FINANCE

Agent-driven FX drift investigation and re-approval write-up

When a foreign invoice trips the FX drift threshold, an agent gathers the invoice, rate history, and counterparty context.

CategoryFinance
Enginepaperclip
Difficultyadvanced
Triggerwebhook
Steps6
Setup~25 min

How it runs

The automated pipeline, trigger to output.

  • TriggerDrift-threshold breach webhookHTTP webhook
  • ActionPull invoice and booked rateStripeStripe
  • ActionRetrieve currency-pair rate historyHTTP webhook
  • LogicAgent explains drift and recommends actionOpenAI
  • ActionLog narrative to audit fileAirtableAirtable
  • OutputPost decision-ready summary to approver in TeamsMicrosoft Teams

What it does

Instead of just flagging a number, this template has an agent investigate why a foreign invoice has drifted. It assembles the booked rate, the current rate, the trend over the invoice's life, and the customer context, then drafts a short narrative explaining the movement with a hold-or-confirm recommendation for the approver.

When to use it

Use it when raw variance percentages aren't enough and approvers want context before they re-approve or write down an invoice. Good for finance teams who want decisions documented with reasoning, not just a threshold breach.

How it works

  1. 1A webhook fires when an invoice breaches the drift threshold.
  2. 2The agent pulls the invoice and its booked rate from Stripe.
  3. 3The agent retrieves recent rate history and movement for the currency pair.
  4. 4It reasons over the data to explain the drift and form a recommendation.
  5. 5The narrative and recommendation are logged to Airtable for the audit file.
  6. 6A decision-ready summary posts to the approver in Teams.

Set it up

What you configure once, before turning it on.

  1. 1
    Connect StripeCustomers, subscriptions, payments.
  2. 2
    Connect AirtableBases, tables, views, automations.
  3. 3
    Connect Microsoft TeamsChannels, chats, files.
  4. 4
    Connect HTTP webhookTrigger any URL on agent actions.
  5. 5
    Connect OpenAIModels, embeddings, files.
  6. 6
    Set each agent's modelWe leave models unset so you pick the tier — fast + cheap, or top-quality.
  7. 7
    Tune it to your dataEdit the prompts, filters, and field mappings so it matches how your team works.
  8. 8
    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.