How we handle GDPR data-subject requests on a deadline
Legal
The finance agent we run on Kortix — connected to Stripe, the bank feed, and Google Sheets. Each month it matches transactions against invoices, flags mismatches, and posts a summary, escalating anything it can't match.
Monthly reconciliation is the kind of task that's simple to describe and tedious to do: line up what came in and went out against what was supposed to, find the things that don't match, and explain them. On a small team it lands on one person for an afternoon, and it's the same afternoon every month.
We handle this with a scheduled agent. Once a month it reconciles transactions against invoices, flags what doesn't line up, and posts a summary. This writes up how we run that on Kortix — the connections, the steps, and the guardrails.
Reconciliation is mechanical until it isn't. Most lines match cleanly, and the value is entirely in the handful that don't: a payment that never landed, a fee that doesn't tie out, an invoice with no matching deposit. Finding those means going through everything, which is slow and easy to get wrong when it's the same repetitive comparison hundreds of times.
A spreadsheet formula catches the exact matches but not the near ones, and the near ones are where the problems hide. So the check either runs shallow and misses things, or runs deep and eats a person's day. Either way it only happens as often as someone has time for it.
On Kortix, a monthly cron triggers a reconciliation agent. Each run spawns its own isolated session — a cloud sandbox — with scoped access to what reconciliation needs: Stripe, the bank feed, and the Google Sheet we track in. It matches transactions against invoices, flags the mismatches, posts a summary to the sheet, and escalates anything it can't match to a person.
Open your Kortix workspace, start a new session, paste this prompt and send it. The agent writes the agent file, the reconciliation-match skill and the monthly trigger, and opens a change request for you to review. Prefer clicks? Choose Use this template instead.
Set up a monthly expense reconciliation agent in this workspace.
Create the agent agents/expense-reconciliation.md. Its job is to match Stripe activity and the bank feed against the invoices in our reconciliation sheet. It may use the Stripe, Plaid and Google Sheets connectors and the PLAID_ACCESS_TOKEN secret.
Create the skill skills/reconciliation-match/SKILL.md with these rules:
Add a cron trigger monthly-reconciliation to kortix.yaml. It runs the expense-reconciliation agent on the 2nd of every month at 06:00 with the prompt "Load the reconciliation-match skill and reconcile the period since the last dated entry in the reconciliation sheet."
Tell me which of these connectors are not connected yet.
Open a change request with all of it. Do not merge it.
Open Connectors in your workspace and connect Stripe, Plaid and Google
Sheets. Complete the Plaid Link step once for the bank account and store the
result as the PLAID_ACCESS_TOKEN secret. Credentials are brokered
server-side and never enter the agent's machine.
Reads charges, payouts and refunds to match against invoices.
Reads bank transactions to match against the ledger.

Reads the invoice ledger and marks matched and open lines.
Merge the change request, then start a session with the expense-reconciliation agent and send this prompt. Check the matches and the escalated rows before you let the schedule take over.
Reconcile last month in the reconciliation sheet.
The change request added this trigger to kortix.yaml. From now on the agent
runs on the 2nd of every month at 06:00, after the prior period closes, each run in a new session on its own cloud computer.
To change the time, edit the cron line, or open Triggers in the workspace.
triggers:- slug: monthly-reconciliation name: Monthly expense reconciliation on: cron: "0 0 6 2 * *" run: agent: expense-reconciliation prompt: Load the reconciliation-match skill and reconcile the period since the last dated entry in the reconciliation sheet.The agent reads Stripe and the bank feed and writes only to the sheet. It never issues a refund or starts a transfer. A line it cannot match within tolerance gets its own escalated row with both records attached, so a person can chase it without pulling the data again. It never rounds off or writes off a difference to make the sheet balance.
Each run ends with a new dated entry in the sheet. This is what a run looks like in the session:
Reconciled 312 lines for the period. 291 tied out exactly and 18 within tolerance, mostly Stripe fees. Escalated 3 new mismatches, including one payout with no matching deposit.
Once it works, paste any of these into the same session.
When a run escalates any mismatch, post the count and a link to the sheet in #finance.
On the 16th of each month, run a read-only reconciliation of the first half of the month and list mismatches without writing to the sheet.
When the same unexplained difference appears three months in a row, propose an explained-mismatch note for me to approve.
Giving an agent access to financial systems is a security question first. The relevant controls on Kortix:
The afternoon that used to go to comparing hundreds of lines now goes to resolving the few the agent couldn't match. The reconciliation runs every month whether or not anyone has time, and the summary is waiting when the finance person opens the sheet.
The setup relies on four pieces: sandbox isolation per run, a secrets manager to broker the Stripe, bank, and Sheets tokens, an escalation path for anything unmatched, and memory that carries recurring explanations forward month to month.