Survey data on AI adoption, error rates, and security policies in accounting

Dynamite Docs, 2026-08-23

Key statistics at a glance

This guide collects statistics on AI adoption, security risk and accounts payable operations in accounting and finance. Read the source beside each number. Survey populations, definitions and dates differ, so two percentages may describe different groups rather than contradict each other.

In a survey of nearly 600 accounting professionals across six continents, 98% reported using AI tools in their work. (Karbon, State of AI in Accounting Report 2026)
AI adoption among tax and accounting firms grew from 9% in 2024 to 41% in 2025, with 72% using AI tools at least weekly. (Wolters Kluwer, Future Ready Accountant 2025)
A data breach in the financial services sector cost an average of USD 5.56 million in 2025, the second highest of any industry. (IBM, Cost of a Data Breach Report 2025)
Only 21% of surveyed accounting firms had an AI policy or documented strategy in place. (Karbon, State of AI in Accounting Report 2026)
Among organizations experiencing data breaches, 20% linked incidents to unauthorized or unvetted AI tools used by staff. (IBM, Cost of a Data Breach Report 2025)
Ransomware was involved in 88% of small and midsize business cyber breaches analyzed in 2025. (Verizon, 2025 Data Breach Investigations Report)
UK accounting practices using AI reported saving an average of 18 hours and 53 minutes per week on routine administrative tasks. (Xero, UK accounting AI economic-impact study)
The global AI in accounting market was estimated at USD 6.93 billion in 2025 and is projected to reach USD 53.45 billion by 2030. (The Business Research Company, AI in Accounting Market Report 2026)

AI adoption rates across accounting firms

Adoption figures change with the question. Some surveys count any use of an AI tool. Others ask about formal deployment, weekly use or planned investment. Keep the population, question and fieldwork date beside every percentage before using it in a budget or policy paper.

In an Intuit survey of accountants, 64% said their firms planned to invest in or upgrade AI software within the year, and 46% reported daily use. (Intuit QuickBooks, Accountant Technology Survey 2025)
Among tax firms using generative AI, 44% reported using it daily and another 29% used it weekly. (Thomson Reuters, Generative AI in Professional Services Report 2025)
CPA firms of all sizes ranked technology and AI adoption as the top issue affecting their practices over the next five years. (AICPA & CIMA, PCPS Top Issues Survey 2026)
In a survey of 1,256 small business owners, 76% reported using AI tools and 93% noted positive business outcomes. (Goldman Sachs, 10,000 Small Businesses Voices AI survey 2026)
The Federal Reserve Banks small business survey found 46% of employer firms using AI and another 15% planning to adopt within 12 months. (Federal Reserve Banks, Small Business Credit Survey 2026)

Market size and industry growth projections

Market forecasts are estimates, not operating benchmarks. Research firms use different category definitions and forecast periods. The gap between the figures below is a reason to read the methodology, not to average them together.

The Business Research Company estimates the accounting AI market will grow from USD 6.93 billion in 2025 to USD 10.4 billion in 2026, reaching USD 53.45 billion by 2030. (The Business Research Company, AI in Accounting Market Report 2026)
Grand View Research forecasts the market to expand from USD 4.87 billion in 2024 to USD 96.69 billion by 2033 under a broader market definition. (Grand View Research, AI in Accounting Market Report)
Mordor Intelligence projects the dedicated intelligent document processing market to reach USD 7.18 billion by 2031. (Mordor Intelligence, Intelligent Document Processing Market Report)

Accounts payable automation benchmarks

Industry benchmarks compare the average accounts payable department with Best-in-Class teams on invoice cost, cycle time, exceptions, and straight-through processing.

The average accounts payable department spent USD 9.84 to process an invoice and took 8.2 days. (Ardent Partners, State of ePayables 2025 benchmarks)
Best-in-Class accounts payable teams reported invoice processing costs and cycle times 79% lower than all other organizations. (Ardent Partners, State of ePayables 2025 benchmarks)
Average accounts payable departments reported an 18.4% invoice exception rate and spent 21.9% of staff time dealing with supplier inquiries. (Ardent Partners, State of ePayables 2025 benchmarks)
Best-in-Class teams processed more than 1.8 times as many invoices straight through as peers. (Ardent Partners, State of ePayables 2025 benchmarks)
Intuit found that 95% of surveyed accountants adopted automation in 2024, led by payroll (47%), accounts payable/receivable (46%), and transaction entry (43%). (Intuit QuickBooks, Accountant Technology Survey 2025)

Cybersecurity and data breach risks for finance teams

Accounting firms and finance departments hold financial records, tax identifiers and bank details. Use the breach figures below to size controls and incident planning, not to predict the cost of a particular incident.

