CwX 2026: How agentic AI is transforming audit and assurance

Caseware’s CwX 2026 global flagship conference series launched last month in Fort Lauderdale, FLA, with a focus on agentic AI and Caseware’s latest innovation, Caseware Verity, an agentic intelligence layer built into the Caseware platform.

At a CwX session titled “Unveiling the Next Generation of Caseware AI,” Andrew Smith, Chief Product Officer at Caseware, opened by acknowledging how quickly the AI landscape is moving, and why that pace matters for the profession. The core argument: generic AI tools aren’t enough for assurance. The unique demands of the work, professional judgment applied across an entire engagement, firm methodology that must be respected at every step and the accountability auditors carry, require something purpose-built.

That’s what Verity is designed to be. It’s built directly into Caseware Cloud, giving audit and assurance teams context-aware guidance to support their workflows, grounded in professional standards, firm methodology and the full context of the engagement they’re working on.  

Why assurance needs its own AI

Smith framed three characteristics of assurance work that shape what AI needs to do to be genuinely useful.

Contextual reasoning. An auditor applies professional judgment across an entire engagement including the trial balance, prior year, board minutes and risk assessments. For AI to support that judgment, it needs to reason across that full body of context.

Methodology governance. In assurance, the methodology governs every step, every procedure, every risk response. AI needs to be embedded within that workflow and not layered on top of it.

Professional accountability. Every conclusion, including those shaped by AI, must be defensible under scrutiny. That requires AI outputs to be designed to be traceable and explainable, with the firm in control.

“The AI model matters,” he said, “but in specialized verticals, once you have a frontier model, what surrounds that model and how you fill the domain matters more.”

What Verity does

Smith walked through a planning review scenario to illustrate Verity in action. A reviewer asks Verity whether anything was missed during planning. Verity loads the trial balance, retrieves materiality, reviews identified risks and surfaces a variance with no associated risk attached. With Verity, that gap is caught at the planning review stage rather than discovered during fieldwork.

“Think of Verity as a second set of eyes on the engagement,” he explained. “It’s brief, contextual and ready when you are. The review points are still yours to call.”

The Verity Agentic Suites

Alongside Verity itself, Smith introduced the Verity Agentic Suites, agents that help streamline specific workflows within the engagement. Three were highlighted.

Disclosure Checklist Agent

The Caseware Verity Disclosure Checklist Agent supports the process of working through a disclosure checklist against draft financial statements. Rather than manually scanning through financials to locate each relevant disclosure, the agent reads the statements and returns citation-backed answers, pointing to where in the financials each answer comes from and providing context for why.

Firms beta testing the Disclosure Checklist Agent reported saving an average of 2.7 hours per disclosure checklist run, with results varying by engagement, according to Caseware's CwX 2026 beta program results.

Firms also found the agent useful as a quality check on completed checklists, in one case surfacing questions that had been answered incorrectly due to misinterpretation and explaining why the correct interpretation differed.

The agent transforms how staff deal with checklists, Smith said. “You’re not searching anymore. You’re reviewing, validating and learning.”

Document Intelligence Agent

The Caseware Verity Document Intelligence Agent, still in beta, automates the extraction of information from source documents into work papers, reducing manual data entry during engagement execution.

Risk Suggestion Agent

A preview of the Caseware Verity Risk Suggestion Agent was shown at CwX. In the demonstration, the agent analyzed an engagement’s trial balance, board minutes, and prior-year file and surfaced five risks for auditor review. One of these was a $118,000 discrepancy between the trial balance and the net profit figure reported in board minutes, highlighted as a potential high-impact issue that may warrant further investigation.

Governance and accountability

A recurring theme throughout the session was the importance of AI governance in an assurance context. Smith believes that governance needs to be built into the platform, not added as an afterthought.

In practice, this means Verity outputs are designed to be traceable and explainable, AI reasoning and artifacts are retained under firm control, engagement boundaries are enforced, and firm methodology is designed to align with platform workflows.

“When your team uses Verity, you never have to wonder whether an AI conclusion was authorized,” Smith said. “You never have to question whether your methodology was bypassed. Every AI output is traceable and explainable, with your firm in control.”

Feedback from beta partner firms

Three audit partners piloted Caseware Verity on live engagements during the 2025 busy season as part of Caseware's beta program, including Kevin McElgunn, Audit Partner at Withum; Jessie Kanter, Partner at Citrin Cooperman; and Jessica Estrella, Partner at EisnerAmper. They shared their feedback during a fireside chat, with several consistent themes emerging.

The first was the importance of engagement context. For Kanter, this was what set Verity apart from generic AI tools she had seen elsewhere:

"When I'm asking questions, it can pull from the context of the engagement, so I get actual answers that matter to what I'm doing right now, versus something generic."

McElgunn noted that the learning dimension of Caseware Verity was particularly valuable, drawing on his own experience as a junior auditor being walked through a disclosure checklist line by line by a partner. Verity, he suggested, provides something similar at scale.

For Estrella, the practical impact came down to time. Her teams were completing work in minutes that had previously taken hours. "We are seeing tasks really go from hours to minutes," she noted.

What’s next

Smith outlined a roadmap with a goal of automating up to 50% of the engagement workflow across the full lifecycle — prepare, plan, validate and report. The platform is also being developed as a composable environment, meaning firms will be able to build their own agents on top of it and connect Verity into their broader AI workflows.

Learn more about Caseware Verity and how it could work for your firm.

‍