For agencies and multi-product teams
How do we split our AI spend by project or client?
Short answer: Give each project or client its own API project or workspace and key, set a spend limit on each, and tag agent sessions with a project id in your own logs. Subscriptions are easier to budget but rarely show which project used what, so measure the heavy work on keys you can attribute. Start with the $150 burn-audit hour.
Why the bill can’t answer it
One invoice covers everything. If several applications or agents share one key, the total for that key can’t say which one made a request; OpenAI’s usage dashboard help says as much and suggests comparing your own request records. Flat subscriptions are predictable, but they usually report usage per person, not per project.
Use the structure the vendors already give you
- OpenAI API projects. Usage can be broken down by project, and each project can have its own monthly spend limit and alert. A hard limit stops requests once tracked spend reaches it (OpenAI’s guide to API projects, OpenAI’s spend limits guide).
- Anthropic Workspaces. Each workspace can have its own spend limit and rate limits, set below the organization’s, and usage reporting carries the workspace id. The Default Workspace can’t have a spend limit (Anthropic’s Workspaces docs).
- One key per project or client. Keys are the unit most reports can split by, so don’t share one across accounts.
Start with the $150 burn-audit hour. One hour with an operator. You leave with a written map of the jobs, tokens and flows to cut or rebuild. Pay the hour, then we schedule the session. See a sample audit (example data) or try the free burn check.
Book the hour — $150 Operator pack — $625 depositAI Concierge — $2,000/mo
The pack is $1,250 for 10 hours; checkout takes the $625 deposit. Questions first? Email daniel@agenthiveinc.com or call our AI front desk demo at (320) 335-6186.
When sessions touch several projects
Agents often work across repos or clients in one session. Two habits make that measurable: start each task with a project tag the agent writes into its log, and record tokens per call rather than per session so a mixed session can be split afterwards. It won’t be exact, but it is good enough to rank projects and spot the one that costs more than it earns. AWS Well-Architected Agentic AI Lens recommends the same idea for agent workloads: attribute cost to the work that caused it and look at cost per workflow.
The three ways to work with us
- $150 burn-audit hour. One working hour. We map the agents, routines and flows your team runs and hand back a written list of what to cut, fix or rebuild.
- $1,250 operator pack ($625 deposit). Ten hours of build and fix work on the flows from the audit. The balance is due as the work is scheduled.
- AI Concierge, $2,000/mo. A senior AI operator on retainer: two 45-minute working sessions a month, unlimited async help between them, and a shared asset inventory. Month to month; cancel any month after the first 30 days.
Common questions
Can we bill clients for their AI usage this way?
You can see usage per project or workspace, which gives you a basis. Whether and how you pass it on is a contract question for you and the client.
We only use subscriptions. Is this still worth doing?
Yes for the heavy jobs. Move the work you most need to attribute onto keys you can split, and keep subscriptions for general use.
Do you need access to our billing?
Read-only access or exports are enough for the audit. Nothing changes unless you approve it.
Start with the $150 burn-audit hour. One hour with an operator. You leave with a written map of the jobs, tokens and flows to cut or rebuild. Pay the hour, then we schedule the session. See a sample audit (example data) or try the free burn check.
Book the hour — $150 Operator pack — $625 depositAI Concierge — $2,000/mo
The pack is $1,250 for 10 hours; checkout takes the $625 deposit. Questions first? Email daniel@agenthiveinc.com or call our AI front desk demo at (320) 335-6186.
Related
- How do we measure what each AI agent workflow actually costs?
- An AI operator for agencies that already pay for AI
- What spend controls should we put on our AI agents?
- Free AI agent burn check
Outside facts link to their sources and were checked in November 2026; product pricing and limits change, so confirm on the vendor’s page. First Deploy is built and run by Daniel Graham at AgentHive Inc, Palm Coast, Florida. We don’t promise a savings figure before we’ve seen your setup.