KPMG's AI audit agents cut up to 3.5 weeks from some audits
KPMG announced the U.S. rollout of KPMG Clara intelligence (KCi), which uses AI agents to draft and coordinate parts of audit work while auditors direct, review and decide. On some engagements, KPMG says timelines shrank by up to three and a half weeks after analyzing full populations of transactions, and clients pulled fewer support documents by hand. KPMG aims to bring KCi to all of its private company audits in 2027.
This story was produced through MarketScale. See how Business Services teams put it to work with Executive Thought Leadership.
Key takeaways
KCi's best reported time savings, up to three and a half weeks on some engagements, come from clients no longer hand-pulling individual support documents. Expect the first changes in how a finance team hands over data.
KPMG aims to run every private company audit on KCi in 2027, so planning conversations for those engagements are the right time to ask which audit areas will get full-population analysis.
Treat the three-and-a-half-week figure as a best case from selected engagements and selected audit areas, not as an average to build into a close calendar.
Free workspace
Turn your Business Services expertise into content.
Record interviews, organize footage, and write with AI on a free trial of the MarketScale platform for qualifying companies. No demo required, no credit card.
Up to three and a half weeks. KPMG says that's how much time its new AI audit agents have cut from some engagements. On those engagements, the firm says, clients no longer had to manually retrieve volumes of individual support documents.
KPMG LLP announced the U.S. rollout of KPMG Clara intelligence, or KCi, on November 17. In the KCi model, software agents draft and coordinate the work, and auditors direct it, review it and make the calls. The firm aims to put KCi into all of its private company audits in 2027, with a separate, phased rollout planned for public company audits. Taylor Portela of Inside Public Accounting put the scale in context: KPMG reported FY25 net revenue of $15.9 billion.
For the CFO of a private company KPMG audits, 2027 is the date that matters. What does an agent-assisted audit ask of the people being audited? And where does the saved time actually come from?
Agents at four points in the audit
KCi runs on KPMG Clara, the firm's global smart audit platform. According to KPMG's announcement, KCi isn't a separate tool. Its specialized agents sit inside the auditor's workflow and connect client data, technical accounting knowledge, risk assessment and substantive procedures.
The agents work at four points. In risk assessment and planning, they analyze engagement data, suggest preliminary risk considerations and draft first-pass documentation, so the team can spot risk profiles before year-end. In engagement orchestration, they track milestones and information requests, surface technical accounting guidance and run team dashboards. In transaction analysis, automated data pipelines review every transaction in selected audit areas and flag outliers, trends and anomalies. In testing, they automate parts of control testing and sample evaluation, and auditors review the results.
Clients will probably notice the coordination layer first, even though it's the least glamorous part. KPMG says the dashboards show client accounting teams more of the requests, milestones and engagement activity. A finance team would likely see a shared list of open requests in the first week, well before anyone sees an anomaly report.
KCi's footprint before the U.S. launch
KPMG LLP
Where the three and a half weeks come from
KPMG says that on some engagements, analyzing the full population of transactions cut timelines by as much as three and a half weeks. It also meant clients no longer had to pull large volumes of individual supporting documents by hand. Because the pipeline reviews every transaction in the selected audit areas, auditors begin with a list of outliers, trends and anomalies to look into.
Agentic AI tests enterprise workflows and org structure
- 16
- threads
- 7
- communities
- Cooling, down from 11
- momentum
- Positive 50%
- Neutral 17%
- Mixed 33%
- Negative 0%
- ai agents36
- anthropic4
- agentic ai1
Who is legally liable when an AI agent agrees to terms of service on a user's behalf?
Which sectors are moving fastest to restrict AI agent bookings?
How can organizations capture and transfer tacit expertise before automating a role?
How to read this
Counts are public professional discussion threads and posts observed in the Business Services listening index, classified for professional relevance; complaints, gossip and off-topic posts are excluded. Sentiment is per thread. Voices are paraphrased and aggregated; no individual, forum or platform is identified.
That could change what the audit needs from the company. If the pipeline reads complete transaction populations, the main handoff could shift toward complete transaction data and away from gathering individual support documents by hand.
