The Audit Podcast — Ep 264: How Internal Audit Teams Are Building AI Tools with No Tech Background Needed (Snowflake)

Host: Trent Russell
Guests: Amrita (Head of Internal Audit, Snowflake), Pooja (Internal Auditor), Ajay (Internal Audit Manager)
Date: November 18, 2025

Episode Overview

This episode explores how Snowflake’s internal audit team is empowering auditors—without deep technical backgrounds—to build powerful AI and analytics tools for audit transformation. The hosts and guests provide hands-on examples, discuss real-world impact, and share lessons learned, focusing on making AI accessible for audits, driving efficiency, and inspiring teams to embrace a data-driven future.

Key Discussion Points & Insights

1. Making AI Accessible to Auditors (00:00—04:42)

• Theme: Auditors as Change Agents
• Amrita urges auditors to “be change agents” by leveraging data and AI to do “more with less,” addressing the challenge of limited resources in audit teams.
• Trent emphasizes that Pooja, a key guest, has no IT or data science background but built a powerful AI tool in one evening. This demonstrates that with user-friendly platforms, even non-technical auditors can build and deploy AI solutions.
  • Trent: “You will not see…any mention of an IT background. What she built, the AI tool that she built, took her an evening to build!” (03:10)

2. Introducing Snowflake’s Native AI Tools (04:42—07:45)

• Snowflake & Cortex: Amrita provides context about Snowflake’s AI platform, Cortex, describing its document intelligence (Doc AI), LLM summarization, text-to-SQL, and search capabilities.
  • Amrita: “Cortex…houses all of these AI solutions…for enterprise data that resides within Snowflake. It has things like LLM functions for summarization, text classification…Cortex Analyst…Cortex Search…Doc AI.” (06:52)
• Business-Friendly UI: Emphasizes that these tools are out-of-the-box and business-user-friendly, requiring no code.

3. Demo 1: AI-Powered Contract Review (Pooja, 07:55—14:12)

• Challenge: Reviewing massive volumes of contracts is labor-intensive and a bottleneck.
• Solution: Using Doc AI, Snowflake’s internal audit automated extraction and review of contract terms with no coding, leveraging OCR and zero-shot models.
  • Pooja: “What if I tell you that you could train an AI first year analyst to review thousands of these contracts with perfect consistency and superhuman speed?” (08:24)
  • Upload contracts (PDF, etc.), ask questions in natural language (e.g., “Is there a volume discount?”), and receive answers with confidence scores—no hallucination.
  • The unstructured data is converted into a structured table for analysis.
• Impact:
  • Processed 7,000+ contracts in 15 minutes; highlighted issues even human reviewers missed.
  • Amrita: “Internal audit is being a catalyst for sparking interest in first line to also leverage AI. Which is the beauty here…” (12:01)
• Learning Curve:
  • Pooja: “Absolute zero coding business friendly user interface… turned weeks of manual work into a half day affair.” (13:06)
  • Built out in a single evening.

4. Demo 2: Automating SOC 1 Report Review (Ajay, 14:41—17:07)

• Problem: SOC reports for vendors are lengthy; manual review and extraction took 8–10 hours per report (200 hours/year for 20 vendors).
• AI Application:
  • LLM reads PDFs and extracts key data (period coverage, subservice orgs, control objectives, exceptions, management response) instantly.
  • Ajay: “Gives us the important information we need in a matter of minutes… anticipate saving around 150 to 160 man hours for a 20 vendor SOC report population.” (16:27)
• Next Steps: Future enhancements will auto-check subservice orgs and manage SOC report rabbit trails.

5. Demo 3: Analytics for Key Reports & Termination Controls (Ajay, 17:07—23:22)

• Data Analytics for Key Reports:
  • Ingests Workday audit logs into Snowflake, flags unchanged reports, benchmarks calculated fields, and reduces need for screenshots.
  • Amrita: “Number of screenshots that you guys have had to share with external auditors goes away. No one likes those screenshots.” (19:38)
• Continuous Monitoring of Termination Controls:
  • Automated Okta log ingestion addresses retention limits (only 90 days kept by Okta), allowing both continuous internal and SOX-specific threshold monitoring.
  • Dashboards show delayed terminations (over 1 day for management, 5 days for SOX), enabling immediate alerts and deeper risk detection.
  • Amrita notes monitoring enables both operational gains and proactive risk mitigation.

