Enterprise ChatGPT outperforms legal AI
Healthcare company associate GC on where legal AI products break down
The core issue is that general legal AI has not yet created enough workflow advantage to beat a secure general model that the company already pays for. In this interview, the associate GC describes Harvey, GCAI, and similar tools as mostly another chat layer with some prebuilt workflows, not a system that materially changes contract review, compliance analysis, or research. For a lean in house team, the missing value is not raw text generation, it is reliable use of internal precedent inside everyday legal operations.
-
The strongest unmet use case is concrete and narrow. Ingest the company’s own agreements, learn its clause positions, then flag when a third party draft departs from that playbook. The interview says current tools, including CLM products like Luminance, promise this but still require too much setup and too much checking to trust in live work.
-
Large firms are buying Harvey, but often for reasons that do not help smaller in house teams. A separate law firm interview says Harvey adoption is driven by client demand, brand recognition, and targeted practice group licenses, not broad proof that it is clearly better than internal AI plus existing systems. That makes the product look bigger in the market than its day to day edge may be.
-
Where Harvey is more tangible is transactional work, where lawyers repeatedly draft, review, summarize, and extract terms from documents. But even there, the edge comes from fitting into Word and document systems, or from access to internal knowledge, not from better base model output alone. Research remains harder because Harvey does not own the underlying legal corpus the way Westlaw or Thomson Reuters tools do.
This category is heading toward two clear lanes. One lane is cheaper enterprise AI that already meets baseline drafting and summarization needs. The other is legal software that is deeply wired into contracts, precedent, research, and approvals. The winners will be the tools that disappear into daily workflow and reliably catch issues against a company’s own documents, not the ones that simply wrap a frontier model.