Frontline Managed Services released its 2026 Law Firm Performance and Future Readiness Benchmark last month, and Legaltech News covered the headline. Average technology spend per lawyer is set to rise from $33,720 in the current fiscal year to $38,460 in the next, an increase the report puts at 15.7 percent. Of the 150 firm leaders surveyed, 76 percent plan to increase per-attorney spend next year. Not a single one plans to cut it.
That is the right instinct, and the same survey backs it up. Firms are getting real value out of what they have already bought. Seventy-four percent have seen operating costs come down over the past twelve months. Seventy-one percent report improved client satisfaction. Seventy percent have lowered operational or cybersecurity risk. The time savings are landing too: more than a quarter of respondents reported AI-driven savings of 25 to 49 hours per attorney per year, an even larger group reported 50 to 99 hours, and more firms came in above 100 hours than below 25.
Anyone still arguing that legal AI is a solution in search of a problem is arguing against the data. The tools work. Spending more on them is a defensible decision.
Then you reach the two numbers that should stop a managing partner cold. Over those same twelve months, only 38 percent of firms saw matter profitability improve. Only 23 percent saw revenue per lawyer go up.
So the benefits are broad and the financial return is narrow. Three quarters of firms are cutting costs and saving hours without moving the number that actually describes the health of the business. That gap is not an argument for spending less. It is an argument for spending differently, and the survey is unusually clear about where the difference lies.
The survey found the reason and buried it
Ninety percent or more of respondents explicitly include AI adoption in their technology strategy. Eighty-one percent include data and analytics. Seventy-one percent include cloud migration and optimization.
Fifteen percent include a plan for reinvesting the time and capacity that technology frees up.
That single statistic explains most of the gap. Firms are successfully buying hours back and then losing track of where those hours went. The savings are documented. The destination is not.
An hour saved is not an hour earned. Once it exists, it becomes one of three things. It becomes billable work the firm would not otherwise have captured. It becomes business development or client service that grows the book. Or it becomes nothing, absorbed into the day without a trace. Absent a decision made in advance, it defaults to nothing.
That default is exactly what the 23 percent figure is measuring. The 15 percent who planned for it are a very good candidate for the 23 percent who got paid for it.
Organized is not the same as accountable
The maturity data looks encouraging at first. Every firm that answered the strategy question had a defined technology strategy in place. Sixty-three percent had strategies with named initiatives, owners, and timelines.
Only 22 percent had strategies tied to business outcomes, investment decisions, and executive review.
So roughly four out of five firms are running technology programs that are organized but not accountable. There are owners. There are timelines. There is no line drawn back to what the firm is trying to earn. And for firms in the 50 to 100 attorney range the picture is sharper still, with only 7 percent reaching that most advanced tier against far better numbers from their larger peers. The firms with the least margin for a wasted investment are the least likely to be measuring one.
This is what tool-first buying looks like from the inside, and it rarely feels reckless. It feels responsive. A competitor makes an announcement. A partner forwards a demo. A vendor runs a genuinely impressive pilot. The firm buys, deploys, trains, and moves on to the next thing, and every one of those steps is defensible in isolation. What never happens is the step before them, where someone writes down what this specific purchase is meant to change about the firm's economics and how anyone will know whether it did.
Look again at what firms agree on. AI adoption, cybersecurity, analytics, cloud. Those are the four items that were already obvious, the ones in every strategy deck in the industry. The two almost nobody is working on are technical debt at 29 percent and reinvestment of freed capacity at 15 percent. Those two are what determine whether the other four pay for themselves.
The herd converges on the categories that are easy to agree about. The return lives in the categories that are not.
Strategy First means the number comes before the purchase order
Strategy First is the first of Adjuria's four tenets, and it is not a statement about planning in the abstract. It is a statement about sequence.
When we scope an engagement, the opening question is never which tools you want. It is what you want your firm to be able to do that it cannot do today, and what that capability is worth in hours, in capacity, and in margin. Every initiative we recommend carries an ROI projection. Every phase is scoped and priced on its own, so you can see what the last phase actually returned before you authorize the next one. That structure exists so the conversation at the end of a phase is about evidence rather than about faith.