In a sample-based test, auditors pick a subset of transactions and ask the client for documents supporting each one. Full-population analysis runs the whole set of transactions in an audit area through automated checks first. Auditors then follow up on the items flagged as outliers or anomalies.
The strongest objection is about scope. Unite.AI pointed out that the three-and-a-half-week figure is a company-reported result from selected work, not an average across engagements. Full-population analysis also covers only selected audit areas, and a flagged anomaly isn't proof of an error. A controller shouldn't take a month off the close calendar based on this announcement. The fair expectation is less document retrieval in the areas where KPMG turns on the pipeline.
Earlier questions, same signature
KPMG is changing how audits are run as well as the software behind them. It's tightening role definitions, moving supervision and review earlier in the audit, and standardizing how data is organized. Christian Peo, Vice Chair of Audit and Assurance at KPMG U.S., said in the announcement that auditor judgment can't be automated. He described the tool's job as giving auditors more time to ask where the risk is and what doesn't add up. The firm says KCi can support earlier discussions with teams and clients, and that spotting risk earlier gives teams more time to address areas requiring significant judgment before year-end.
Responsibility stays with people. KCi runs under KPMG's audit methodology, professional standards and its Trusted AI framework. Engagement teams can trace any output to its underlying data, evaluate it and corroborate it independently before it enters the firm's system of record. The underlying technology combines KPMG's own capabilities with enterprise tools from its relationships with Anthropic, Google Cloud, Microsoft and OpenAI.
Training is part of the rollout. KPMG Lakehouse hosted nearly 1,000 Audit and Assurance interns earlier this year. They trained in critical thinking, professional skepticism and how to evaluate AI-assisted work. The firm is also broadening its learning and development programs so engagement leaders can supervise AI-enabled audits and lead their teams as working methods change.
Research offers a useful check on fee expectations. A 2022 study in the Review of Accounting Studies used more than 310,000 resumes from the 36 largest audit firms to measure their AI investment. It found that a one-standard-deviation increase in recent AI investment was associated with a 5.0% lower likelihood of an audit restatement and a 0.9% drop in audit fees. The 17 Big Eight partners the researchers interviewed said quality was the main goal of audit AI, ahead of efficiency. The study came before agent-based tools like KCi, so it can't predict KCi's pricing. Still, it suggests that AI's measurable payoff in audit has shown up more in quality than in the size of the bill.
Questions for 2027 audit planning
For private companies whose KPMG fieldwork falls in 2027, the planning meetings now on the calendar are the natural time to ask how KCi will change their engagement. The useful questions are concrete: which audit areas get full-population analysis, what data access or extract format the pipeline needs, and how the request-tracking dashboard will be shared with the finance team.
Ask the engagement partner which audit areas KCi will analyze in full and what data feed it needs from your ledger. Those two answers tell you where document requests should shrink and where your team's prep work moves instead.
KPMG says it is planning a structured, phased rollout for public company audits. In the meantime, the signal to watch is simpler: whether time savings like three and a half weeks hold up as KCi grows from more than 125 U.S. teams to all of KPMG's private company audits in 2027.
Sources
- KPMG Unveils Next Era of AI-Enabled, Human-Assured Audit with U.S. Launch of KPMG Clara intelligence (KCi) ↗
- KPMG Launches Clara Intelligence in U.S. ↗
- KPMG launches AI-enabled audit platform in the US ↗
- KPMG Brings AI Agents Into U.S. Audits With Clara Intelligence ↗
- Is artificial intelligence improving the audit process? ↗ · Springer
Your experts belong here
Every story in MarketScale Business Services starts with a company putting its consultants, practice leads, and account teams on the record. Buyers are already reading this topic. The only question is whose experts they find.
Clients hire the firm whose thinking they have already read, which means fewer cold conversations for your partners.
About the author
The MarketScale Newsroom reports on the companies, technologies, and trends shaping 16 B2B industries. It turns primary sources and expert commentary into clear, useful coverage for the people doing the work.