Consilio’s 2026 survey finds leadership-driven legal organizations further along in AI deployment—and carrying heavier integration burdens. Moving beyond the pilot is only the first test of transformation.
For the first time in Consilio’s six-year global survey, the legal profession’s leading strategic challenge is not the volume of work. It is understanding, selecting and deploying the technology intended to relieve it.
Fifty-four percent of the 678 legal professionals surveyed cited technology selection and deployment as a primary challenge, narrowly surpassing the 52 percent who cited work volume. The margin is modest. The reversal is not. AI has not displaced legal’s capacity problem; it has added a technology-management problem to it.
Teams that were already stretched must now evaluate a crowded market, identify credible applications, coordinate with IT and security, train users and explain what the investment is producing. Consilio examines the role of executive pressure in that process in The AI Mandate in Legal, a companion analysis of legal, risk and compliance professionals from in-house departments and law firms around the world.
The report classifies 279 respondents as mandated because they selected “leadership vision” as a driver of change in their legal function; 399 did not. The label therefore captures leadership-driven adoption, not necessarily a formal directive from the board or C-suite. Within that definition, those organizations report greater progress moving AI from experimentation into production.
They also report more fragmented systems, more manual workarounds and less satisfaction with the technology environments they are building.
The mandate is producing movement. The harder question is whether the organization can absorb it.
Link to Moving AI into production Moving AI into production
Organizations in Consilio’s leadership-driven cohort reported higher full deployment across nearly every core legal use case examined.
Contract-management AI was fully deployed by 13.3 percent of those organizations, compared with 5.3 percent of the non-mandated group. Document review and analysis stood at 17.5 percent versus 9.9 percent. Low- or no-code automation produced the widest gap, at 6.7 percent versus 1.5 percent.
The multipliers are striking. The absolute figures are more sobering. Even where leadership was actively driving adoption, fewer than one in five respondents reported full deployment in any use case. When partial deployment is included, contract management reaches 34 percent among mandated organizations and 25 percent among the comparison group.
The one exception is telling. In intake and triage, non-mandated organizations report higher deployment rates. That kind of AI tends to be adopted in direct response to workload pressure rather than strategic direction—evidence that operational necessity can move an organization faster than an executive directive.
The survey does not establish that executive pressure caused those differences. Organizations with engaged leadership may also have stronger infrastructure, larger budgets or a greater appetite for change. What the findings show is that leadership-driven organizations are further along in putting AI into live workflows.
They also report better outcomes. Sixty-six percent cited increased efficiency and productivity, compared with 54 percent of non-mandated organizations. They were more likely to report lower operating costs, greater automation of routine work and improved decision-making.
“On paper, that looks like a clean win for leadership-driven adoption,” said Michael Pontrelli, Consilio’s chief client experience officer. “But unfortunately, what we’ve seen is trust doesn’t grow with it.”
The survey bears him out. Accuracy and trust concerns appeared at nearly identical rates in the two cohorts—49 percent of mandated organizations and 50 percent of the rest named them as barriers. The mandate accelerates deployment. It does not resolve the reasons people are cautious about what is being deployed.
Nor does a coherent operating environment necessarily grow with it. Production is not the end of the transformation. It is where the next set of problems becomes visible.
The report does not support the easy conclusion that executives are merely buying tools while leaving their organizations untouched. Seventy-six percent of the leadership-driven cohort said they had intentionally redesigned how they use AI, compared with 57 percent of non-mandated organizations. Half had automated workflows or processes, versus 30 percent. Forty-six percent had changed team structures, compared with 25 percent, while 39 percent reported a leadership or cultural shift—more than twice the 17 percent recorded in the comparison group.
These organizations are changing. They are also moving quickly enough to expose the weaknesses of technology environments that were not designed for the tools now entering them.
Half of the mandated cohort reported manual workarounds between systems, compared with 32 percent of non-mandated organizations. Forty-seven percent cited fragmented systems that do not integrate, versus 37 percent. Nearly one-third lacked a single source of truth, compared with 21 percent.
Only 10 percent of the leadership-driven group said they were satisfied with their current technology environment. Satisfaction among non-mandated organizations was higher, though hardly impressive, at 17 percent.
That is the tension at the center of the report: the organizations moving fastest are also carrying the largest integration burden.
Link to The integration bill The integration bill
A contract tool may accelerate analysis, a review platform may reduce document volume and a research product may shorten an assignment. But information still has to move between systems. Records may remain scattered across competing repositories, while permissions and responsibilities are divided among legal, IT, security and risk.
The pattern resembles integration debt. Each new product may improve a task while making the wider environment more difficult to manage.
“What we’re actually seeing is pockets of innovation, not complete transformation,” Pontrelli said. “It’s happening at the task level rather than the system level.”
