- Clio’s June 2026 report found 65% of firms using AI save five hours weekly per professional, recapturing $1,500 in billable capacity.
- AI’s time savings challenge hourly billing, leading firms to shift to alternative fee arrangements for profit growth.
- Despite advanced AI tools, a 2026 industry report shows 43% of firms still lack a formal AI policy.
AI is forcing law firms to confront a problem built into their own business model. Clio‘s June 2026 Legal Trends Report found that 65% of firms actively using AI tools are saving roughly five hours per professional each week, worth around $1,500 in recaptured capacity at a $300 billing rate. That sounds like a win, but under hourly billing, it is also a direct revenue cut. The firms gaining the most from AI are the ones restructuring their pricing before the arithmetic catches up with them.
A recent industry analysis found that 65% of firms actively using AI tools are saving roughly five hours per professional each week, which, at an average billing rate of around $300 per hour, represents approximately $1,500 in recaptured capacity weekly.
The Billable Hour Problem
For decades, the billable hour has been the foundation of legal pricing. AI is cracking it. When a task that once took four hours now takes one, a firm billing by the hour must charge for that one hour, and eats the difference. That is not a hypothetical: it is the arithmetic reality many firms are already working through.
Firms using AI-powered document review have seen review time cut by more than two-thirds, translating into significant annual cost savings.
The Shift to Alternative Fee Arrangements
The firms moving fastest are restructuring around alternative fee arrangements: fixed fees, capped matters, and value-based pricing. Under these models, AI efficiency flows directly to margin rather than evaporating as lost billable hours. A matter that once required 40 hours of associate time and now requires 12 still bills at the agreed fixed fee. The ten hours saved are profit.
Outcome-based pricing in professional services is poised for significant growth over the next decade, as AI reduces the delivery cost of those outcomes.
Where AI Is Being Deployed
A 2026 Relativity webinar poll found it was the top-cited area of AI impact among attendees, with document review and legal research also cited by a large share of respondents, alongside summarisation.
The platforms firms are using have matured considerably. Thomson Reuters‘ CoCounsel, integrated with the Westlaw database, covers research, document review and drafting. LexisNexis+ AI, anchored to the LexisNexis corpus, offers natural language querying, contract analysis and summarisation, with a stated emphasis on source attribution to reduce hallucination risk. Both represent a departure from general-purpose AI: they are trained on legal datasets and return verifiable citations to primary law, which matters when the output could end up in front of a judge.
Harvey has seen rapid adoption among AmLaw 100 and Magic Circle firms, supporting contract analysis, regulatory research, deal diligence and drafting, and allowing firms to load their own precedents and style guides. For in-house teams, platforms like GC AI are positioning as end-to-end solutions, covering contract review, legal research, regulatory analysis and outside counsel management within a single interface. In a May 2026 In-House Legal Bench test, GC AI is reported to have scored 86.8% across 100 in-house legal tasks, outperforming general-purpose AI platforms. LegalOn is specifically focused on contract review, reporting reductions in review time of up to 85% through automated risk flagging and redline generation based on attorney-built playbooks, according to the company.
The Governance Gap
Adoption is running well ahead of governance. A 2026 industry report from 8am found that more than half of respondents said their firm had provided no training on responsible use of generative AI and had no plans to introduce any. 43% of firms had no formal AI policy at all; only 9% had an enforced written policy. That gap is not just an operational risk. Unverified AI outputs and unsecured client data create direct exposure under professional conduct rules.
The risks are not theoretical. Courts have already sanctioned lawyers for submitting fabricated case citations generated by AI, prompting firms to introduce verification protocols, mandatory human review and disclosure policies. The American Bar Association’s Formal Opinion 512 clarifies that competence, confidentiality, client communication, supervision and reasonable billing remain duties when lawyers use generative AI. That opinion provides a framework, but a policy document is not a substitute for firm-level training, which most firms still have not delivered.
The gap between individual attorney adoption and institutional readiness is where the real risk sits. An associate using an AI tool without guidance on data handling or output verification is a liability regardless of how capable the tool is.
What Changes for Legal Professionals
AI is not eliminating legal roles. It is redistributing where lawyer time goes. Routine document-intensive work, manual review, first-pass research, intake processing, is increasingly handled by AI, which frees attorneys for work that requires judgment: strategy, client counsel, negotiation and advocacy. That shift is also visible in hiring. Demand is growing for professionals who combine legal knowledge with working familiarity with AI systems, sometimes called legal technologists or legal data intelligence specialists, reflecting the skills now needed to bridge legal reasoning and AI-assisted execution. This links to a broader pattern in enterprise AI adoption: as the SHRM survey on AI rollouts found, the firms managing this transition most successfully are the ones that communicate changes to staff before implementation, not after.
The economics are straightforward. Firms that continue billing by the hour for work AI can do in a fraction of the time will face client pressure, audit scrutiny and margin erosion. Firms that restructure around fixed fees and value-based models can treat AI’s time savings as direct profit. Clio’s data quantifies the efficiency gain; whether a firm captures it or loses it is a pricing decision, not a technology one. For more analysis on enterprise AI strategy, visit our Enterprise AI section.
