Context, governance, and trust: the future of legal AI

Beyond context: why the future of legal AI will be defined by context, governance, and trust

At ILTACON 2026, one idea dominated conversations across the legal technology landscape: context has become the critical ingredient for effective AI. Across sessions, customer discussions, and vendor announcements, there was growing recognition that AI is only as valuable as the knowledge, relationships, and information it can access.

Simply put: AI is only as good as the data underneath it. Its quality, its structure, who is allowed to see it, and what it costs to use.

They're right, and context matters, but context alone is no longer enough.

iManage CEO Neil Araujo echoed the premise in the company's ILTACON update: "It's not the model. It's not the layer that goes above the model. It's your data, your knowledge, your experience. That's what makes you different."

As organizations move from experimentation to enterprise adoption, the conversation expands beyond how to make AI more effective. Legal leaders increasingly grapple with a different set of questions: How do organizations activate institutional knowledge for competitive advantage? How do they govern AI across multiple platforms, keep agents within guardrails, and maintain client trust while accelerating innovation?

This points to a larger reality.

• Context delivers stronger AI outcomes.

• Governance is what makes AI manageable and scalable.

• Trust determines whether AI can become transformational.

That is why the future of legal AI cannot be defined by context alone, but by the combined forces of context, governance, and trust.

The missing layer: institutional memory

Much of the context conversation has focused on helping AI find and understand information more effectively.

Important, absolutely. But not entirely sufficient.

The most valuable knowledge inside an organization is stored in experience: Why was a particular strategy pursued? How did a negotiation shift direction? Why did one argument succeed where another failed? And which lawyer successfully navigated a similar issue?

Institutional memory shapes legal judgment every day, yet much of it remains fragmented across individuals and practice groups, buried in emails, undocumented conversations, and accumulated experience.

As we embed AI in everyday legal work, we raise the stakes for preserving and operationalizing this knowledge. That’s why industry observers recognize capturing institutional memory as the next major objective beyond traditional document retrieval and search. The overarching goal is to ensure that both people and AI agents can benefit from the organization's collective intelligence.

But, in addition to being the last to be captured, institutional memory is the least likely to be governed. While documents have metadata, permissions, and retention rules, their reasoning rarely does. And governance is key to using AI safely and productively.

The organizations with access to the most models won’t be the ones that create lasting advantage from AI. It will be the ones that best capture, connect, preserve, and activate their institutional knowledge.

Because documents tell you what happened.

Institutional memory helps explain why.

And in legal work, that distinction matters.

The foundation exists, but is it ready?

Most legal organizations are not short of knowledge.

Decades of documents, emails, matters, and expertise already sit inside document and email management systems, classified, permissioned, and audited.

One of the recurring themes at ILTACON was how to bring that knowledge to life for AI without rebuilding it elsewhere and managing the redundancy.

That is as much a data architecture challenge as it is an AI challenge.

Search, metadata, relationships between people, matters, and documents, and the security model that governs them transform stored content into usable context. If those foundations are weak, no amount of model capability can compensate. If they are strong, every AI tool built on top of them becomes more effective.

There is also an economic reality emerging. Every AI platform that needs to draw on the same content must index it, understand it, and process it. Organizations deploying multiple AI platforms are discovering that the cost of understanding the same data can multiply quickly.

The real advantage will go to organizations that don't chase the newest model. It will go to those whose existing foundation is ready to be queried, trusted, governed, and built upon.

Context is only valuable if it can be trusted

As the context discussion heats up, we can begin to make an important distinction.

For some, context is employed to improve AI outputs. Better answers. Better retrieval. Greater efficiency.

Those outcomes matter. But for most organizations, context must serve a broader purpose. This is where the conversation needs to evolve.

The challenge facing most legal leaders today is more complex than generating better answers. It is ensuring that people and agents can access the right information, at the right time, with the right permissions, within an environment both lawyers and clients can trust. The challenge is generating better context for those answers.

Context is not just the information AI consumes. Context without governance introduces risk. Governance without context limits value.

Organizations need both. And both depend on the same thing: data that is correctly described, enriched, permissioned, and retrievable.

At iManage, this distinction is reflected in a framework that connects documents, metadata, activity, relationships, expertise, security controls, and institutional knowledge within a governed foundation for legal work. This framework is what we call iManage context fabric™ In it, rather than treating context solely as a source of information for AI systems, we aim to create an intelligence layer that helps organizations safely activate the knowledge they already possess.

