The promise of generative artificial intelligence (AI) in legal practice is seductive: speed up document review, contract drafting, legal research, and thereby shave down hours billed. Yet the reality for many law firms is different. A recent survey by the Association of Corporate Counsel (ACC) and Everlaw found that nearly 60% of in-house counsel reported “no noticeable savings yet” from outside counsels’ use of generative AI. (Bloomberg Law News) Among those who did see some benefit, only 13% pointed to fewer billable hours and 20% to faster turn-around.  That suggests the headline of “AI slashes bills” is premature.

One major reason is that law firms remain where they always were: a patchwork of experiments instead of a unified transformation. The business model based on time-spent (“billable hours”) is deeply embedded. As a Harvard Law piece puts it, large law firms’ productivity gains from AI clash with the traditional billable hour model. (clp.law.harvard.edu) When a firm charges by the hour, there is a disincentive to reduce hours spent; improvements in efficiency don’t automatically translate to fewer billable hours. (2Civility) Until the billing model evolves, firms have less motivation to push AI’s full potential into cost-reducing workflows.

Compounding the billing-model friction is uneven adoption of AI across practices and firms. Some firms or practice groups test tools for document review; others do contract-drafting automation; many lag behind entirely. Legal tech firms struggle to sell their AI products to large law firms because the billable hour model skews incentives. (Legal.io) Put differently: the technology is advancing but the institutional deployment lags. An Everlaw survey showed lawyers enabling generative AI report saving up to 260 hours annually, but such gains don’t necessarily map to billable-hour reductions if those hours are reallocated rather than eliminated. (everlaw.com)

From the client side the pressure is mounting. Many in-house legal departments expect outside counsel to adopt generative AI tools. According to a survey by LexisNexis, 67% of in-house counsel said they expect their law firms to use these tools. (legaldive.com) Meanwhile the ACC-Everlaw data show that 64% of respondents expect to bring more legal work in-house because of generative AI. (everlaw.com) In short, clients are signaling change and may drive pricing shifts, even though many firms aren’t ready.

One more factor is the measurement gap. Even when AI is deployed, law firms struggle to track and demonstrate savings to clients. AI might reduce time on a task but still require review, validation, or supplemental work by senior lawyers—so billable hours don’t fall as expected. The SSRN article “How the Billable Hour Can Survive Generative AI” argues that hours may drop but other factors (rate, staffing, utilization) change to offset that drop. (SSRN) Thus efficiency gains aren’t automatically visible or bill-reducing.

Looking ahead, AI’s role may push pricing models to evolve. Several thought-leaders suggest the billable hour’s grip is loosening. For instance a Thomson Reuters article on “Pricing AI-driven legal services: The billable hour is dead, long live…” observes that generative AI may accelerate shifts to flat fees or output-based billing. (Thomson Reuters) Similarly, research from Wolters Kluwer points to 67 % of corporate legal departments and 55% of law firms expecting AI-driven change to the billable hour model. (Wolters Kluwer) The inconsistency across firms means we are in transition rather than arrival.

In sum, AI is real, law firms are adopting tools, and some work is faster. But the core obstacle to billable-hour reduction is structural: the business model built on hours, inconsistent deployment of technology, lack of measurement/discounting mechanisms, and a client-driven push for change from outside. Until law firms coordinate practice-wide workflows, redesign billing, and reflect AI-driven efficiencies in their invoices, clients will continue to ask “where’s the savings?” and firms will nod and say “we’re working on it.”

Well… on the bright side, at least we didn’t say “It depends.”