For years, the default justification for legal technology has been time savings. How many hours does it eliminate? What's the cost reduction? Where's the efficiency gain? At LegalWeek 2026, multiple sessions challenged that framework head-on, arguing that conventional ROI metrics—cost reduction, efficiency, time savings—are insufficient because AI often changes workflows rather than simply accelerating them.
The Expanded ROI Framework
One panel proposed an expanded ROI evaluation that encompasses strategic capability development, quality improvements, risk reduction, competitive differentiation, and new capabilities AI makes possible. The shift in framing was captured in a single question: instead of asking "How much time does AI save?", firms should be asking "What new capabilities does AI make possible?" That's a fundamentally different lens, and it changes the business case entirely.
Traditional ROI Metrics (insufficient for AI):
Cost reduction and efficiency gains Time savings per task Labor arbitrage
Expanded ROI Metrics (necessary for AI):
Strategic capability development Quality and accuracy improvements Risk reduction and defensibility Competitive differentiation New capabilities made possible Institutional knowledge preservation
Understanding Total Cost of Ownership
A session on Total Cost of Ownership added nuance to the investment picture. The TCO formula presented—License plus Operational Cost plus Maintenance—spans acquisition and implementation, ongoing operating costs, extensive training, AI governance overhead, risk and error costs, platform flexibility considerations, vendor lock-in risks, and the often-overlooked cost of policing shadow AI. The panel stressed that AI adoption is fundamentally about people—accelerating adoption, reducing shadow IT risk, supporting talent strategy, and creating internal AI literacy.
The biggest cost of AI adoption isn't the software. It's training, governance, and managing people's expectations about what the tool can actually do.
What made these sessions particularly valuable was their honesty about the current state of measurement. Multiple panelists acknowledged that measuring AI's business impact is currently more art than science. Structured evaluation frameworks are essential but still under active development across the industry. The key is not to wait for perfect metrics before investing—it's to invest with a broader definition of value that includes capabilities, quality, and risk reduction alongside traditional efficiency metrics.
The Investment Decision Framework
A practical investment decision framework was presented covering must-have versus optional tools, vendor suitability, and scale and payback analysis. One of the more provocative themes was AI becoming "invisible" embedded infrastructure within workflows—something that just works in the background rather than requiring active engagement. When AI reaches that stage, measuring its ROI becomes less about isolated time savings and more about overall matter outcomes.
The firms that are measuring ROI most effectively are the ones that have moved beyond task-level metrics and are looking at matter-level outcomes. How much faster do you reach settlement decisions? How much better is your early case assessment? How many fewer surprised discovery disputes do you have? These are the kinds of metrics that actually move the needle.
The Paradox of Capability
One session highlighted an interesting paradox: as AI reduces friction, stakeholders generate more legal work, not less. Contracts that were too small to review now get reviewed. Questions that were too expensive to research now get researched. This is a feature, not a bug—but it means the ROI of AI isn't always captured in reduced hours. It's captured in expanded capability. You're doing more legal work with the same team, which means higher quality, less risk, and better outcomes.
This reframing is crucial for how firms think about AI investment. If you evaluate AI purely on time savings, you miss the real value. If you evaluate it on capability expansion—on what becomes possible that wasn't before—the business case becomes much clearer.
The Path Forward
The firms getting the most value from AI aren't the ones counting saved hours. They're the ones using AI to see the full picture of a matter—to identify risks earlier, make more informed strategic decisions, and carry institutional knowledge forward instead of rebuilding it. That's the kind of return that doesn't fit neatly into a spreadsheet, but it's the kind that actually transforms how litigation is practiced.
If your AI ROI analysis starts with "how many hours does this save?", you're asking the wrong question. Start with "what does this make possible?" and the answer becomes much more compelling.
This article draws on reporting from LegalWeek 2026, held March 9–12, 2026 in New York City. The views expressed are those of Advocacy.