The State of AI in Law and Professional Services in September 2025
Clients are moving first
In-house legal teams are no longer waiting for external counsel to show them what AI can do. Companies like National Grid and Accenture now use AI agents across routine work, saving measurable time and shifting their cost base. This is changing the brief. Clients are beginning to press firms to explain how AI is used on their matters, what governance is in place, and how efficiencies are passed through to pricing. The direction of travel is clear. Clients expect transparency and demonstrable value. Law firms that cannot evidence AI-enabled productivity will lose ground in competitive pitches and panel reviews.
Law firms under pressure to show value beyond advice
Several firms are experimenting with new business lines that look more like consultancy than pure legal practice. Some are acquiring AI boutiques or recruiting data scientists to own intellectual property and speed up internal adoption. Others are offering AI advisory services directly to clients. These moves are blurring the boundary between law and consulting. For managing partners, the question is how far a firm should stretch: building consultancy arms, partnering with technology providers, or sticking to legal advice while embedding AI in delivery. Each choice carries implications for conflicts, pricing, and culture.
Some firms are now going further by trying to develop and sell AI products directly to clients, positioning themselves in competition with technology providers. This is a risky path. Building products requires sales support, marketing, client success and constant maintenance, which are disciplines most law firms do not possess. Unless ring-fenced as a distinct business line with its own budget and operating model, the returns are likely to be marginal and the brand diluted if the product disappoints. More importantly, it risks confusing innovation with technology for its own sake. The starting point should always be the business problem clients need solved, not the allure of a new tool. Firms that treat innovation as business-led change, in pricing, governance, or delivery, will create real value. Firms that chase products without a clear problem to address may end up carrying more cost than benefit.
Platforms are entering legal service delivery
We are beginning to see technology platforms cross into regulated service provision. Lawhive’s acquisition of a UK law firm is one example of a legal tech company becoming an authorised provider. At the same time, AI-enabled legal service companies like Eudia are buying alternative legal service providers. This signals a new competitive dynamic. Delivery models that combine regulation, technology, and scale are emerging. For traditional firms, the risk is not just losing efficiency battles but being displaced by platforms that reshape how clients buy legal work altogether.
The Big Four and other consultancies are embedding legal AI
Consulting firms are rolling out AI agents internally and in client engagements. Deloitte’s partnership with Legora and its widespread use of in-house audit chatbots show how the Big Four are integrating legal-specific AI into their wider professional service models. McKinsey, BCG, and others are doing the same in adjacent domains. This threatens to normalise the expectation that professional services come bundled with AI-enabled efficiency and governance by design. For law firms, this creates pressure both in multidisciplinary practices and in pure legal competition.
Regulation and privilege risk are front of mind
New national laws in Europe, such as Italy’s AI statute and Ireland’s designation of AI Act enforcement authorities, show that AI regulation is moving from draft to enforcement. UK regulators and judges are also beginning to signal how AI use should be disclosed in practice. At the same time, cases in the US and Australia highlight the risk of fake citations and of privilege being undermined when generative AI tools are used without proper safeguards. For firms, this makes governance a client-facing issue. The ability to evidence control, audit, and supervision will become as commercially important as the work itself.
Determining the value of AI investment
Much has been written about law firms and corporates investing in AI, but fewer have been able to point to tangible returns. Too often the story is anecdotal: “some of our lawyers like it,” or “it frees up hours each week.” These comments suggest greater efficiency but do not in themselves create more revenue or profit. Making lawyers’ lives easier is welcome, but unless the firm adapts its model, the financial benefits are rarely captured.
The missing link is that technology is often deployed before the business case has been resolved, and without the supporting elements that allow efficiency gains to turn into economic value. Unless the firm adapts its structures, the investment will remain under-realised. Key shifts include:
-
- Business model adjustment. Freed capacity must be converted into additional matters, reduced write-offs, improved fixed-fee margins, or shorter WIP to cash cycles. Without this flex, time saved simply dissipates.
- Incentive alignment. Partners need reasons to adopt AI in ways that improve margins and not just convenience. This may mean changing how profitability is measured and rewarded.
- Talent management. As routine work diminishes, training structures must be rethought. Junior lawyers need opportunities to learn through oversight, prompt and review skills, and higher-value tasks.
- Team structure. Firms will need to move from traditional pyramids to flatter models, supported by new skills and roles.
- Client engagement. Firms must actively demonstrate to clients how AI delivers measurable value, both to reinforce trust and to capture greater share of wallet.
- Process redesign. Any work that can be systematised should be pushed down to lower-cost resources, with the cost base adapted quickly.
The conditions for value are straightforward. Value appears when at least one of the following is achieved: more matters are completed with the same cost base, write-offs are reduced in hourly work, margins on fixed-fee engagements improve, WIP is converted to cash more quickly, or risk-related costs fall. Where benefits are strategic rather than immediate, they should be treated as protected revenue or avoided cost, with lead indicators such as panel wins, lower claims, or reduced attrition.
Measurement itself must be part of deployment design. For each AI use case, firms should lock three or four KPIs before rollout: cycle time to first draft or key milestone, realisation and write-offs, quality delta such as error rate or clause compliance, and at least one client or risk metric. Without these baselines, even successful deployments will struggle to show impact.
The correct approach is to begin with the business proposition, not the promise of efficiency for its own sake, and then deploy technology to support that proposition. Deploying AI without this is unlikely to generate measurable returns – something many firms are now coming to realise.
Talent and the future pipeline
There are early signs that AI is reshaping recruitment. PwC UK has already cut its graduate intake, citing structural change in how work is done. Surveys suggest job postings in AI-exposed professions are slowing compared to less exposed roles. For law firms, this raises two questions. First, how to train juniors when routine work is reduced. Second, how to develop new skills around supervising, verifying, and integrating AI. The firms that can solve for talent development in an AI-augmented world will have a lasting advantage.
Final Word
AI is a structural force reshaping how legal and professional services are delivered, priced, and judged. Clients are already moving faster than many firms. Competitors, of all sorts, are redefining the edges of the market. Regulators are beginning to draw the boundaries.
For managing partners, the choice is less about whether AI belongs in the business and more about what kind of firm will emerge once it is embedded. Some will reduce costs but fail to capture value. Some will chase technology but lose sight of client problems. A smaller number will reshape their models, align incentives, and convert efficiency into strategy.
The firms that endure will not be those that waited for clarity but those that created it.