AUDIPY/Playbooks

AI Playbooks.Auditing by plan.

Standardised analysis workflows, open to any AI agent.

AMY sets off a chain reaction
Define once. Run the same way every time.

With Playbooks you define once which steps AUDIPY and AMY should perform, then run the workflow at the push of a button with the AI agent of your choice. Through the open MCP standard, Playbooks work with any MCP-capable AI agent, such as Claude Code, Claude Cowork or OpenAI Codex.

Whether it is an audit of annual financial statements, a compliance audit or a travel expense analysis: Playbooks record the audit path in readable form and execute it exactly the same way every time, with no differences in how individual users operate the software.

New colleagues follow the same audit path from day one, and recurring audits run without having to learn the process again.

Travel expense compliance 2025 · with your AI agent
# source
sap_export.csv · filter: posting_date >= 01/01/2025
# audit steps
1. flag weekend and public holiday postings
2. list amounts > €1,000 without receipt
3. find duplicate receipt numbers
4. review cash payments
Run the playbook travel-expense-compliance-2025
AMYStep 1/4 ✓ 18 weekend postings flagged
Step 2/4 ✓ 7 amounts without receipt
Step 3/4 ✓ 3 duplicate receipt numbers
Step 4/4 ✓ 12 cash payments reviewed
AMYFinal report: 40 anomalies in 8 seconds. Report sent to the audit team's inbox.
How it works
Three steps.

Define

Once, in plain language instead of a chain of clicks: which data, which filters, which reports.

Version

Stored centrally, traceable and shareable. Every authorised user can use the workflow.

Execute

At the push of a button with your AI agent. AMY reports back on what happened at each step.

Connected via MCP
Your agent, your choice.

AUDIPY provides its functions via the open MCP standard (Model Context Protocol). An AI agent such as Claude Code, Claude Cowork or OpenAI Codex can use it to load data, run analysis steps and work through Playbooks while AUDIPY runs in the background.

The agent works through the playbook step by step, and every step remains traceable.

The MCP connection is included in all AUDIPY packages. You bring your own AI agent and its access.
AI agentClaude CodeClaude CoworkOpenAI Codexother MCP agents
MCP
AUDIPY
  • List files
  • Read data structure
  • Run analysis step
  • Create report
Why Playbooks
Routine becomes knowledge.

Use again and again

Defined once, repeated as often as you like. No matter who starts it, the result is reproducibly the same.

Documented and traceable

All steps remain readable, no black box. This simplifies quality assurance, handovers and further development.

Open to any agent

Thanks to MCP, you are not tied to a single agent. You decide which agent runs your Playbooks.

Insights
From audit step to playbook.

Automate your audit path.

On request, we show you online in English, free of charge, how your workflow becomes a playbook.