Agentic AI

At ILTACON 2025, a recurring theme emerged across multiple sessions: the wave of AI advancement following generative is agentic. All told, its impact on the legal industry may be even more profound. While GenAI tools have been focused on producing text, extracting data, and answering prompts, agentic AI is designed to work more like a trusted colleague: setting goals, planning and executing multi-step workflows, adapting along the way, and checking in with humans when needed.

What Is Agentic AI?

Traditional AI tools in law have been built to perform discrete tasks, such as extracting clauses from a contract or summarizing a deposition. Agentic AI is goal-oriented. It uses a range of tools, data sources, and reasoning processes to figure out how to accomplish a broader objective, rather than just executing a single command.

Panelists from Aderant, Thomson Reuters, Litera, and Harvey emphasized that agency exists on a spectrum. In some use cases, you might “dial up” the AI’s autonomy to adapt workflows on the fly. In others, you might “dial it down” for predictable, tightly controlled steps. Either way, transparency is essential: lawyers and staff need to see what the agent is doing, review the sources it’s relying on, and verify that the process meets professional and ethical standards.

Human in the Loop, By Design

Agentic AI will not, and should not, operate in a vacuum. Building trust means designing verification and check-in points at the right moments in a workflow. Whether drafting a first-pass contract, researching case law, or flagging unusual contract changes, the agent should surface its reasoning and allow the human to adjust course.

The design challenge for legal technology providers is to make these review cycles seamless: embedding firm-specific voice and playbooks, citing authoritative sources, and showing the path the AI took to reach its conclusions. This human-in-the-loop approach balances efficiency with accuracy, a critical requirement for law firms.

Why Start in the Back Office?

While some law firm leaders are eager to explore agentic AI in legal workflows, others see back-office and operational functions as a safer proving ground. HR, IT, business development, billing, and compliance all present opportunities to deploy agents to automate multi-step, manual processes. This approach allows firms to gain comfort with the technology before applying it directly to client-facing legal work.

Lessons from Other Industries

If the software development sector is any indication, the legal industry is on the cusp of significant change. Agentic AI has already transformed coding by automating complex, multi-step tasks and amplifying the contributions of top talent. Legal technologists predict a similar amplification effect in law—freeing professionals from repetitive work so they can focus on the most complex, high-value aspects of their roles.

Preparing for the Next 12-24 Months

Speakers agreed on several near-term priorities for firms considering agentic AI:

• Get your data house in order. Agentic AI is only as effective as the context you provide. Structured experience data—matters handled, clauses used, workflows followed—will be foundational.
• Map your workflows. Agents can’t optimize a process that hasn’t been documented.
• Design for transparency and trust. Build in the right review points, permission controls, and source-citation practices.
• Stay flexible. Foundational models are evolving rapidly; design agents that can take advantage of the best available model for each task.

Beyond Hype

Is agentic AI just the latest tech buzzword? Industry leaders don’t think so. They expect it to change not just the speed of legal work, but the nature of the human role in producing it. As one panelist put it, “However big you think this will be in five years, it will be bigger.”

For law firms willing to experiment, capture the correct data, and build thoughtful workflows, agentic AI could represent the most significant leap forward since the first wave of generative AI. This technology is young, but the opportunity is here—and those who start now will be best positioned to lead the next wave.