AI agents for the close and the plan
An agent that cannot write back to your model is a demo. PlaniFi AI Agents run real process against your live SAP Analytics Cloud model, versioned and auditable like any human action. This is the third of four places AI has to land, and it is the one that closes the gap while SAP’s roadmap catches up.
What one of them actually does
Take budget entry. Someone types what they want in plain language. The agent validates it against the live model, distributes it to the leaf members while respecting your hierarchy, writes it back through real SAC APIs, and hands back the job ID so the action can be traced. It is OAuth-secured and it runs today.
Real SAC APIs. Not a widget, and not a screen-scrape.
That distinction is the whole thing. Most agent demos read your data and produce a sentence. These take an instruction and change a number in a system your auditors look at, which is why the write path matters more than the language model.
The agents
Project Worker Bee, the delivery accelerator, is an agent too. It has its own page, because what it accelerates is the project rather than the process.
Governed like a human action
Every agent action is versioned and auditable. Nothing writes to production without the same trail a person would leave. That is not a feature we added late. It is the reason a CFO can let one of these near a close at all.
Almost nobody is doing this yet
McKinsey found 23% of organizations actively scaling an agentic AI system in at least one business function, with another 39% experimenting. But in any given function, no more than about 10% are scaling agents, and roughly two-thirds have not begun scaling AI across the enterprise at all.
Read that as the opportunity it is. This is early, and being early in the close is worth more than being early almost anywhere else in the business.
See the budget-entry agent run Month-end close agent
Where this sits
See it run against a live model
It is a short demo against a real SAC tenant, not a slide.