AI tools raise efficiency and accuracy in accountancy and business advisory, delivering measurable time and cost savings for adopters. However, uneven implementation, data quality issues and integration barriers limit the scale and consistency of those productivity gains.

Main Finding

AI adoption in accountancy and business advisory materially improves operational performance — the study reports a 40% reduction in process times, a 60% reduction in error rates, and a 10‑point increase in client satisfaction — while adoption is constrained by costs, skills gaps, organisational resistance and ethical/governance concerns.

Key Points

• Technologies in use: machine learning and digital assistants are the most frequently adopted AI tools in accountancy/advisory.
• Measured impacts (reported from regression analysis):
  • Process times reduced by ~40%.
  • Error rates reduced by ~60%.
  • Client satisfaction increased by ~10 points.
• Theoretical framing: Technology Acceptance Model (TAM) and Diffusion of Innovations explain drivers of adoption (technological readiness, perceived usefulness).
• Barriers to adoption: high implementation costs, need for specialised skills, organisational resistance, and ethical/governance issues (transparency, accountability, regulatory compliance).
• Practical recommendation: combine technology investment with training and cultural change; ensure AI systems meet ethical and regulatory standards.
• Future research suggested: longitudinal studies, cross‑industry and cross‑country comparisons, and study of AI integration with blockchain and big‑data analytics.

Data & Methods

• Data source: secondary data (PwC’s "Sizing the Prize" report) used to evaluate AI impacts on operational metrics.
• Analytical approach: quantitative/statistical methods including regression analysis, paired t‑tests, and ANOVA; descriptive statistics to profile technology adoption.
• Key empirical outcomes (as reported): 40% reduction in process times, 60% reduction in error rates, +10 client satisfaction points.
• Limitations (implicit in method):
  • Reliance on secondary, aggregate data (PwC report) — potential issues for firm‑level causal inference and generalisability.
  • No primary data collection reported (e.g., firm surveys, interviews) or experimental/longitudinal design to establish longer‑run effects or dynamics.
  • Abstract does not report sample size, error margins, or specification details for the regressions.

Implications for AI Economics

• Productivity and service quality: Large reported efficiency gains imply substantial productivity effects in professional service firms; AI can shift time allocation away from routine tasks toward higher‑value advisory work.
• Labor market effects: Findings are consistent with skill‑biased technological change — demand will grow for data/AI‑literate accountants and for tasks requiring judgment and client interaction, while routine roles may contract or be reskilled.
• Returns to investment and adoption heterogeneity: Benefits depend on technological readiness and strategic alignment, suggesting uneven returns across firms and potential widening of performance gaps in the sector.
• Market structure and pricing: Cost reductions and higher throughput could compress prices or enable new service models (subscription/automated advisory), altering competitive dynamics among firms.
• Policy and regulation: Ethical, transparency and accountability risks highlight needs for governance frameworks, standards for explainability/auditability, and targeted upskilling support to manage transition costs.
• Research agenda for economists:
  • Estimate firm‑level causal effects using primary panel data or quasi‑experimental methods.
  • Model diffusion dynamics and complementarities between AI and complementary investments (training, IT infrastructure).
  • Quantify welfare implications: consumer surplus from improved services vs. distributional impacts on employment and skills.
  • Study interactions between AI adoption and regulatory regimes, and potential market power effects from platformized accounting tools.

Overall, the paper provides indicative quantitative evidence that AI can substantially improve firm performance in accountancy/advisory, while emphasising adoption barriers and the need for governance — both important inputs for economic analysis of AI diffusion and policy design.