That challenge is unfolding in a market Pontrelli describes as “frothy.” At a recent e-discovery conference, nearly every booth appeared to belong to an AI company—some built around the technology, others rapidly repositioning around it. The scene reminded him of the dot-com boom.
For legal departments, the consequence is not merely vendor risk. Teams are making architectural choices before many products have established a durable record. A tool that performs well in isolation may still prove expensive to integrate, difficult to govern or disruptive to replace.
The important buying questions therefore extend beyond what a product can do today. Legal leaders need to understand what systems and data it must touch, who will review its output, how it fits into the wider workflow and whether it can be replaced without dismantling the process built around it.
Those questions are less compelling than the demonstration. They are more likely to determine whether the investment survives.
Consilio’s strongest-performing survey cohort combined leadership vision with what respondents identified as an emerging technology opportunity. That 173-person group reported the highest deployment rates and strongest skills development in the survey. Twenty-three percent had fully deployed AI for document review, while 20.5 percent had fully deployed it for contract management. Eighty-three percent said they had developed AI-evaluation or prompt-engineering skills during the previous 12 months.
Consilio interprets the results as evidence that leadership performs best when its direction is accompanied by practitioner engagement and a recognizable technology opportunity. The contrasting mandate-only cohort points in the same direction, although its small size—just 34 respondents—makes the finding directional rather than definitive. Only 14 percent of that group reported developing process-design skills, compared with 19 percent of non-mandated respondents and 38 percent of those combining leadership vision with emerging technology opportunity.
“Leadership pressure creates the conditions for capability,” Pontrelli said, “but it doesn’t create any capability itself.”
A directive can create urgency, permission and executive support. It cannot teach a lawyer how to test an output, redesign a process or recognize when automation has moved a bottleneck rather than eliminated it.
Governance exposes the same limit. Only 6 percent of mandated organizations reported having no formal AI governance structure, compared with 19 percent of non-mandated organizations. Leadership-driven organizations were also more likely to use cross-functional committees.
They were no more likely to have reached maturity. Just 7 percent of the mandated cohort had a documented governance framework that was actively followed, compared with 6 percent of non-mandated organizations. Security concerns, expertise shortages and limited budgets likewise appeared at nearly identical rates in both groups, echoing the trust findings.
The mandate may overcome resistance from leadership. It does not create safety, expertise or funding.
“Governance can’t formalize around a target that’s still—and always will be—moving,” Pontrelli said.
Perhaps. But an organization’s obligations do not move with the technology. Legal departments still need standards for data access, confidentiality, validation, human review and accountability that can survive a change in product, provider or underlying model.
Link to Where strategy begins Where strategy begins
Consilio describes the missing coordination layer as orchestration: a common foundation connecting technology, data, governance and people, supported by a platform and partner strategy built around the legal function rather than a succession of isolated purchases. The idea follows logically from the fragmentation documented in the research: the next challenge is no longer whether legal teams can experiment with AI, but whether they can make its use coherent, accountable and sustainable.
It is worth noting that orchestration is also Consilio’s business: the firm that documented the fragmentation sells services designed to address it. The diagnosis does not depend on the cure, but readers should weigh the prescription with that interest in mind.
Pontrelli pointed to Consilio’s Innovation Lab as one practical response. Clients evaluate emerging technologies and use cases through an annual in-person meeting and virtual collaboration during the year. The model spreads the evaluation burden and can give participants a more informed recommendation to take back to their organizations when seeking funding for broader deployment.
That outside support can widen the available expertise and reduce duplicated effort. It does not transfer accountability.
A provider can assess tools, connect systems and help a legal department understand its options. It cannot determine the company’s risk tolerance, assign internal authority or decide when a lawyer must reject a machine’s work. Those responsibilities remain with the legal function, as does accountability for the outcome.
Deployment therefore cannot become the principal measure of progress. More licenses, pilots and automated tasks may indicate momentum. They may also conceal duplicate systems, manual handoffs and responsibilities no one has clearly accepted.
The harder work is deciding which processes should change, what information the technology may access, who owns its output, and how value will be measured across the legal function rather than within a single application.
A mandate can require a legal department to move. It cannot determine the destination or build the operating discipline required to get there.
Until those choices are explicit, the mandate is pressure looking for a strategy.
—
Link to Michael Pontrelli Michael Pontrelli
Michael Pontrelli is chief client experience officer at Consilio, where he leads the firm’s client experience strategy and works with corporate legal departments and law firms to navigate technology, operations, and complex legal matters. His work spans e-discovery, investigations, workflow design and emerging technology, with an emphasis on helping legal teams improve efficiency while strengthening integration, governance, accountability and the practical adoption of AI across global legal organizations.