Stripe, Microsoft, and BlackLine Put Agent Permissions Into Production
AI agents are entering finance systems with real permissions, and the winners will be the teams that can control access, approvals, and auditability.
What is this trend?
Finance vendors are moving AI agents into live ERP and treasury workflows with granular permissions, approvals, and audit trails, making control design the key enabler of automation.
- Agents are now executing finance actions, not just drafting analysis.
- Permissions, expirations, and approval gates are becoming standard controls.
- Close, reconciliation, and treasury workflows are the first production use cases.
- Audit trails and confidence scores decide what can auto-run vs. needs review.
- Teams that govern agent access well can scale automation without losing control.
What’s the latest?
Stripe’s rollout this week is the clearest sign that finance execution is moving into production agent workflows, not just analysis.
How it developed
- Controlled finance workflows, continuous refinancing risk, and sharper treasury readiness
- Liquidity Tightens, Refinancing Speeds Up, and Treasury Becomes a Real-Time Command Center
- Agentic Finance Tightens Controls, Credit Risk Goes Structural, and Climate Stress Testing Goes Supervisory
- Agent Permissions Go Live, Collateral Quality Tightens Refinancing, and Connected Planning Reshapes FP&A
Go deeper
Curated long-form picks on this trend — podcasts, videos, and analysis, by seniority.
If you're an individual contributor

Observability, Cost, and Efficiency Practices for AI Agents
How-to podcast on running AI agents securely at scale, covering observability, guardrails, and cost metrics amid tighter rules.
The Everything Feed - All Packet Pushers Pods · Podcast
Listen from 22:42 →
Building Agent Capability From Real-World Practice Scenarios
Substack explainer on agent training via practice & assurance—mapping real finance workflows into testable layers.
Shift*Academy · Substack
Read →
Essential Guides for Developing and Governing AI Agents
How-to weekly digest on agent observability and governance, aligning AI coding agents with tightening bank rules.
Creators' AI · Substack
Read →If you manage a team

AI Copilots Speed Up Steps But Keep the Bottleneck
An opinion essay contrasting copilots with agents, with a refund-process case study on automating routine workflows.
Innovation Unpacked · Substack
Read →The Work Between the Systems: Where AI in the Close Needs a Reviewer and a Record
News analysis featuring Adam Riccoboni on AI finance-close workflow audits and human review under tighter agentic rules.
Cpapracticeadvisor News · News
Read →
AI‑Enhanced Editing and Curation Standards for Better Decision Briefs
How-to Substack guide on building a “company of curators” for supervised AI review amid tighter agentic finance rules.
Operating by John Brewton · Substack
Read →If you lead the organization
The automation paradox in finance
News analysis on CFO AI governance: why legacy oversight fails and how continuous monitoring layers manage risk.
Accounting Today · News
Read →
Beware the black box: Finance puts AI to the audit test
Promotion news feature with Jeremy Ung on AI audit governance for automated finance workflow layers.
AFR · News
Read →How Organizations Can Build Audit-Ready AI Governance in Revenue Cycle Management
Explainer on audit-ready AI governance for revenue cycle automation, aligning controlled workflow layers with risk traceability.
Nasscom · News
Read →