Insights
How do I allocate AI spend in Microsoft Dynamics 365 Business Central?
To allocate AI spend in Business Central, define Dimensions for Department, Project, and (optionally) Agent, post metered usage as a General Journal with dimension values on each line, use Allocation Keys for shared agents whose cost is split by a fixed percentage, and enter the provider invoice against the accrual so the true-up lands on the same dimensions. Business Central's dimension model fits agent spend better than most ledgers because every line can carry several attributes without a chart-of-accounts explosion.
Set up once
- G/L accounts. An expense account "AI Services" (with sub-accounts if needed) and a liability "Accrued AI Services."
- Dimensions. You already have Department. Add a dimension "AI Agent" with one value per agent, and make it a shortcut dimension so it appears as a column in journals. If agents do project work, Project (Job) is the third dimension.
- Allocation Keys. For shared agents, create an allocation key that splits by percentage across departments (e.g., Finance 40, HR 35, Legal 25). The key can be adjusted monthly from request counts. Allocation Keys split the general journal accrual; for the purchase invoice, set up an Allocation Account with the same percentages, since that is where Business Central splits a single invoice line.
- Vendors. One per provider. For Microsoft Copilot Studio or Azure OpenAI, the charge arrives inside the Microsoft or Azure invoice; the allocation rule below applies to that line.
- Recurring General Journal. Set up the accrual as a recurring journal with the "Reversing" method so it reverses automatically on the first of the next month.
The monthly cycle
- Month end. Post the recurring accrual journal: AI Services debited per Department and Agent dimension, Accrued AI Services credited.
- Invoice receipt. Post the purchase invoice to AI Services with the same dimensions. For a shared agent, post one line to the Allocation Account, which splits it across departments by the same percentages.
- True-up. The reversed accrual and the actual invoice net to the variance, which stays on the same dimensions.
Worked example
September metered usage $6,400: Sales agent $3,000 (Department: Sales), Support agent $2,200 (Department: Service), Shared assistant $1,200 (allocation key: Finance 40 / HR 35 / Legal 25).
Recurring General Journal, September 30, reversing
| Account | Debit | Credit | Department | AI Agent |
|---|---|---|---|---|
| AI Services | 3,000 | Sales | SALES-AGENT | |
| AI Services | 2,200 | Service | SUPPORT-AGENT | |
| AI Services | 480 | Finance | SHARED-ASSIST | |
| AI Services | 420 | HR | SHARED-ASSIST | |
| AI Services | 300 | Legal | SHARED-ASSIST | |
| Accrued AI Services | 6,400 |
Purchase Invoice, October 6, $6,480: same lines scaled by 6,480 / 6,400; the $80 variance follows the same dimensions.
Copilot Studio and Azure OpenAI
Copilot Studio bills by message or credit packs; Azure OpenAI bills by token inside the Azure subscription. Both arrive as lines on a Microsoft invoice with no department information. The allocation has to come from your own metering (Copilot Studio analytics, Azure cost exports tagged by resource group, or a gateway). Post the Microsoft invoice line to AI Services with an Allocation Account built from that metering, and reconcile its percentages monthly.
Reports
Analysis Views over AI Services by Department and AI Agent give the per-agent cost trend; Account Schedules (Financial Reports) can show AI spend as a line under each department's operating expenses.
Related
Copilot Studio cost management for finance teams · How do I do chargeback for AI usage? · What is the journal entry for AI token usage?
Feature names reflect Business Central as of this writing. Dimension design should be agreed with your controller and BC partner.
Not sure which of your AI costs are being booked? Run the free Agent Spend Assessment.
Onaro Meridian is FinOps for agentic AI: the system of record that attributes, controls and books what AI agents spend.

Brian Diamond
Brian Diamond is a fractional Chief AI Officer and founder of Onaro. He has spent 30 years running infrastructure operations and founded LANStatus, a Connecticut managed services provider and Microsoft partner, in 2001. He holds a Chief AI Officer certification and writes the CAIO Brief on AI leadership for finance and operations.