Audit Process Automation

A mid-tier UK accountancy firm cut routine audit documentation time by 58% by integrating AI into their sampling, evidence gathering and working paper generation — while satisfying FRC audit quality requirements throughout.

The Challenge

The firm's audit teams were spending approximately 40% of total engagement time on routine documentation: populating working paper templates, cross-referencing trial balance entries, documenting sampling rationale and writing standard-form explanations of audit procedures. This was consuming senior staff time that should have been directed at professional judgement and client advisory. The FRC's March 2026 guidance on generative and agentic AI in audit added a new governance requirement: any AI-assisted element of audit work had to be documented, with the auditor's professional judgement clearly recorded.

The Trovix Approach

Trovix integrated an AI documentation assistant into the firm's existing audit management platform. The system ingests trial balance data and prior-year working papers, generates first-draft working paper sections for routine audit areas, proposes sampling selections based on materiality parameters set by the engagement partner, and cross-references evidence against assertions. Every AI-generated section is clearly marked, with a mandatory auditor review and sign-off step before the working paper is finalised. The governance layer produces a complete record of AI involvement for FRC documentation purposes.

Technical Architecture

Outcome

Routine audit documentation time fell by 58% in the first audit season. Senior staff redirected the time saved to higher-value advisory work. An FRC quality review of three audits completed using the system returned no findings. The firm is expanding the deployment to all engagement teams for the next reporting season.