The average cost of a data breach in financial services reached USD 5.56 million in 2025. (IBM, Cost of a Data Breach Report 2025)
Organizations with unvetted shadow AI tools experienced breach costs averaging USD 670,000 higher than organizations with clear AI security policies. (IBM, Cost of a Data Breach Report 2025)
Customer personal data was exposed in 65% of shadow AI breach incidents. (IBM, Cost of a Data Breach Report 2025)
Third-party software vulnerabilities were involved in 30% of breaches analyzed by Verizon. (Verizon, 2025 Data Breach Investigations Report)
Ransomware accounted for 88% of small business cyber incidents. (Verizon, 2025 Data Breach Investigations Report)

Security standards and regulatory guidance for tax preparers

Accounting practices and tax professionals in the United States must comply with federal security rules designed to protect client financial records.

The FTC Safeguards Rule requires covered tax preparation firms to maintain a written information security plan with administrative, technical, and physical safeguards. (Federal Trade Commission, Safeguards Rule guidance)
Covered entities must notify the FTC within 30 days of discovering unauthorized access affecting 500 or more consumers. (Federal Trade Commission, Safeguards Rule guidance)
The IRS Security Six checklist advises tax professionals to maintain active anti-malware software, firewalls, two-factor authentication, routine data backups, drive encryption, and secure VPNs for remote work. (Internal Revenue Service, Security Six guidance)

Frequently asked questions

These answers explain how to compare the numbers and turn them into a practical review.

What percentage of accounting firms use AI in 2026? Survey results vary by definition. Karbon found 98% of surveyed professionals used AI tools, while Wolters Kluwer reported 41% formal adoption and 72% weekly use.
How much does a financial data breach cost? IBM found the average financial sector breach cost USD 5.56 million in 2025. Encryption and strict access limits reduce that risk.
Do tax preparers need a written security plan? Yes. Under the FTC Safeguards Rule, many US tax preparers qualify as financial institutions and must maintain a written information security program.
What is shadow AI? Shadow AI refers to employees using unauthorized or personal AI tools for work without administrative oversight, which can risk leaking client data to public AI models. Firms mitigate this risk by enforcing zero-retention policies or routing extractions through a local Ollama companion using our local document AI workflow.

Original research sources and citations

The evidence register links each statistic to the published survey, report or government guidance used on this page. Check the source date and methodology before quoting a figure elsewhere.

Use accounting AI statistics without flattening the evidence

Record five details beside any number you reuse: publisher, fieldwork date, population, question wording and geographic scope. A survey of accounting professionals is not automatically a survey of firms. Reported use of any AI tool is not the same as a controlled deployment inside an accounting workflow. Market forecasts belong in a separate category from observed adoption or operating benchmarks.

Keep security figures in context too. An industry average breach cost does not predict the loss at one practice, and an incident share does not show the probability that a particular firm will be attacked. Use these numbers to ask concrete questions about access, approved tools, retention, encryption, backups and incident response.

For accounts payable, pair cost and cycle-time benchmarks with exception rate, straight-through processing and correction work. A lower processing cost means little if the team sends incorrect invoice data downstream or removes the evidence a reviewer needs.

Separate observed survey results from forecasts.
Do not compare percentages until the populations and questions align.
Keep the source year and publication date with every statistic.
Translate the number into a control, test or budget decision.

Turn the 2026 statistics into a short accounting security review

List every approved AI tool that can receive client or company documents. Record its owner, data types, access method, retention setting, model-training policy and deletion path. Then look for unsanctioned browser tools, shared credentials and exports that sit outside the firm's normal retention controls.

Choose one document workflow and trace it end to end. Check who can upload, which provider receives the file, how uncertain values are reviewed, where exports go and what the audit log records. Test account removal and deletion instead of relying only on policy text.

The accounting AI and data security statistics on this page help set priorities. They do not replace a risk assessment. Use the cited evidence to frame the review, then base the final controls on the firm's documents, systems, legal duties and tolerance for error.

Sources checked for this note

Keep reading

Try it yourself. Upload a PDF, scan, or image and let Dynamite Docs infer the schema.