AI in Law and Professional Services: Harvey to Big Four — 2026-09-18
This week, the legal AI sector saw significant product updates from Harvey and Intapp, while new regulatory guidelines in Germany and risk assessments in Australia highlighted the growing tension between AI efficiency and professional liability. Courts continue to penalize hallucinated citations, with new tools emerging to mitigate these risks, as Big Four firms navigate the shift from billable hours to AI-driven value models.
AI in Law and Professional Services: Harvey to Big Four — 2026-09-18
Top developments
Harvey Releases September Product Update for Personalization and Review Speed
On September 16, 2026, Harvey released its monthly product brief, introducing features that allow lawyers to personalize the AI's output style and accelerate critical document reviews. The update also enhances context awareness across broader systems and knowledge sources, aiming to move AI from individual task assistance to integrated workflow support. This reflects a strategic shift from generic LLM capabilities to tailored, firm-specific utility, addressing the need for consistency in high-stakes legal drafting.
New Tools Aim to Prevent AI Hallucinations in Court Filings
On September 16, 2026, reports highlighted the emergence of new technology tools designed specifically to help lawyers avoid sanctions for AI-generated fake citations. As AI adoption grows, attorneys are facing "fake cases that look real," prompting the development of verification layers that sit between the generative model and the final filing. This development is critical for law firms seeking to balance efficiency gains with the strict liability standards of judicial systems, where courts do not require proof of AI use to impose sanctions for inaccurate authority.
IDW Issues Binding AI Guidelines for German Audit Firms
On September 17, 2026, the Institute of Public Auditors in Germany (IDW) released two new binding notes regulating AI use in audit firms and companies. The guidelines clarify that while AI agents can assist in reviewing accounts and bookings, the professional responsibility remains entirely with the auditor. This move by a key regulatory body signals that the "Big Four" and mid-tier firms in DACH markets must implement rigorous human-in-the-loop protocols to comply with new standards, impacting how audit workflows are automated.
Big Four Face New Risks from Rapid AI Adoption
On September 17, 2026, The Australian Financial Review reported that rapid AI adoption by firms like KPMG and EY is creating new risk profiles for the Big Four. While firms trumpet their commitment to AI, the technology still requires human oversight, and the industry is grappling with how to manage the liability of automated outputs in tax and consulting services. This underscores a broader industry challenge: moving beyond pilot programs to firmwide transformation without compromising quality control or client trust.
Local view
Germany: The German legal tech community is closely monitoring the shift in billing models. A report from beck-aktuell on September 17, 2026, noted that legal departments are increasingly using AI, forcing external counsel to adapt or risk losing business. The article highlights that AI is no longer an experiment but a standard expectation for efficiency. Additionally, LTO (Legal Tribune Online) published a piece on September 15, 2026, detailing the legal pitfalls of AI in legal practice, emphasizing that efficiency gains must not compromise data protection or professional secrecy.
Context & numbers
- Sanctions Tracker: As of July 2026, the largest publicly reported penalty for AI hallucinations was a $110,000 sanction in a federal district court in Oregon for submitting 23 fabricated citations. Another tracker notes roughly $109,700 in combined sanctions in a recent case involving two lawyers.
- Valuations: Harvey is valued at approximately $11 billion with ~$300 million ARR, while rival Legora is valued at $5.6 billion with ~$150 million ARR.
- Adoption: 83% of lawyers are now using AI tools, according to recent statistics.
On the radar
- Intapp Celeste Plug-in: On September 18, 2026, Intapp launched a Celeste plug-in for ChatGPT Enterprise, signaling deeper integration of legal-specific AI into general-purpose enterprise platforms.
- AI Regulation Debates: Political alliances are shifting regarding AI regulation, with unlikely bedfellows like David Sacks and Lina Khan appearing in the same discourse, potentially influencing future compliance requirements for legal AI vendors.
This content was collected, curated, and summarized entirely by AI — including how and what to gather. It may contain inaccuracies. Crew does not guarantee the accuracy of any information presented here. Always verify facts on your own before acting on them. Crew assumes no legal liability for any consequences arising from reliance on this content.
Powered by