Governance is the strategic control point

As organizations seek to operationalize AI, governance is rapidly moving from a compliance conversation to a business imperative. The idea that governance may become as important as AI capability itself was also communicated and reinforced at ILTACON.

Security incidents, client scrutiny, and the rise of increasingly autonomous AI systems have elevated governance to the executive agenda. Organizations no longer ask whether they should govern AI; they ask how they can do so without slowing innovation.

More specifically, they ask which AI agents should have access to client information and what actions they should be permitted to take. They also wonder how organizations maintain information barriers across multiple AI systems while monitoring activity and proving compliance.

Questions around what data exists, who and what may access it, under which policy, and with what audit trail are data governance questions. More importantly, they are questions about trust.

While many vendors focus on helping AI access information, far fewer focus on helping organizations govern how AI — particularly agentic AI — uses it. Governing agent activity across the enterprise will be pivotal to successful AI adoption.

iManage has historically concentrated its expertise, as well as a significant proportion of its overall investment, on governance, which has long been recognized as a primary strength of the iManage platform. Today’s capabilities already extend beyond information governance toward agent governance, and the company vision and roadmap go further still.

Through iManage Security Policy Manager, iManage Threat Manager, and the broader security, governance, risk & compliance portfolio, iManage helps organizations create the necessary controls to govern both human and AI activity. This includes enforcing policies around which agents can access content, what actions agents can take, how AI activity is monitored, and how AI-generated outcomes are audited.

AI doesn't create trust. Governance does.

The future isn't one AI. It's many.

And governance must evolve to accommodate another clear signal from ILTACON: that organizations are not adopting one AI, they are moving toward many. Copilot, Harvey, Legora, organization-developed agents, and other tools — even ones that haven't been built yet — are all in the mix.

As the industry embraces a multi-agent future of multiple AI systems coexisting across the organization, we face an entirely new challenge.

Connecting a single AI tool to organizational knowledge is one thing. Creating a trusted foundation that can support an evolving ecosystem of AI systems is something else entirely.

Organizations must manage governance, security, permissions, and auditability consistently across every AI tool they deploy. And as the number of agents grows, the challenge of enabling access shifts to governing outcomes.

Governance sits in the layer that holds the data, not inside any single AI tool, so it can apply the same rules to every request, wherever it originates. Because no single tool sees everything.

The conversation around context needs to evolve to meet the governance moment. The question is no longer whether an AI platform can access information. The question is what happens next?

Can the AI system identify the right expert? Does it understand relationships between people, matters, and documents? Can it take action within business workflows and operate within established governance and security controls?

As the strategic control point shifts and open standards are adopted, differentiation will rely less on connectivity and more on the richness of context and a governance framework that supports real workflows. Instead of residing in any individual AI application, the real long-term value may lie in the platform that synthesizes, governs, secures, and orchestrates context across all of them.

More than a destination for legal work, the iManage vision is deeper and broader. It offers a platform where professionals can work, collaborate, and engage with AI, built on a trusted foundation of context, governance, and security.

iManage enables organizations to choose the best AI experiences while maintaining a single source of governed context and institutional knowledge across the enterprise.

In a multi-agent world, AI vendors that answer the question fastest or best will come and go. For lasting ROI, look for the platform that understands the context, governs the outcome, and enables the next action.

The trust layer for the next-generation legal organization

Success depends on more than selecting the right models or deploying the latest agents.

Organizations need to preserve institutional knowledge, govern access across a growing AI ecosystem, maintain client trust, and build a foundation that evolves with technological change. Having spent the last several years on AI capability, the legal industry’s next phase is defining AI accountability.

This is the challenge facing every legal organization.

It is also the vision behind the next-generation iManage platform.

Providing context for AI, but also the governance and trust that enable organizations to operationalize AI with confidence.

Context delivers the outcomes. Governance makes AI scalable. Trust decides whether any of it becomes transformational.

In the end, competitive advantage won't come from having the most AI. It will come from having data that's governed, connected, and trustworthy enough to build on.

Join the conversation

Join us for the iManage Innovation Webinar and learn how leading organizations are preparing for the next generation of legal work. Register today.

For the full story behind the ILTACON announcements, read the iManage company update recap.