Insights

The Success Metrics That Matter for Legal AI

How law firm leaders are moving beyond adoption to measure real impact with legal AI.

by Harvey TeamMay 19, 2026

In most law firms, the first question around legal AI is simple: Are people using it? Adoption rates, active users, and prompt volume are often the earliest — and most visible — signals of progress.

But as firms move beyond initial rollout, those metrics start to fall short. Usage alone doesn’t capture whether AI is improving the quality of work, changing how lawyers operate, or delivering better outcomes for clients.

Across our Innovation Spotlight series, we’ve spoken with leaders responsible for deploying AI at scale inside their firms. While their approaches to adoption vary, their thinking on success is notably more nuanced. Rather than relying on a single metric, they look at a broader set of signals, connecting usage to workflow integration and ultimately, client value.

Below, we highlight how four innovation leaders are redefining what success looks like for legal AI, and the signals they use to measure progress.

Start With the Outcome

For Al Hounsell, National Director of AI, Innovation & Knowledge at Gowling WLG, success with legal AI starts with the client. That means asking a distinct set of questions: Are we delivering higher-quality work? Are we responding faster? Are we reducing risk in meaningful ways?

Those outcomes then shape how success is measured. Instead of tracking Harvey usage in isolation, the team looks at indicators that map directly to client value: improvements in responsiveness, stronger and more consistent work product, and more reliable risk management. Adoption still matters, but primarily as a supporting signal. If AI is truly improving outcomes, it should show up naturally in how lawyers work day to day.

This creates a clearer link between AI investment and business impact. Success isn’t defined by how often the tool is used, but by whether it meaningfully improves the service delivered to clients.

Read the full conversation with Al →

Build a Multi-Layered Framework

At Mallesons, measuring legal AI success starts with structure. Chief Innovation Officer, Michelle Mahoney, and her team take a deliberately data-driven approach, defining clear goals and then breaking them down into specific objectives, each tied to a metric, a baseline, and a target. Rather than relying on a single headline number, this creates a system for tracking progress across the entire AI program.

That system spans multiple dimensions. The team tracks adoption and participation, but also looks deeper: at how frequently Harvey is used, how broadly its functionality is applied, and how usage translates into client engagements, value generated, and overall change impact across the firm.

Crucially, Mahoney draws a clear distinction between activity and impact. High usage alone isn’t enough. What matters is whether that usage leads to better outcomes — whether teams are forming new habits, working more effectively, and ultimately delivering more value to clients.

The key is to be clear about what goals you are seeking and what metrics are the most relevant indicators of success.

Michelle Mahoney

Chief Innovation Officer at Mallesons

To capture that, the firm combines both leading and lagging indicators. Early signals like training participation and usage patterns help track momentum, while downstream metrics like client impact and value creation show whether that momentum is translating into meaningful results. Success, in this model, comes from seeing the full picture of how AI is adopted, how it is used, and what it ultimately delivers.

Read the full conversation with Michelle →

Measure Momentum and Maturity

For Pierre Zickert, Counsel and Head of Legal Technology at Hengeler Mueller, AI success is less about initial adoption and more about whether usage becomes a habit. The team closely tracks recurring users and overall prompt volume, treating both as signals of momentum: are people coming back, and are they incorporating Harvey into their regular workflows?

But momentum alone isn’t enough. The next layer is maturity, and how usage evolves over time. Early experimentation is expected, but the real indicator of progress is when teams move beyond one-off prompts and start building structured, repeatable workflows. An increase in more sophisticated use cases, as well as the development of internal playbooks, signals that AI is becoming embedded in how work gets done.

Zickert also pays close attention to qualitative signals. Positive feedback, peer-to-peer sharing, and growing enthusiasm across the firm all point to a deeper level of engagement — one where AI is not just used, but actively championed.

Read the full conversation with Pierre →

Align Metrics to Time Horizons

One of the challenges with measuring legal AI success is that the definition doesn’t stay fixed for long. Steve Johns, Partner, Co-head Technology & Digital Economy at Hall & Wilcox, believes that success needs to be evaluated in stages. What matters at the beginning of an AI rollout is fundamentally different from what matters once it’s embedded, and different again from what matters at scale.

In the short term, the focus is on behavior change. Are lawyers using Harvey regularly? Has it become part of daily work rather than a one-off experiment? Metrics like active usage, frequency, and training completion help answer those questions.

As adoption grows, the focus shifts to integration. At this stage, success is more about how deeply AI is embedded, and whether it’s built into workflows, templates, and processes. The relevant metrics begin to change as well, emphasizing efficiency gains, faster turnaround times, and improvements in the quality and consistency of work.

Over the medium term, success means that AI is embedded in the business and integrated into core workflows, templates, and processes.

Steve Johns

Partner, Co-head Technology & Digital Economy at Hall & Wilcox

Over the longer term, the lens widens again. Success becomes tied to outcomes and differentiation: delivering better results for clients, developing more advanced use cases, and ultimately rethinking how legal services are delivered and packaged.

This approach recognizes that no single metric can capture success across the full lifecycle of AI adoption. The key is to continuously recalibrate, ensuring that what you measure reflects where you are in the journey.

Read the full conversation with Steve →

Want to dig deeper? Explore 10 metrics that show your firm’s AI transformation is happening. If you want to learn how Harvey can help drive meaningful impact across your legal organization, contact our team:

Request a Demo

Yes, I would like to receive marketing communications regarding Harvey's products, services, and events. I can unsubscribe at any time

For details about how we collect, use, and protect your information, please see our Privacy Policy.