6. Team Enablement & External Audit Partnerships (23:22—27:50)

• Auditor Upskilling:
  • Amrita is pushing business auditors to learn and use these tools, citing converging skills and easy learning curves.
• External Audit Coordination:
  • Trent raises the need for alignment with external audit on AI tools and outputs, especially for critical SOX testing.
  • Amrita describes fitting use cases into three pillars: analytics, AI agents, and document intelligence—choosing the right tool for the problem.

7. AI and the Audit Profession: Augment, Don’t Replace (27:50—30:20)

• AI as an Augmenter:
  • Amrita dismisses fears that AI will replace auditors, viewing it instead as an augmentation tool that enables auditors to focus on value-added and advisory tasks.
  • “Jobs will evolve…the flip side is, there are risks too. We should also be a trusted advisor to educate stakeholders on those risks.” (29:00)
• Strategic Approach:
  • Start with the problem, not the tech; AI is not always the right answer, but offers huge ROI when applied strategically.

8. Three High-Impact AI Use Cases For Audit (30:20—30:45)

• Doc AI for contract review
• SOC report auto-summarization
• Automated key report analytics & terminations

9. Final Advice & Notable Quotes (30:45—33:53)

• Amrita’s Closing Message:
  • “I encourage all auditors to focus on being change agents… leverage data and AI tools to do more with less… move to more automated, AI- and data-driven techniques. The future of audit is here—and it’s data-driven.”
  • Combine data with business acumen: “Data doesn’t matter if you don’t have a good understanding of your business strategy and objectives.” (32:05)
  • On personal note: Balance tech enthusiasm with human connection.
    • “Enjoy technology but take a break, invest in relationships as well.”
    • Shares that she uses ChatGPT for parenting advice: “I’m also focusing on…how best to develop my parenting skills, learning patience, how to be a kid again with my kid!” (33:18)

Notable Quotes & Timestamps

• “What she built, the AI tool that she built… took her an evening to build.”
  — Trent, (03:10)
• “You can basically go in and ask questions around volume discount, is there a price cap… relevant when you are assessing contracts for revenue assessment purposes.”
  — Pooja, (09:25)
• “It’s absolute zero coding business friendly user interface… turned weeks of manual work into a half day affair.”
  — Pooja, (13:06)
• “We were able to analyze 7,000+ contracts within 15 minutes.”
  — Amrita, (12:01)
• “Gives us the important information we need in a matter of minutes… anticipate saving around 150 to 160 man hours...”
  — Ajay, (16:27)
• “Number of screenshots… you guys had to share with external auditors goes away. No one likes those screenshots.”
  — Amrita, (19:38)
• “Jobs will evolve… We need to embrace this. There is an opportunity, and the flip side is, there are risks too… we [should] also play that trusted advisory role.”
  — Amrita, (29:00)
• “The future of audit is here—and it’s data-driven. And so it’s very important to have that data-driven mindset.”
  — Amrita, (32:05)

Timestamps for Key Segments

| Segment | Speaker(s) | Timestamp | |--------------------------------------------|---------------|------------| | AI for Auditors: Change Agent Mindset | Amrita | 00:00–04:42| | Snowflake AI Platform Overview | Amrita | 04:42–07:45| | Doc AI Demo: Contract Review | Pooja | 07:55–14:12| | SOC Report Automation | Ajay | 14:41–17:07| | Key Reports Analytics & Terminations | Ajay, Amrita | 17:07–23:22| | Team Skills & External Audit Coordination | All | 23:22–27:50| | AI as Augmenter, Not Replacement | Amrita | 27:50–30:20| | Recap of Use Cases | Trent, Amrita | 30:20–30:45| | Final Thoughts & Personal Side | Amrita | 30:45–33:53|

Tone & Takeaways

The conversation is highly practical, optimistic, and demystifying—emphasizing real results, accessible innovation, and the evolving, human-centric nature of audit work. AI is presented not as a threat but as a toolkit that, when applied thoughtfully, can liberate auditors to focus on insight, advisory, and stakeholder impact.

Key Takeaway:
With modern AI tools like those in Snowflake, even non-technical audit teams can build powerful automation and analytics solutions. The future of audit is already here — data-driven, accessible, and collaborative.