For law firms to remain competitive, they must meet or exceed client expectations around service. So, what do corporations expect of law firms when it comes to using generative AI?

According to a 2024 survey of executives in corporate legal departments at Fortune 1000 companies, 68% of in-house counsel said they approve of their outside counsel using generative AI tools on their company’s legal work. Why? Because 80% of these executives anticipate that AI tools will reduce their outside counsel bills.

But whether generative AI will decrease, increase, or have no substantial effect on law firm profits remains to be seen.

The efficiency gains that generative AI produces may actually result in lower bills to corporate clients — but, at the same time, create new business opportunities. In fact, the survey above found that 70% of Am Law 200 law firm leaders believe AI solutions will enable new work product for clients.

Consistent with this optimism, 53% of Am Law 200 firms have purchased generative AI tools, and 45% are using them for legal work.

But getting the most out of this technology — without creating any ethical risks — depends upon data structuring and permissions changes that firms sometimes overlook.

Why structured data is the key to safe, effective use of generative AI at law firms

For generative AI tools to effectively read and understand your data, your data must be structured — that is, defined and organized in a specific way. Without this structure, it will be difficult to train your AI solution, and the results it generates will be missing information or not specific to your firm.

But there’s an additional reason that law firms in particular must have organized data. If your data isn’t identified in a way that’s clear to the AI, there’s a risk that the AI will surface information that is behind an ethical wall.

To avoid this outcome, your data must be tagged to indicate what client and matter it pertains to. Moreover, your firm must have permissions set across all applications.

Firms without structure and permissions in place have already seen problems arise during testing of AI solutions. For example, one firm informed Intapp about an issue they experienced while testing Microsoft Copilot, which for most law firms is the most natural generative AI choice. As the firm was piloting the solution, they discovered that it surfaced protected information because the firm hadn’t set their permissions correctly.

How to structure your firm’s data and ensure permissions are set correctly for AI

How can your firm most easily achieve structured data, set with appropriate permissions, so you can take advantage of generative AI?

Intapp Workspaces software is specifically designed to enable laws firms to organize and structure their data with identifying information that helps AI read it. In particular, Intapp Workspaces assigns client and matter codes to teams and SharePoint sites.

If you combine Intapp Workspaces with Intapp Walls, you’ll also ensure that Microsoft Copilot doesn’t pull from sources that a requestor doesn’t have permission to access. Microsoft Copilot communicates with Intapp Walls to check whether a user can access particular files and data points.

It’s not enough to have only one of these solutions. For Intapp Walls to correctly apply its policies to the right client and matter teams in Microsoft SharePoint and Teams, Intapp Walls needs to know which Microsoft data corresponds to which clients and matters. Intapp Workspaces solves this problem by facilitating consistent application of client-matter metadata in Microsoft 365.

Likewise, having Intapp Workspaces alone without Intapp Walls will enable structured data, but your permissions won’t be set correctly and automatically enforced across all your applications. As a result, Microsoft Copilot may surface information behind an ethical wall.

Prepare for AI by partnering with Intapp

Whether you’re an Intapp client or new to Intapp, schedule a demo to learn how we can help you assess your readiness for AI. We’ll evaluate your systems to determine whether your data is properly structured and whether you have an effective information governance system in place.