Alwin by DataSnipper: End-to-end agentic automation, built for the realities of audit

Audit teams are being asked to cover more work as experienced talent becomes harder to find and retain, and AI has an obvious role to play in taking on more of that work and doing it quickly. But for audit teams, speed alone isn't enough; you cannot sign your name to work you never saw. Auditors need to be able to defend the conclusion regardless of how fast they got there. What I find interesting is that technology is no longer the limiting factor. Agents can already take on increasingly complex work. The harder question is how you make that work fit the realities of an audit. 

Much of the AI available today creates a difficult trade-off; you give agents more freedom to work across large volumes of data and complex procedures, and it becomes harder to see exactly what they did. Keep every interaction tightly controlled by an individual auditor, and you limit how much work agents can take on at once. 

Neither is a particularly good answer for a profession where someone ultimately has to review the work, put their name to it, and be able to explain how a conclusion was reached. 

With Alwin, agents can execute larger, end-to-end procedures and multiple agents can work in parallel, without losing the visibility and human oversight that make the work defensible. 

“Alwin is already showing tangible benefits for our audit work. In early testing, one cash-testing workflow was around 80% faster, reducing hours of work to minutes of agent run time followed by review.” Gert-Jan Brouwer, Partner Accountancy, CROP Accountants & Advisers 

What an end-to-end procedure looks like in Alwin

End-to-end doesn’t mean handing a procedure to an agent and waiting to see what comes back. When we talk about agentic automation, this means agents execute multi-step procedures, carrying each one from start to finish and handing the completed result to a human for review. That makes it different from a chatbot that answers questions or a tool that summarizes documents. Alwin keeps the auditor involved at the points that matter, while agents take on more of the execution.  

Alwin works across the full audit lifecycle: 

• Plan. Alwin can create a plan for the procedure before work begins, so the auditor can review the approach, make adjustments and approve it before the agent starts.
• Collect. Rather than working from a limited set of manually uploaded files, Alwin can connect to the systems you already use, from file storage platforms to ERPs, and pull in hundreds to thousands of files at once. Agents turn that unstructured documentation into a clear, traceable starting point for verification.
• Verify. Multiple agents can execute procedures in parallel, asking for input when needed and surfacing exceptions that require professional judgment. Evidence stays linked back to its source through Snips.
• Report. Alwin produces workpapers and supporting deliverables ready for review. Auditors can see what the agent did, leave notes, provide additional evidence or instructions, and send the agent back to work when needed.

The Audit Workbench brings that activity together across clients and engagements, giving teams visibility into what’s running, where input is needed and what is ready for review. Spaces and Projects let teams organize work around the way engagements are actually structured, with access controlled across the team. 

FGMK began using Alwin through the design partner program. As part of this partnership, they worked closely with the product team, testing Alwin on audit materials for busy-season and year-end work, and providing valuable feedback as the platform was being developed.

“I immediately tried Alwin on some of our more difficult tasks, like hundreds of pages of GL extraction, and it worked really well. It can also keep tasks running in the background and pick up right where you left off.” Nicholas T. Robbins, Director, Audit and Accounting at FGMK 

Build it once. Put it to work across the firm.

That means auditors can work with agents already defined for a particular procedure, rather than recreating the instructions every time. Firms can manage those changes centrally and roll out new versions in a controlled way. 

The audit trail can’t stop where AI starts

As agents take on more of the work, auditors still need to see how that work was done. In Alwin, agent activity stays tied to the procedure. Reviewers can see the steps taken, the interactions with the team and the evidence behind each conclusion. Through Snips, that evidence remains linked back to its source. 

Review notes and additional instructions stay with the work too, creating a reviewable record of both human and agent activity. 

Meet you where you are

DataSnipper is built to support firms wherever they are in their AI journey, from task-based automation and AI-assisted work through to agents and end-to-end agentic workflows. For firms ready to put agents to work across larger procedures, that shouldn’t mean giving up the controls their teams rely on. 

Alwin’s agents run in the cloud, where they can work across large volumes of evidence and multiple procedures in parallel without relying on an auditor’s laptop. With regional infrastructure and enterprise security and data governance controls, that work can scale without compromising how sensitive audit data is managed. 

The work can then move back into familiar environments, including Excel and existing Audit Management and workpaper systems. Alwin connects with systems including SharePoint, Box, OneDrive, SAP Concur, Xero, QuickBooks and PBC Uplink, bringing evidence directly into the workflow without manual exports and handoffs. 

“Alwin gives us the potential to run this work in parallel in the cloud, complementing the Excel processes we use for high-volume tasks. Just as importantly, every result remains traceable to its source.” Lukas Homola, Director, Head of Audit Technical Team, BDO Czech Republic 

Built for the governance audit requires

As agentic work scales, firms also need control over where client data is processed and who can access the work. Alwin runs agents in the cloud with regional data governance in mind, rather than requiring auditors to download client documents and run AI locally on individual laptops. Access to projects and agents can be controlled centrally as teams work across engagements. 

That allows firms to scale agentic work without giving up the governance expected around sensitive audit data. 

A platform built with auditors, for audit

The name Alwin goes back to the beginning of DataSnipper. Nearly a decade ago, an auditor named Alwin shared the repetitive, manual parts of his work with the three friends who went on to start the company. Naming the platform after him brings us back to where DataSnipper started, building for auditors and the work they do. 

Alwin has been shaped closely with our design partners, testing what we build against real procedures, evidence and review processes.