TemplatesGet a demo →Book a meeting

A period reconciled employer by employer, exceptions only

Enter a period and the flow matches bank receipts against the contribution schedule, keeps only employers that don’t reconcile, and lists them largest difference first. Owed and overpaid are labeled plainly, and every difference is the database’s arithmetic, not a model’s.

Finance · Pensions & retirement · Someone asks · 5 steps · Your database

AI Factory · Payment reconciliationthe flow as it opens in the builder
QuestionWhich period?
Query a databaseMatch receipts to what was due
Work out the numbersKeep what does not match
AnswerList the exceptions
Answer shownThe answer
5 steps in the flowDrag the canvas, or any step, to move it around.

Why teams run it

Matching bank receipts to expected contributions is still a spreadsheet exercise at many funds, repeated for every employer and every period. It is slow, it depends on whoever built the spreadsheet, and one sorting mistake can hide a missing payment.

How it is governed

  • Read-only: SELECT queries only, through an administrator’s connection.
  • The period is bound as a value, so input cannot alter the query.
  • Differences are computed by the database and table step, never by the model.
  • It reports and posts no adjustment to any ledger.
Who runs it
Pension accountants; Contributions and finance teams; Finance controllers.
What you supply
A read-only database connection holding both the bank receipts and the contribution schedule.

It ships with test questions

  • 2026-08
What “a person approves” looks like in the product: the run stops, the banner says what approving will do, and until someone decides, nothing leaves the system. A demo workspace on synthetic data.

From template to a live agent

The same builder, tests and release review apply to a template as to anything built from scratch.

Open it as a working flow
Every step, source, branch and approval is already in place, and every one of them can be changed before it runs on your data.
Prove it before it ships
An evaluation set can gate the release, so a change that breaks an answer never reaches anyone. A question it cannot answer from the sources leaves the gate unproven, not passed.
Keep a person on the decisions
A run that needs a decision stops and says exactly what approving will do, and who it is waiting for.
Publish where people work
A form, an embedded widget, Microsoft Teams or a scheduled batch — with its own access and limits per deployment.
What teams usually change
Map the query to your own receipts and schedule tables and references; Ignore differences under a tolerance by adding a filter condition; Put it on a schedule to run after month-end.
Step 02, in the product: an evaluation set that gates the release. A question it cannot answer from the sources leaves the gate unproven, not passed. A demo workspace on synthetic data.

Questions buyers ask

What does it need to connect to?
A read-only PostgreSQL connection that holds both your bank receipts and the contribution schedule, with the query rewritten for your tables. The database host must be on your approved list before a connection opens.
How does it treat overpayments?
A positive difference is reported as money still owed and a negative one as an overpayment, exactly as the database computed it. The model is instructed never to recalculate either.
Does it correct the ledger?
No. The connection is read-only and the flow produces a report; any adjustment is made by your finance team in your own systems.

See it run on your own data

We open it in the builder, point it at a sample of your documents and systems, and run it with you.