By Sam Davidoff, CEO and Co-Founder, Align
The original article was published on Law.com, here.
Here is what is happening at law firms across the country. Clients are demanding that law firms use AI to bring down costs. The law firms respond by increasing their technology budgets and dedicating personnel to improve the use of AI across practice areas.
Those teams then research and sit through demos of dozens and dozens of AI products and decide which ones to pilot. At the end of the pilots, they pick one or two products to roll out firmwide, often based more on what they see peer firms doing than on internal analysis.
Is this the right way to capture the value from technology investment?
It’s clearly not, and you don’t need to take my word for it. McKinsey recently surveyed nearly 2,000 organizations across every major industry to answer precisely the question: how do organizations best capture value from new technology? It tested roughly thirty possible answers including bigger budgets, better models, stronger governance, more training, deeper talent benches, and executive engagement.
The thing that mattered most was whether the organization had fundamentally redesigned its workflows.
According to McKinsey, companies that had redesigned workflows were nearly three times more likely than their peers to capture meaningful EBIT impact. This probably sounds familiar. The same pattern has shown up across every major wave of legal technology in the last decade, including legal-research, document management, e-discovery, knowledge management. Firms that buy first and ask questions later get a fraction of the value. Firms that diagnose first capture nearly all of it.
There’s more. MIT’s NANDA initiative, analyzing roughly $30–40 billion in enterprise generative AI spending across 300 deployments, found that 95 percent of pilots produced no measurable P&L impact. The reason is simple. Piloting to find out if a product will help you is shooting first and asking questions later. You’re buying the tool before you know what problem you are trying to solve. It’s the same thing McKinsey’s study found.
Want more evidence that this problem exists at law firms? Thomson Reuters’ 2026 Peer Monitor Report found that average law firm technology spending grew 9.7% in 2025, the fastest growth rate on record. That’s what I noted above. Clients demand AI; firms increase their budget.
What’s harder to find is evidence that the spending is producing what it promises. Wells Fargo’s Legal Specialty Group reported that revenue growth at major firms in the first half of 2025 came almost entirely from rate increases—standard billing rates climbed 9.2%—while demand for legal services actually contracted 2.1%.
Clio’s 2025 Legal Trends Report found that the average lawyer still bills three hours of an eight-hour day—a 38% utilization rate. BigHand’s 2025 Legal Workflow Report found that 87% of firms still delegate work manually, and 31% report that their lawyers are doing more administrative work than they were a year ago, not less.
Let’s put that together. Law firms are spending more on technology, and law firms are making more money. But the two are entirely disconnected. The “spending more on tech” isn’t leading to “less busy work” (what the clients want). And the “making more money” is coming from increased rates, not efficiencies.
No lawyer would let a client buy litigation strategy off the floor of a conference. The premise of legal practice is that you assess the matter first, identify what’s actually at issue, and then decide how to attack.
And yet this is precisely how technology decisions get made at most firms. A managing partner sees a demo, hears that a peer firm has the same product, and authorizes the purchase. The diagnostic step—what is the actual workflow problem this is meant to solve and is it the one most worth solving—gets compressed into whatever can be inferred from a thirty-minute presentation or a pilot where the participants are often nonrepresentative, tech-forward lawyers. The result is predictable: tools that solve real problems, but rarely the most important ones, layered on top of a workflow no one has examined in years.
The firms that get this right work in the other direction. They start with the question that should always come first: where, specifically, are our lawyers losing the hours they can’t bill? Or relatedly, where can we cut hours that our clients shouldn’t have to pay for (even if they currently do). Not in the aggregate, but in the day-to-day shape of the work. Which hand-offs are taking too long? Which coordination tasks are falling on the wrong people, at the wrong hours, in the wrong way? Where is the ratio of time-spent to value-delivered the worst, and what is structurally causing that?
Those are workflow questions. They produce technology decisions, but only at the end, and only after the harder analysis is done.
Consider two litigation departments at firms of similar size and profile. The first commits roughly $300,000 a year to an enterprise AI deposition platform (I’m making up this tool to avoid picking on any product in particular, though I wouldn’t be shocked if such a thing exists). It performs as advertised. But deposition workflow isn’t where this litigation team is actually losing hours—and no one asked that question before the contract was signed.
The second department spends a quarter of that budget. But before it spends it, the firm reviews its time keeping records for the past two years and asks the following question: in the seventy-two hours before a major deposition, hearing, or filing, where is time spent and what amount of that time gets written off? The answer is consistent. It is not reading documents. It is not writing outlines. It’s not summarizing depositions. It is the coordination layer—getting the right version of every document into the right hands, on the right device, in the right order, when the partner reorganizes the binder at 9 p.m. the night before. That layer is invisible in any conference demo. It is also where the hours go.
The second department buys a tool that addresses what the diagnosis surfaced. The first buys a tool that addresses what was being marketed. A year later, only one of them has measurably moved its utilization rate.
The example above is just an example. It is a real one, however. I’ve focused particularly on this issue of predeposition and pretrial work, and see the same problems over and over:
These are not problems that you are going to find an answer to sitting through AI-product demos or in AI-based solution at all.
The overall point is bigger. It’s easy to sit around and speculate that this or that product (AI or otherwise) will bring value. It’s harder to identify in advance the places you are losing value.
Hard, but not intractable. You have the data. It is in firm billing and accounting records. The analysis of this information doesn’t require legal training—though it does require a good understanding of how legal teams work. Also, while AI isn’t the solution to every legal problem, it is a good solution to a lot of these analysis problems. Your AI model and harness of choice will do quite a good job of helping you break down and categorize your accounting and billing records to answer these questions. (And no, you don’t need a specialized, legal-specific tool to do that).
That is what a workflow-first diagnosis looks like in practice. It is not glamorous. It is also the only honest way to know whether the next technology purchase will actually move the number.
There is a reason McKinsey’s finding generalizes across industries. Workflow redesign is harder than software procurement, which is why so few organizations actually do it. It requires firms to look at how work moves through the organization with the same rigor they apply to a client matter. It requires partners to ask uncomfortable questions about where their own hours go. And it requires acknowledging that the most expensive inefficiencies are usually not the ones that everyone is talking about or that the most heavily marketed products are addressing.
Technology can solve real problems. But it has to be pointed at the right ones. The firms pulling ahead in this market are not the ones with the largest tech stacks. They are the ones who made the diagnosis first and chose the tools second.
Sam Davidoff is the CEO and co-founder of Align, a legal technology company focused on litigation workflow and trial preparation. He previously practiced as a litigator at Williams & Connolly.
Align is the only digital case binder platform built exclusively for litigators. Founded by a former top-tier litigator, Align delivers a familiar, secure, and reliable way to prepare, organize, and present your cases.
.png)
