Meet the team that works while you sleep.
The FP&Agents are Cube's autonomous AI workforce for finance: four growing agent teams, orchestrated by one super agent, with read and write access for anything you build.
The problem
The problem one assistant can't solve.
Up to 80% of the week is still assembly, and every finance leader knows the 11pm feeling of stitching tool outputs together the night before the board meeting.
FP&A is a dozen different jobs
Mapping, forecasting, variance, board decks. Each needs different context, and a general assistant is shallow at all of them.
You became the router
Someone splits the work, prompts each tool, stitches results back, and checks it all ties. Coordination is the new manual work.
Handoffs drop the numbers
Data to analysis to deck, each step in a different tool. Figures drift, and nothing ties at the end.
Charlie & the teams
The most advanced AI in FP&A.
Ask anything about your financials. Charlie plans the work and puts the right team on it.
Plans the work, orchestrates every team, and returns the decision-ready answer.
What are FP&Agents?
FP&Agents are Cube's autonomous AI workforce for FP&A. They take on whole jobs (reconciling the data, refreshing the forecast, explaining the variance, drafting the board deck) and carry them to a finished, decision-ready answer. Charlie, the super agent, directs the work, and every figure traces to the source transaction.
Charlie
The super agent orchestrating every team.
- Plans the workBreaks your question into tasks
- Routes to the teamsPuts the right specialists on each piece
- The agents run itPull, reconcile, model, and draft, autonomously
- Checks the answerEvery figure verified against the source
Two ways in
Run our agents, or build your own.
Either way, the work runs on the same clean, decision-ready data, and every answer traces to the source.
Use the FP&Agents
Charlie and the four teams come with the platform and run the workflow from data to boardroom, ready on day one.
Build on the Cube MCP Server
Cube's APIs give your agents read and write: pull decision-ready figures out, push plans and updates back, all governed and logged. Building on ungoverned data is a faster way to build the wrong thing.
The Agent Exchange
Proven plays, ready to run.
Browse the exchange, pick a play, and Charlie puts the right team on it. Edit anything before it runs.
Variance bridge to budget
Refresh the rolling forecast
Draft the investor update
Map the new entity
+ dozens more plays in the exchange, across every team
The Analyst Agents at work
Variance, explained before the meeting starts.
Smart Variance ranks the quarter's variances by amount and percent, flags what moved, and every figure opens to the transactions behind it.
Why it holds up
Analysis gets the week back.
The assembly work, handled
Agents pull the data, reconcile it, and build the first draft. Your team starts at the review.
Built for finance, at every endpoint
Purpose-built agents that know your chart of accounts and business logic, working in the tools you already use.
Trace to the source
Every agent answer maps back to the specific GL transaction, so you can defend it in the room.
Use cases
Where teams put the agents to work.
Variance analysis
Every line over threshold flagged, with timing separated from structure.
Rolling forecasts
Refreshed with the latest actuals the moment assumptions change.
Board reporting
This quarter's board financials drafted from the same traced numbers.
Month-end close
Ask what's blocking close and get the punch list, already assigned.
Headcount planning
People plans and financial targets kept in the same conversation.
Ad hoc answers
Every department has questions. Finance has answers, with sources.
See it on your own workflow.
Pick the use case that eats your week and watch the agents run it end to end.
Security & governance
AI your auditor can live with.
Agents work inside the same controls as your team. Read the security overview for the full posture.
Cell-level access control
The CFO sees everything. A department head sees their cost center. Agents follow the same rules as the person asking.
Every action logged
Agent activity lands in the audit trail alongside human edits.
Trace to the source
Every agent-generated figure maps back to specific GL transactions.
Finance stays in control
Agents propose changes to mappings and models. Your team reviews and approves before anything lands.
Trace to the source
Every number opens.
Any figure an agent hands you drills to the transactions behind it.
| Jun 04 | AWS | usage · June | NetSuite | $92,410 |
| Jun 12 | Datadog | annual true-up | NetSuite | $38,750 |
| Jun 27 | Snowflake | compute · June | NetSuite | $27,900 |
| + 38 more transactions | ||||
The heavy lifting is Charlie's job now.
Book a demo and run a real FP&A question through the agent teams in a live sandbox.
Prefer to look around first? Explore the platform.
