Navigating the Generative AI Era
LegalWeek was strong this year. The move to the Javits Center was a welcome upgrade—easy to navigate, plenty of informal seating for conversations, and a friendly staff. More importantly, the education sessions I attended were substantive and featured genuine thought leaders in the space.
Over the next few posts, I’ll share several takeaways from the conference—all focused on GenAI. But before diving into specific tools and workshops, it’s worth stepping back and looking at how GenAI is changing a core principle of legal operations. The legal ops playbook is changing.
For years, legal ops followed a simple rule:
Get the process right first. Then select the tool.
That logic still applies when evaluating traditional point solutions. But GenAI platforms are starting to collapsing that sequence.
Today, process design and tool development can evolve together.
You should still map the process—but the timeline has collapsed. A workflow can move from concept to prototype in days. Legal teams can iterate in real time instead of waiting for the “perfect” process map before building.
For legal departments with enterprise access to general-purpose GenAI tools—such as Anthropic’s Claude, OpenAI’s ChatGPT, or Google’s Gemini—the best starting point is often surprisingly simple:
Build in the general tools first.
Prototype.
Identify gaps.
Then add legal-specific tools to address gaps.
This also refocuses the platform vs. point solution debate.
Point solutions will still have an important role—especially in litigation, where specialized workflows remain critical. But for most mid-to-large organizations, a flexible platform approach that supports multiple use cases will often prove more scalable and cost-effective.
The long-standing skepticism in legal operations about relying on a single enterprise legal platform still applies. The future will likely remain a hub-and-spoke model: a central platform integrated with data sources and specialized tools. The difference is that the hub itself will increasingly be GenAI-first.
Choosing among legal AI platforms is therefore less about identifying a single “best” product and more about understanding how a provider fits your organization’s profile. Key factors typically include:
- Initial license count and cost
- The provider’s model training approach (fixed vs. adaptive)
- Integration with your primary legal research platform
- The provider’s product roadmap and development philosophy
A client portal feature may also influence platform selection. Many legal departments—if not law firms—will want to build client-facing GenAI applications. That capability is becoming standard and will likely be broadly available across major legal platforms in the near future.
At Hearst, our approach was shaped by a few specific factors.
The Office of General Counsel has extensive experience working with HearstLab portfolio companies, so our legal ops team values platforms where we can actively influence the roadmap. Other organizations may prefer vendors whose development priorities are shaped more broadly by the market.
We also operate with a small but highly capable operations team, and we’ve invested significant effort in educating our lawyers and paralegals on AI. That meant we wanted a platform with a strong library of legal workflows—allowing citizen builders to prototype solutions quickly while our dev-ops function helped refine and scale them. We also needed a provider that would adapt training to where the legal team is, rather than retracing introductory AI concepts.
Existing technology investments mattered as well. Because Hearst is a committed user of Thomson Reuters’ Westlaw Advantage and Deep Research, integrations with other research platforms such as vLex or LexisNexis were less compelling differentiators.
Finally, the ability to fully audit AI outputs was essential for our legal use cases.
For those reasons, Hearst prioritized a pilot with Legora, which I’ll discuss in more detail in a future post.
In the next installment, I’ll focus on what may be the most important GenAI challenge for legal organizations: training lawyers and allied legal professionals for the AI era.
Then I’ll also cover two excellent workshops led by Epiq on:
• Total Cost of Ownership for GenAI
• Strategic leadership of AI initiatives
And I’ll close the series by revisiting my 2021 post “Is the Golden Age of Legal Ops Dead?” in light of the discussion at the LexisNexis GC Workshop on what GenAI disruption may mean for both law departments and law firms.
One final note: I’m now available to consult in this space.
I originally published Ops in a Box: Legal Edition – A Magical Kit because my full-time role prevented me from consulting. That’s no longer the case.
If I can help accelerate your organization’s GenAI journey—whether by narrowing the 800+ GenAI legal products to a practical shortlist, or by organizing an agent-builder day to jump start experimentation—I’d be happy to connect.
