access can be revoked from the connected platform.
Is customer data used to train models?+
No. Customer finance data is processed to complete the requested work and is not used to train general models.