INVOICE PROCESSING

Catch mis-coded invoice lines before they post

Reviews newly-coded invoice lines against department budget patterns in BigQuery, flags GL codes that look wrong for the vendor or amount.

CategoryInvoice Processing
Enginesim
Difficultyadvanced
Triggerwebhook
Steps5
Setup~25 min

How it runs

The automated pipeline, trigger to output.

  • TriggerCoded invoice event via webhookHTTP webhook
  • ActionPull vendor spend pattern from BigQueryGoogle BigQueryBigQuery
  • ActionScore each line for coding anomaliesOpenAI
  • LogicRoute clean vs anomalous invoices
  • OutputOpen Linear issue for flagged linesLinearLinear

What it does

Acts as a second set of eyes on GL coding. When an invoice is coded, it checks each line against historical spend patterns by vendor, GL account, and amount range in BigQuery, and surfaces lines that deviate sharply from the norm.

When to use it

Use it when coding is already happening (manually or via another workflow) and you want a control that catches expense-misclassification and wrong-cost-center errors before the GL closes. Aimed at controllers and FP&A.

How it works

  1. 1A coded-invoice event (webhook) triggers the run.
  2. 2A BigQuery query returns the typical GL account, amount distribution, and cost center for that vendor.
  3. 3An OpenAI step compares the submitted coding to the pattern and scores each line for anomaly risk with a short rationale.
  4. 4A logic step passes clean invoices through and routes anomalous ones onward.
  5. 5For flagged invoices, a Linear issue is opened for the controller with the suspect lines and suggested correct codes.

Set it up

What you configure once, before turning it on.

  1. 1
    Connect HTTP webhookTrigger any URL on agent actions.
  2. 2
    Connect BigQueryDatasets, queries, schemas.
  3. 3
    Connect OpenAIModels, embeddings, files.
  4. 4
    Connect LinearIssues, projects, cycles, triage.
  5. 5
    Set each agent's modelWe leave models unset so you pick the tier — fast + cheap, or top-quality.
  6. 6
    Tune it to your dataEdit the prompts, filters, and field mappings so it matches how your team works.
  7. 7
    Test, then turn it onRun once against a sample, confirm the output, then enable the trigger.

Run this workflow in your colony.

14-day trial. No DevOps. No Sales call. Provisioned in under a minute.