What happened
Snowflake's Internal Audit team cut fieldwork time for evidence collection and report production by 50%—from months to weeks—saving 3,500+ hours in 2026 and 1,500+ for control owners.
Why it matters
The time freed up lets auditors shift from gathering evidence to validating it and focus on high-risk strategic areas, though auditors still own every conclusion under a human-in-the-loop model.
WHO IT HITSInternal audit teams at large enterprises stand to benefit most, as the workload reduction frees auditors to focus on high-risk strategic work and shrinks manual evidence requests for business control owners.
Summaries like this, every morning.
Snowflake's internal audit team had a head start on this shift because billing, revenue, HR, access, and audit log data were already live in Snowflake. That setup let the team test against full populations without stitching together disconnected sources—a prerequisite the team stresses as essential.
The change runs deeper than time savings. For key controls such as revenue stream scoping, order-to-cash testing, and key access controls monitoring, testing moved from random statistical sampling to 100% population coverage with near real-time visibility into exceptions. On payroll reconciliation alone, a CoCo-built app gave control owners back 500+ hours a year. The team says it has also increased the number and complexity of audits it runs.
Accountability remains with people. The team operates under a human-in-the-loop model: AI prepares and auditors conclude, reviewing AI output before anyone relies on it. Third-party tools like Petual.ai run as Snowflake Native Apps, keeping processing inside the company's Snowflake account. The team frames the role of the internal auditor as evolving from historical record-keeper to strategic risk advisor.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
