Generative AI is already reshaping legal practice—particularly across arbitration services and commercial dispute resolution. The question for practitioners is no longer if to engage with AI, but how and how quickly. AI is beginning to change how legal work is performed, how disputes are resolved, and how clients assess value in areas ranging from the arbitration process to broader AI in arbitration and mediation services.
At Legal Geek Europe, I led a workshop based on the American Arbitration Association’s Practicing Law Institute course, AI Transformation, which draws directly on lessons from the AAA’s internal and market-facing evolution. The session brought together practitioners from both law firms and corporate legal departments, with seven full tables of engaged participants focused on a shared challenge: moving from experimentation to execution.
The conclusion was clear. The legal industry does not lack awareness of generative AI. It lacks a structured, intentional approach to implementation.
From AI Awareness to Action in Legal Practice
Across all groups, participants described a similar starting point:
- “We’re using AI individually, but not systematically.”
- “Adoption is happening in pockets—mostly drafting, summarizing, and research.”
- “There’s no clear path forward at the organizational level.”
These responses reflect a broader pattern across legal practice. AI adoption is underway, but fragmented. Experimentation is occurring, but not coordinated. Without clarity on the underlying objective—whether improving efficiency, enhancing client service, or expanding access—organizations struggle to prioritize and scale.
At the AAA, early progress came from reframing the problem. Rather than asking where to deploy AI tools, the focus shifted to identifying points of friction within existing workflows—areas where AI could create meaningful leverage. This problem-first approach enabled targeted experimentation that could be evaluated, refined, and expanded.
Structural Barriers to AI Adoption in Arbitration and Dispute Resolution
Participants consistently identified barriers such as risk concerns, lack of time, reluctance to make an upfront investment, and limited AI literacy. However, discussion quickly revealed that these challenges are structural rather than situational.
Legal systems—including litigation, private arbitration, commercial arbitration, and broader dispute resolution frameworks—are designed around risk mitigation, precedent, and individual autonomy. These characteristics support high-quality outcomes, but they also slow adoption of new technologies.
More fundamentally, the traditional model of lawyer development depends on repetition. Foundational tasks—research, drafting, document review—serve as the primary mechanisms for building expertise.
As one participant noted: “There’s a generational difference: more established practitioners can be reluctant to share their models and templates, but the new generation is focused on instant delivery and collaboration.”
This tension is particularly relevant in arbitration and dispute resolution, where expertise is developed through repeated exposure to structured case work. As generative AI absorbs elements of that work, organizations must deliberately redefine how skills are developed and how value is created.
At the same time, a gap in AI literacy persists. Many practitioners are experimenting with AI tools, but relatively few understand their capabilities and limitations in depth. This produces skepticism of output and uneven adoption—overconfidence in some areas, hesitation in others.
Structuring AI Innovation Across Arbitration and Mediation Workflows
A consistent theme in the workshop was the absence of structure.
Participants described environments where:
- ideas are generated but not captured
- experimentation occurs but is not shared
- use cases are identified but not advanced
The AAA’s approach offers a contrast. AI transformation was treated as an organizational capability, supported by:
- distributed experimentation across teams
- centralized mechanisms for idea intake and evaluation
- regular governance to prioritize initiatives
This model enables broad participation while maintaining alignment with institutional objectives.
Equally important was a cultural shift from “quiet experimentation” to shared learning. As practitioners began exchanging use cases and insights, adoption accelerated. This transition—from individual use to collective capability—is essential for scaling AI across complex environments such as arbitration and mediation.
High-Impact Use Cases in AI-Assisted Arbitration and Commercial Disputes
Participants identified a range of use cases that fall into two categories.
Efficiency-focused applications:
- drafting arbitration agreements and clauses
- summarizing case materials
- improving internal knowledge access
Strategic applications:
- testing legal arguments
- identifying overlooked risks
- supporting decision-making in complex disputes
This second category is particularly significant. It reflects a shift from using AI solely for automation to using it as a partner in reasoning and analysis.
The AAA’s own experience illustrates this progression. In arbitrator selection, AI reduced time while improving outcomes by surfacing a broader pool of candidates. Human judgment remained central, but the quality of inputs improved.
Similarly, tools such as ClauseBuilder AI, for dispute resolution clause-drafting, required a shift in mindset. Because generative AI is probabilistic, outputs are inherently variable. Rather than seeking perfect accuracy, the organization defined thresholds for usefulness and implemented governance mechanisms—human review, disclaimers, and controlled rollout—to manage risk.
This reflects a broader principle: AI adoption requires moving from risk avoidance to risk management in service of value creation.
Expanding Arbitration Services Through AI-Enabled Innovation
The AAA’s transformation has extended beyond internal operations into the development of new capabilities for the market.
One example is the creation of what can be described as an AI-enabled arbitrator, an early instance of AI arbitration in practice. This capability is not intended to replace human adjudication in complex cases. Instead, it expands access by enabling faster, more scalable resolution for lower-complexity disputes—particularly where traditional processes may be cost-prohibitive.
This represents an important shift. AI is not only improving existing arbitration processes. It is enabling new forms of dispute resolution.
For many participants, this perspective was new:
- “We’ve been focused on internal efficiency—we haven’t thought about this externally.”
Organizations that approach AI intentionally—aligning internal transformation with new service models—are better positioned to create differentiated value.
Client Expectations and the Future of AI in Arbitration and Mediation Services
Participants from corporate legal departments emphasized that client expectations are accelerating change.
Organizations are seeking:
- greater efficiency in arbitration processes
- increased transparency in cost and timelines
- clearer articulation of value
This is contributing to a gradual shift away from time-based billing toward value-based models. As AI reduces the time required for certain tasks, the link between effort and value becomes less direct.
In this context, co-designing AI solutions with clients—through pilots, feedback loops, and shared experimentation—emerges as a critical strategy. AI initiatives that are not aligned with client needs risk becoming internal optimizations with limited external impact.
From Intention to Execution
The workshop concluded with participants committing to a single action to be undertaken within the next 30 days:
- “Build an agent based on procedural handbooks to reduce repetitive questions.”
- “Map the corporate lifecycle system and test AI against it.”
- “Streamline the process of building monthly board minutes” and “arbitrator appointment.”
- “Auto populate forms with data from correspondence files.”
These actions are intentionally modest. AI transformation does not begin with large-scale programs. It begins with structured, targeted experimentation.
The central lesson is that most legal organizations already possess the foundational elements required for transformation. The constraint is not capability, but coordination.
Closing the gap between intention and execution—even in a single use case—creates momentum. Sustained over time, that momentum becomes transformation.
What Practitioners Should Do Next
For practitioners working across arbitration, mediation, and broader dispute resolution:
- Identify one high-value use case
- Define success criteria clearly
- Establish appropriate governance and acceptable risk
- Begin testing
The organizations that lead in AI in arbitration and legal practice will not be those with the most tools, but those with the most intentional approach to applying them.