Legalweek 2026 Session Recap

Key Takeaway: Agentic AI is shifting how legal departments operate, moving beyond experimentation to real operational impact. During the Legalweek 2026 session “Strategic Legal Leadership in the Age of Agentic AI,” industry experts discussed why legal leaders must focus on high‑value use cases, lead cross‑functional adoption, and put clear governance in place. Panelists stressed that these strategies will balance innovation, accountability, and risk as AI agents become embedded in legal workflows.

With their ability to automate complex tasks, reason, and adapt to evolving legal requirements, AI agents represent the future of high-performing legal operations.  Legal departments seeking to integrate agentic AI into their operations must identify high impact use cases, manage organizational change, and determine appropriate safeguards.

This was the focus of the Legalweek 2026 legal operations workshop session presented by Epiq, “Strategic Legal Leadership in the Age of Agentic AI.” Jon Kessler, Vice President and General Manager, Information Governance at Epiq moderated a panel of four in-house leaders who are spearheading agentic AI adoption in their organizations:

• Jessica Escalera, Head of Legal Operations — Americas, HSBC
• Nicole Langston, Head of eDiscovery, Counsel, Barclays
• Lydia Petrakis, Assistant General Counsel, Microsoft
• John Zhu, Sr. Director, Legal and Compliance Tech, GSK

Driving Use and Adoption 

At the outset, the panelists spelled out the key difference between generative AI and agentic AI, with Nicole explaining how the latter can perceive context, plan steps, interact with multiple systems, make decisions about how to proceed, and work toward a goal. With that in mind, the panel emphasized three points regarding the adoption of agentic AI.

AI Adoption Extends Beyond Legal

The first is that adoption is not confined to legal departments. Other parts of the business are using AI and seemingly every day their teams are coming up with new ideas for the use of AI. Therefore, legal departments need to look beyond their own use of AI for legal workflows and instead collaborate with other business units on AI for business impact. This represents a sharp shift for legal, both in its role and process-wise. Rather than try to dictate how other departments use AI, legal departments need to take an agile and flexible partnership-based approach that balances legal risk with strategic business benefits.

Leadership as the Catalyst for Adoption

Second, leadership is paramount, especially since, as Lydia put it, “If you're not using AI, you’re falling behind.” AI adoption is not a one-off. Rather, it requires taking the entire organization up the AI learning and implementation curve. Executive leadership must be the catalyst for involving all the right stakeholders and taking a cross-functional approach. 

They should also leverage individuals within the organization who are already using AI. For example, have them lead internal training sessions, share practical use cases, and act as peer advisors to equip team members to harness AI in their day-to-day work.

Mindset Matters as Much as Technology

Third, a shift in mindset is necessary. Part of this is education in how agentic AI works, what it can and cannot do, the importance of process maps, and its pitfalls and risks. It also requires understanding that, with agentic AI, iteration is crucial. Finding the best use cases and which of those can scale involves trial and error, and leaders should encourage experimentation within appropriate guardrails.

Use Cases and Key Considerations

The panel highlighted the practical and legal considerations in identifying and pursuing agentic AI use cases. As a practical matter, not all agentic AI use cases are equal. The best approach is to start with a use case that has a large and compelling impact for a particular team and use that as a showcase and starting point. Often, the ideal candidate will involve streamlining a process or getting information beyond what is possible without AI. 

Governance, Accountability, and Human Oversight

Governance is crucial. Jon suggested a “lego bricks” approach whereby guardrails exist in every step and for each component of a workflow. Information governance goes beyond the accessibility of data to a particular agent; it also determines the question of what rights the agent has to share what types of data, and with whom. Delegation to agents creates novel risks. How much latitude should organizations give agents to make decisions, and when must human judgment intercede? Jessica noted that it is easy to become complacent with AI, creating dangers of overreliance on such tools. This makes it critical that a human will ultimately be accountable.

John stressed that, in all cases, it must be crystal clear what a particular agentic process and workflow do, who owns the process in question, and who is responsible for its output. Driving accountability also requires validations and human checkpoints, especially when agents interact with other agents. Parties must understand why they're getting the outputs they're getting and understand AI enough to know when to challenge those outputs. 

Agentic AI is increasingly able to replace traditional software applications and provide organizations with greater cost efficiency, improved decision-making, enhanced competitiveness, and more streamlined legal workflows. Legal department leaders and their teams will play an indispensable role in ensuring its appropriate and safe use.