How Kansaro transforms audit testing with AI | Kansaro Briefs
audit testingaudit technologyAI automation

Audit testing with technology

Scott Newcomer 7 min read

The gap in audit technology today

Walk into any audit team and you might find a similar stack: DataSnipper for document matching, Suralink for requests, maybe TeamMate or AuditBoard for workflow and source of truth system. These tools are helpful because they organize your work, streamline documentation, and keep everything in one place.

But here's what they don't do: they don't actually perform the testing.

You still pull samples. You still chase down supporting documents. You still match invoices to payments, trace approvals through email threads, and validate that controls actually happened. The tools help you organize that work, but the substantive testing? That's still on you.

And that's the problem. Because audit testing isn't just time consuming, it's where risk lives.

What auditors actually need

You need technology that does the hard part: analyzing messy evidence, performing the actual tests, and delivering results you can defend.

Let's think about a typical access provisioning audit. You're testing whether new system access follows the OIDC principle - only authorized users get access, based on documented requests, with manager approval, within the right timeframe.

You've got:

  • 147 access grants across Active Directory, Salesforce, NetSuite, and three legacy systems
  • Tickets scattered across ServiceNow, Jira, and email
  • Some requests have approval chains, some have a single email, some have nothing
  • Termination dates don't sync between HR systems and IT
  • A contractor who got admin rights due to an "urgent" need

Traditional audit tech helps you store all that chaos but it doesn't necessarily solve it. You need something that augments your work.

How Kansaro is different

At Kansaro, we built something fundamentally different. We don't just help you test, rather we do the testing for you.

Our AI doesn't just match documents or flag potential issues. It performs deterministic, repeatable tests on messy real world evidence and delivers clear, defensible workpapers.

How Kansaro performs audit testing

Let's walk through that access provisioning test.

You upload your access data. Maybe it's an Active Directory export, user lists from your SaaS platforms, whatever you pulled from IT. And the data is messy. Different date formats, inconsistent naming conventions, privilege levels coded six different ways.

Kansaro reads it all such as email archives, ticketing systems, and whatever evidence you have supporting this test. Then it performs the actual tests you have defined: Was there a documented request? Did the right person approve it? Was it provisioned within your policy timeframe? Does the access level match what was requested?

Kansaro doesn't force you into predefined templates or rigid data structures. You bring your attributes and control data. The platform adapts to how your organization actually works, not the other way around. It maps your data, finds the connections between access grants and approvals, and tests against your specific control criteria. Did the right person approve it based on your authorization matrix? Was it provisioned within your policy timeframe? Does the access level match what was requested for that role in that department?

The output isn't a list of documents to review. It's a completed workpaper showing which access grants passed all controls, which failed, and the specific evidence supporting each conclusion. Every result is tied back to source documents and thus everything is defensible.

You're not spending two weeks manually matching 147 user accounts to approval emails. The testing is done.

Audit testing

Real examples across audit types

Purchase order approvals

You're testing whether POs over $50K have VP approval. You upload your PO data from the ERP, probably a CSV with inconsistent formatting, notes crammed into description fields, some dollar amounts with currency symbols and some without.

Kansaro identifies which POs cross the threshold, finds the approval documentation you provided based on your attributes (email threads, workflow system records, scanned signature pages), and tests each one. Did a VP approve it? Was it timely? Does the approver's authority match the amount?

Done. Workpaper ready for review with exceptions flagged and evidence attached.

Segregation of duties

You're validating that no single employee can both create vendors and approve payments. You've got user access data from your financial system: role assignments, permission matrices, historical access changes.

Kansaro maps every user's combined permissions, identifies conflicts, and documents which users have inappropriate access combinations. It traces when the access was granted and whether compensating controls exist.

You get a clear list of SOD violations with supporting evidence and not a spreadsheet you need to further analyze.

Expense report compliance

Testing whether expense reports over $5K have itemized receipts and VP approval. You're working with expense system exports, PDF receipt packets, approval workflows, and probably some email forwards from the CFO's assistant.

Kansaro processes every expense report in your sample, extracts receipt data (even the blurry or crunchy ones), validates approval chains, and tests compliance with policy limits. It catches the $4,800 expense split across two reports to avoid approval requirements. It flags missing receipts. It documents everything.

Your workpaper shows pass/fail by report, with drill-through capability to see the source documents.

Drill down

You stay in control

Here's what's critical: you're still the auditor. Kansaro doesn't make judgment calls about materiality or risk. Kansaro doesn't decide what's significant. Kansaro accelerates the testing, it doesn't replace your expertise.

What it does is handle the mechanical work. The sampling, the matching, the testing, the documentation so you can focus on what actually requires your expertise. Is this exception part of a pattern? Does this control failure indicate a bigger issue? What does management need to know?

At every step, you can see exactly what Kansaro tested, how it tested it, and what evidence it used. You can drill into any transaction, review the logic, validate the conclusion. The AI handles the volume and you get to handle the judgment.

That's how audit technology should work.

What this means for your audit process

When testing takes hours instead of days, a lot can change. You can test more controls. You can expand sample sizes without adding weeks to your timeline. You can respond to new risks mid-audit without derailing your schedule. Your team stops burning hours on data cleanup and document matching and they start spending time on analysis, root cause investigation, and actually talking to business partners about what matters.

And when findings do surface? You've got the evidence already documented, the workpapers already built, the audit trail already clear. No more scrambling to recreate your work.

Building vs. buying

Some teams are tempted to build this themselves. I get it. You've got smart people, everyone's excited about AI, how hard could it be?

Really hard. I wrote about this after the IIA conference - building technology products is expensive, complex, and full of hidden costs. Technical debt piles up fast. Requirements never stop coming. Team continuity is a constant challenge.

And here's the thing about audit testing specifically: you need deterministic results. Not "pretty good" answers and definitely not "the AI tried its best." Defensible, repeatable, auditable conclusions that will hold up under scrutiny.

That takes years of refinement, extensive testing, and deep domain expertise in both audit methodology and AI engineering. It's not a side project for your analytics team.

Getting started

If you're curious how Kansaro might work for your team, let's talk. We'll walk through a real test from your audit plan and show you what the platform can do.

No demos with fake data, no generic use cases. Just your actual work, done faster and with confidence you can defend.

At Kansaro, we're building AI for the front lines of audit where real work happens and every detail matters. Our platform delivers deterministic, tested results from messy evidence, turning hours of manual digging into clear, defensible workpapers.

You know as an auditor, you don't want to guess. You need answers you can trust. Kansaro gives you the tools to get there faster and with confidence.

Contact us to see Kansaro in action

← Briefs
audit testingaudit technologyAI automation

Meet a collaborative AI for auditors.
Work will never be the same.

Join the waitlist

Be the first to know when the personal plan becomes available. We'll notify you as soon as spots open up.