About Snap
Rooted in auditing. Built for audit execution.
Audit testing (vouching transactions, analyzing journal entries, evaluating controls) is repetitive, high-stakes, and increasingly hard to staff.
It eats the hours of junior staff during busy season and gets uneven attention as deadlines compress. We started Snap because we believe modern AI can perform this work with documentation quality that makes review fast and confident. Every conclusion is traceable to the source document it came from, with the reasoning visible at each step.
Why we exist
Auditors should review evidence, not chase it.
Snap is built by auditors and engineers who understand both what a workpaper has to do for a reviewer and what AI can actually be trusted to perform today. Getting AI to perform reliably in production is genuinely hard, because AI is fundamentally non-deterministic. That understanding shapes how Snap is built: not as a demo that works on a good day, but as a system designed for the documentation, traceability, and review discipline that audit work actually requires.
Mike Kim, CPA
Co-founder
Mike is a CPA with six years of audit experience, including time at KPMG, followed by six years working for an AI research lab in Vancouver. He bridges deep audit experience with AI and drives Snap's quality and standards alignment.
Dhirendra Singh
Co-founder
Dhirendra built enterprise-grade systems at Microsoft and Zillow before co-founding Snap. He leads Snap's engineering and platform architecture, with a focus on building AI systems that meet the reliability and documentation standards audit firms expect.
The moment
Why now
Three things are converging.
Staffing pressure
Audit firms are losing experienced staff faster than they can replace them. Manual testing work falls on junior staff during busy season and gets uneven attention as deadlines compress.
AI maturity
AI has become reliable enough to perform structured testing work with documentation that holds up to review, provided the system is built for traceability and human oversight.
Regulatory alignment
Regulators in the UK, US, and Canada have published research and guidance on AI in audit. The common thread: AI is supported provided the work is explainable, supervised, and properly documented.
Built for this moment
Snap sits at the intersection of these three shifts: an execution layer designed to perform audit procedures the way the profession needs them done.
Where we are
Snap is built for this moment.
We are starting with the procedures that are hardest to staff and easiest to get wrong under pressure: vouching, journal entry testing, and controls. The goal is audit work that is consistent, traceable, and ready for CPA review from the start.