Episode Overview

Podcast: The Audit Podcast
Episode: IA on AI – Claude Cowork Walkthrough w/ Brian Kuenzi
Host: Trent Russell
Guest: Brian Kuenzi
Date: February 26, 2026

This episode dives deep into Anthropic’s new Claude Cowork feature and its relevance for finance, accounting, and internal audit professionals. Trent Russell and guest AI expert Brian Kuenzi walk through the latest Anthropic AI offerings, focusing on real-world use cases for internal audit, the emergence of plugins and agent-based automation, and practical advice for implementing these tools securely and effectively in audit workflows.

Key Discussion Points & Insights

1. Introduction to Claude Cowork & Plugins

• What is Claude Cowork?
  • Claude Cowork is a new desktop application/UI for Anthropic’s AI (Claude), designed to be as user-friendly as possible—especially for non-programmers. It brings together Claude’s code agent capabilities into an accessible desktop tool.
  • "I would think of Claude Cowork as an actual personal assistant agent that lives and can operate on your desktop, it can get to your files, it can help you do tasks, et cetera." — Brian ([01:36])
• Use Cases:
  • Conventional to-do automation: prioritizing tasks, assigning responsibilities, potentially even performing actions like booking appointments.
  • For auditors: integrating with work management tools (like Asana, Jira), automating documentation, data analysis, and more.
• Plugin Marketplace Explained:
  • Plugins act as packaged toolsets that let Claude interact with external tools (ERPs, SaaS systems, etc.), similar to APIs.
  • As Brian explains, "It's essentially just a package of tools that allow the LLM to connect to the world... you can create your own (plugins)... but there's a package of all the things that you would need to have Claude talk to the world." ([03:07])

2. Open Source & Customization Power

• Anthropic hosts its plugins openly on GitHub as a “starter kit,” inviting anyone to tailor or create their own versions.
• The marketplace is likened to a farmer’s market:
  • "You create your own stuff, you bring it, people can come use, buy." — Trent ([07:43])
• A key insight: the skill sets and task automations available in Cowork are not exclusive—users or other developers can build and share their own specialized tools.

3. Why This Matters for Internal Audit

• Rapid Evolution
  • The year-over-year progress in agent-based automation is highlighted, emphasizing that developments are moving faster than expected.
  • "Whatever crazy thing you think is going to happen in... whatever time period is probably going to happen quicker." — Brian ([08:44])
• Practical Skills for Auditors:
  • Out-of-the-box skills include audit support, SOX 404 testing, control documentation, sample selection, classification of control deficiencies, and more.
  • These "skills" are structured as markdown instruction files, giving context and guidance to Claude for generating relevant outputs.

4. Security, Deployment, and Practical Guidance

• Caution on Deployment:
  • Should you use these tools on a work laptop? Both Trent and Brian urge caution.
  • "All you have is a work laptop, spend 200 bucks, go get a laptop and do this on that." — Trent ([10:41])
  • Brian’s counterpoint highlights Anthropic’s security reputation and the virtualized, sandbox approach, but he still recommends limiting file access and careful granting of permissions.
• Key Takeaway:
  • No matter your comfort level, experiment with these tools in a safe, controlled environment—doing something is better than waiting and being left behind.
  • "I think doing something is better than sitting and waiting, as some other people want to do." — Brian ([13:20])

5. Accessibility & The Future

• Lowering the Barrier for Non-Programmers:
  • Anthropic, OpenAI, and others are racing to create agent-based AI tools that are production-ready and easy to use, even for nontechnical users.
  • "They are clearly making access to the non programming world... more accessible. And for the accountants, the auditors, finance people... they're basically saying hey, we're, we're going to give you guys the starter kit." — Brian ([17:32])
• Market Signals & The Coming Wave:
  • The move toward user-friendly, SOC-compliant, 'out-of-the-box' solutions for enterprise AI is accelerating.
  • Anticipate widespread, production-grade offerings within a year from all leading vendors.

Notable Quotes & Memorable Moments

• On the Future Pace:
  "Whatever crazy thing you think is going to happen in... whatever time period is probably going to happen quicker."
  — Brian ([08:44])

• On Plugin Ecosystem:
  "You create your own stuff, you bring it, people can come use, buy."
  — Trent ([07:43])

• On Security and Experimentation:
  "All you have is a work laptop, spend 200 bucks, go get a laptop and do this on that."
  — Trent ([10:41])

  "I think doing something is better than sitting and waiting, as some other people want to do."
  — Brian ([13:20])

• On Skills for Auditors:
  "They've literally just Anthropic has basically said here's context... It's basically taking the vast LLM knowledge base and then shrinking it down and saying use this stuff, use this context."
  — Brian ([16:13])

Timestamps for Important Segments

• 00:39 — Introduction & Definitions: Claude, Claude Code, Claude Cowork explained
• 03:07 — What are plugins and plugin marketplace? How they connect AI with your tool ecosystem
• 07:43 — Open source plugins, customization, and the “farmer’s market” analogy
• 08:44 — The acceleration of agent-based AI, impact on finance & audit, signal for the industry
• 10:14–13:18 — Security discussion: where and how to try Cowork safely
• 15:00–18:00 — Walkthrough of actual audit skills: SOX testing, audit support, shaping context for AI
• 17:32 — Making enterprise AI accessible to non-programmers and signals about coming enterprise-grade solutions

Conclusion / Takeaways

• Anthropic’s Claude Cowork and plugin ecosystem represent major leaps forward in making powerful AI tools accessible, customizable, and effective for audit and finance professionals.
• The open-source approach empowers teams to develop, share, and refine sector-specific automation and analysis capabilities.
• Caution on deployment: experiment in a safe, segregated environment before rolling out in production.
• The landscape is evolving rapidly; auditors and accountants should get hands-on now to be prepared for the industry-wide adoption of AI agents and plugins.

For more resources and walkthroughs, Trent and Brian encourage listeners to check out relevant GitHub links and stay proactive in experimenting with AI tools for audit.