What Legal Teams Wish They’d Known Before Deploying AI

Not long ago, firms were focused on whether to experiment, where to begin, and what guardrails might be needed to initiate AI deployment. Today, the question is more practical and more pressing: how can legal organizations turn AI into meaningful, durable business impact?  

That question was at the center of a recent ILTACON panel led by Simon Weierman, Senior Director of Segment Management at LexisNexis, featuring Vishal Agnihotri, Managing Director of Innovation & Knowledge at Alston & Bird; Cindy Bare, Chief Data & Innovation Officer at FBT Gibbons; and Krista Ford, Director of Knowledge, Research & Information Services at Steptoe & Johnson.  

The next phase of adoption will not be defined by a single model, a single platform, or a clever prompt. It will be defined by how well firms prepare the conditions around AI technology. 

AI Exposes the Need for Operational Change    

The most revealing effect of AI may be its ability to show firms what was already uneven beneath the surface. Fragmented knowledge practices, siloed information, and highly individualized ways of completing the same task can be manageable when work stays manual. They become much harder to ignore when a firm tries to scale an AI-enabled workflow.  

When a team can see that a shared process accounts for most of the work while a smaller portion varies by individual preference, it has a clearer starting point for a thoughtful conversation about standardization. That kind of visibility can strengthen more than technology adoption; it can improve process design, legal project management, resource planning, and eventually, the ability to understand the true cost of legal services. 

The Promise of AI Is Only as Strong as the Information Behind It 

In an enterprise legal environment, that expectation meets a more complicated reality. Important knowledge may sit across multiple systems, be protected by permissions, or require careful information governance before it can be safely used in an AI experience. 

This is why data readiness has become an increasingly strategic issue. “As we start bringing these information repositories together, the output is going to be a lot better, and then we can start learning and training and deploy better agents as we connect all of our data sources,” said Ford. Firms that have invested in bringing information together, clarifying ownership, and improving access controls have a stronger foundation for using AI to understand their experience and make institutional knowledge more actionable. The work can be demanding, but it unlocks value that was previously difficult to see or use.  

Security is a core part of trusted legal AI. Better outputs require the right information, but also the right controls around who can access it, how it is handled, and what client expectations apply. 

Governance Is an Ongoing Discipline 

New AI features are arriving constantly, often embedded in existing tools. Each one can introduce questions about retention, vendor risk, training, client consent, and the way a capability interacts with the rest of the technology stack. 

That means governance must be a repeatable, cross-functional rhythm involving innovation, information security, knowledge management, IT, legal, and business. The right pace will differ by firm and by client. Some clients are actively asking their outside counsel to use AI; others require a more cautious approach. A practical framework acknowledges that variation while giving attorneys and staff clear boundaries they can follow with confidence. “We've sort of opened that aperture, and we continue to open that aperture as new features come along. So, there's an Infosec conversation, there's an AI governance conversation, there's re-education of our own users, our own attorneys, because every feature may work very differently,” said Agnihotri.  

Adoption Is as Much a People Challenge as a Technology Challenge 

There is a growing burden on firms to vet new capabilities while also helping attorneys understand what has changed, how to use it effectively, and where human review remains essential. 

Clear communication matters just as much when a tool has not yet been approved. Shadow adoption often comes from an understandable desire to respond to client needs or move faster, but firms also need a credible path for users to raise needs, understand the risk, and see how decisions are made. That transparency makes governance feel like an enabler of progress rather than a barrier to it. 

Human Judgment Still Defines Trusted Legal Work 

Many current agentic workflows are deliberately instruction-driven rather than fully autonomous, which gives firms a way to explore more sophisticated ways of working without losing control of the process, the output, or the validation required before work is used. 

Technology does not eliminate the need for experienced professionals. Their expertise is what helps firms distinguish a plausible answer from an authoritative one, a fast workflow from a trusted workflow, and a useful experiment from a scalable capability. 

The Next Advantage Will Be Built, Not Bought 

Firms that gain the greatest advantage from AI will not simply deploy more tools. They will strengthen how work gets done—improving knowledge, connecting information, establishing governance, and building adoption around real practice needs. 

To do that, firms need legal AI built on trusted legal intelligence and designed for the standards of accuracy, security, and accountability legal work demands. With Lexis+® with Protégé™, LexisNexis helps legal teams move from experimentation to meaningful impact with greater confidence. 

Learn more about Lexis+® with Protégé™.