Is Just Claude Sufficient for Legal Teams?

Spent some time yesterday afternoon with a dozen legal peers talking about AI implementations in legal. Not AI in the abstract, but the actual work of rolling it out inside a legal team.

The conversation kept circling back to a common theme: should we invest effort in horizontal tools like Claude or in vertical tools built specifically for legal work, like Spellbook or Legora?

We identified the following barriers that impact all AI deployments:

• Tool access restrictions from IT & compliance departments

• Lack of time for the legal team to explore and test capabilities

• Uncertainty about viable use cases (does the effort to create a GPT/Gem/Skill outweigh the effort just do the work in a traditional workflow)

• Difficulty creating repeatable workflows

• Complex sharing mechanisms (many times requires API/OAuth-type approaches that exceed the legal team's skill sets)

Horizontal tools carry more of these burdens because vertical legal AI vendors knock many of them down by design: pre-built use cases, managed sharing, and integrations that don't require API fluency to deploy. The billions in PE and VC capital flowing into vertical legal AI right now suggests the market agrees.

That said, vertical tools don't replace the horizontal ones. The teams getting the most from AI are running both. Well…today, anyway. The future "winner" is far from clear.

It's a fun time to be an AI attorney!