We also plan the part that 85 percent of firms are skipping. If a new workflow gives a partner four hours back every week, we decide in advance where those four hours go, and we write it into the roadmap. Maybe it is additional matter capacity without additional headcount. Maybe it is the client development work that has been sliding for two years. Maybe it is a change in how a practice group prices its work, because the cost of production dropped and the billing model should reflect that. Any of those can be the right answer. Leaving it undecided is the one answer that reliably produces the 23 percent outcome.
That is also why our consulting practice and our platform travel together. Strategy without a delivery capability produces a roadmap nobody executes. A platform without strategy produces exactly what the Frontline numbers describe, which is real savings and an invisible return.
A comprehensive platform changes where your budget goes
There is a cost in most legal AI engagements that never appears in the quote. The first several months go to building the obvious things, the capabilities every firm needs and no firm is distinguished by. Document analysis. Drafting. Conflict checks. Intake. You fund that groundwork, you wait for it, and when it is finished you have parity.
We did that work already, and we did it deliberately so our clients would not have to pay for it again.
AI Associate arrives with the common work covered. Document analysis and drafting with multi-model QA built into the revision flow. Citation checking that pulls real case metadata without exposing it to a model, so hallucinated authority gets caught before a filing does. Conflict of interest review against your own maintained database, run manually or dropped into an automated workflow. Email analysis that routes messages to the right matter, extracts what it needs, flags new leads, and sends the boilerplate. A meeting organizer that works your calendar and contacts parties directly to find a time. Collaborative redlining, e-signature, and secure intake webforms that update matter records on submission. Web search and crawling as inputs to your analysis workflows.
The Frontline survey found that firms have already automated the front end of this list, with intake, collections, and conflict checking almost universally automated in whole or in part. Fair enough. The question is what the next dollar buys, and for most firms it buys another round of the same table stakes from a different vendor.
If your firm needs what most firms need, we treat that as configuration rather than construction. Which means the customization budget, the money and the time and the partner attention, goes to the work that is specific to you. The demanding client whose intake carries an extra approval step. The practice area with a filing calendar nobody else in the industry uses. The precedent bank that took thirty years to build. The report your managing partner actually reads on Monday morning.
Those are the parts that produce advantage, precisely because they are the parts your competitors cannot buy.
Count the hours, do not estimate them
Most of the firms in the Frontline survey are reporting their time savings from impression and memory. It is why the ranges in the report are so wide, and it is why the connection between hours saved and dollars earned stays fuzzy enough to lose.
AI Associate logs every task with time-saved metrics, and every client gets a monthly ROI report showing hours recaptured, measured against the baseline we take before rollout. Not a projection from the sales cycle. A running total from the work itself. When it is time to renew, or to expand into another practice group, or to justify the line item to the compensation committee, the argument gets made from a log instead of from a feeling.
The same discipline runs through the delivery work. We optimize the process, not just the task inside it. Automating one step of a workflow that contains three unnecessary steps gets you a faster version of the wrong process. So we map how matters actually move through your firm, including the exceptions and the handoffs and the part everybody quietly works around, then cut what should not be there before automating what remains. That is where the second and third rounds of savings come from, and it is why our engagements keep returning value after the initial deployment stops feeling new.
Staying at the peak of the curve
Technology spend in this industry is going up next year, and it should. The survey settles that question, and the firms raising their budgets are reading the evidence correctly. Lower costs, happier clients, less risk, and real hours back are all worth paying for.
The open question is the second half of the trade. Both of the curves that matter here, hours saved and dollars returned, have a peak, and right now most firms are climbing one of them while ignoring the other. You do not reach the top by spending first and measuring later. You do not reach it by waiting for the market to resolve itself while competitors build a two-year lead in capacity, either. You reach it by deciding what the hours are for before you buy the thing that produces them, then building on a platform where the common work is already finished so your investment goes toward what makes your firm worth hiring.
The averages in this report are what happens without that discipline. They are not a ceiling.
Talk to us about strategy and where AI genuinely fits your firm's goals, or request a demo of AI Associate to see what the everyday work looks like when it is already handled.