Modernizing Government Audit with Automation and AI: Key Insights from DataSnipper’s Public Sector Event

Government and public sector audit teams sit at the center of financial accountability, responsible for overseeing enormous volumes of documentation while meeting rising expectations for transparency, traceability and compliance. In this webinar, leaders from DataSnipper and Bryan Nelms, Partner at Kearney & Company discussed what's changing in public sector audit, how agencies are approaching automation vs. AI, and what practical modernization looks like without disrupting existing audit methodology. 

Below are the key insights and takeaways. 

Right from the start, the message was clear: the opportunity in government isn't to "disrupt" how agencies work. It's to modernize core workflows so teams can keep up with growing oversight expectations. 

Several macro forces are converging: 

• Massive oversight scope: Over $7 trillion in public spending is subject to oversight annually.
• Rising expectations: More pressure for transparency, traceability, and stricter compliance.
• Shrinking talent pipeline: Fewer accounting graduates (down roughly 20% since 2010), declining CPA participation and significant attrition from the profession.
• Investment surge: Analysts expect $98.1 billion invested into AI/automation expansion by 2033, suggesting rapid adoption ahead, including in government.

The bottom line: audit demand is increasing while capacity is constrained, creating more work with fewer people and making technology adoption less optional and more essential. 

From DataSnipper's public sector conversations, most organizations are pursuing some combination of hiring, automation to reduce manual work and increase throughput, and innovation including AI exploration and rollout. 

A recurring theme: hiring alone can't close the gap. Technology is now central to modernization strategies, particularly tools that integrate with existing processes rather than forcing wholesale change. 

Across customer feedback, DataSnipper emphasized four themes that repeatedly come up in government and public sector adoption: 

1. Embedded in Excel: Teams want to work where they already operate.
2. Full audit traceability: Evidence needs to be reviewable and defensible.
3. Alignment with existing methodology and compliance: Modernization cannot compromise compliance.
4. Productivity gains without additional headcount: The core driver behind adoption.

Government certifications and accreditations also emerged as a requirement for scaling adoption. DataSnipper highlighted achieving Texas StateRAMP accreditation as one example. 

Drew Warren, Account Executive at DataSnipper, drew a clear line between the two. 

Automation is rules-based and deterministic. It follows defined logic and repeatable steps. Examples discussed include cross-referencing and document matching, recalculation checks, and standardized extraction from similarly structured documents. The value is efficiency through rules, high auditability, and easier governance approval. 

AI is context-based and interpretive. It handles variability and unstructured content. Examples discussed include extracting from non-standard documents such as invoices where fields appear in different places, pattern and anomaly identification, and summarizing findings. The value is efficiency through interpretation, especially helpful when documents and investigations are less structured. 

This framing resonated strongly for public sector adoption because many agencies need to start with deterministic capabilities before layering in probabilistic AI. 

The webinar highlighted how IT environments, compliance requirements and governance structures vary widely across government, and adoption patterns reflect that. 

• Executive branch agencies often take an automation-first approach. Common realities include FedRAMP or similar requirements, restrictive IT approval processes and risk-averse adoption culture. Automation works well here because it is deterministic, Excel-based and fully auditable, fitting existing procedures with minimal disruption. Common use cases mentioned include bulk document matching, 100% population testing, invoice-to-contract cross-referencing, and standardized workpaper documentation.
• Legislative and oversight environments tend to explore AI earlier. AI becomes particularly valuable when document structure is inconsistent, work is investigative such as performance audits, or reporting cycles are tight. AI-oriented use cases mentioned include extracting clauses from contracts, handling non-standard PDF extraction, anomaly detection for investigative review support, extracting clauses from contracts, and investigative analysis support.

The recommendation: agencies don't need to choose automation or AI. Modernization can be phased, adding AI as readiness evolves. 

Former auditor and Solutions Engineer, Luis Mejia demonstrated how DataSnipper supports evidence work directly in Excel via an embedded ribbon. The workflow begins by importing documents into the workbook, which is important for controlled scope and traceability. 

Key capabilities demonstrated included: 

• Foundational snips allow auditors to embed evidence directly in Excel. Text Snip pulls text into cells with clickable traceability back to source evidence. Sum Snip ties multiple amounts to a single total with visible components. Table Snip extracts tables from PDFs into Excel, adjustable via column and row controls. Markups enable annotations, checkmarks and signatures within documents without exporting.
• Automation features include automated table extraction across pages, and Document Matching, which matches source documents such as invoices to Excel rows with traceability to evidence, options for thresholds and fuzzy matching, and clear handling of exceptions. Form Extraction scales extraction across similarly structured documents and saves templates.
• Financial Statement Suite is a specialized workflow for financial statement review that includes mathematical accuracy checks, internal consistency cross-references, spellcheck and prior-year comparisons. The module supports exporting a PDF with a summary of findings and version comparison, where unchanged tick marks can carry forward so reviewers focus only on what changed.
• AI-enabled workflows include DocuMine, a Q&A tool over documents that delivers traceable outputs auditors can approve, reject or edit, reinforcing a human-in-the-loop approach. AI ExtractionS suggestions can pull fields from variable document formats, such as checks, with cell-level traceability intact.

• Agentic capabilities were also previewed, where an Excel Agents can read the procedure template in the workbook, plan steps in natural language, execute extraction and reconciliation tasks, and write results back into Excel with linked evidence and exception notes.

Brian Nelms, partner at Kearney & Company, shared lessons learned from adopting DataSnipper over approximately two years in federal financial statement audits. 

• The first objection: security. In federal environments, the immediate concern is how tools handle sensitive information. Nelms emphasized the need for governance and security alignment, partnering closely with internal IT leadership including the CIO, and ensuring policies, procedures and compliance frameworks support adoption. He also noted his firm's achievement of CMMC Level 2, underscoring how technology adoption must map to federal security expectations. 
• Starting small and scaling deliberately. Nelms’ approach followed a staged path: proof of concept using mocked data, a pilot with limited engagements, limited deployment, then broader scaling. This phased approach reduced risk and increased adoption confidence. 
• Where they saw immediate value. The firm eliminated repetitive manual processes in testing workflows and found efficiencies in financial statement review for tick-and-tie and footing work. An unintended benefit was improved collaboration and faster review. Snips made evidence easy to navigate for reviewers, staff were pushed to be explicit about support for attributes, and gaps became visible sooner, prompting earlier coaching and clarification. 
• A practical client impact. Automation helped teams identify missing or unclear support earlier, allowing auditors to return to clients faster with a targeted list of gaps and giving clients more time to remediate before deadlines. 

In the webinar's live poll, the top barrier to adopting AI tools in audit functions was compliance and regulatory concerns, by a large margin. Most respondents said their agency is "just exploring" AI. 

This aligns with the overall theme: public sector organizations want the benefits, but adoption must be defensible, governed and incremental. 

The clearest thread across the webinar was that modernization in government audit must meet agencies where they are. That means automation-first for deterministic, high-auditability workflows; AI layered in where document variability and investigative needs justify it; human-in-the-loop review to preserve accountability; and Excel-native workflows to minimize change management and improve adoption. 

For public sector audit leaders facing increasing workload and shrinking capacity, the message is pragmatic: start with high-impact workflows, prove value safely and scale with governance without compromising compliance.