Inside the battle in legal tech to 'OpenAI-proof' its business (www.businessinsider.com)

A wave of big legal‑tech funding this week — roughly $320 million so far — has gone to startups explicitly trying to be “OpenAI‑proof.” Crosby (which announced $20M) reviews contracts as a service using LLMs plus human triage to deliver sub‑hour turnarounds; Spellbook raised $50M to focus on contract drafting and a Word plug‑in tuned on some 10 million contracts and firm house styles; EvenUp ($150M Series E) and Eve ($103M Series B) are backing niche plaintiff/litigation workflows. The rush follows disclosures that OpenAI uses internal helpers like a “contract data agent” and fueled market jitters (e.g., a drop in DocuSign stock) about infra providers moving from powering SaaS to competing in verticals. For the AI/ML community this highlights the strategic value of vertical specialization, human‑in‑the‑loop pipelines, curated domain corpora, and workflow integration (e.g., Word plugins, fine‑tuned triage models) as durable moats against generalist LLMs. It also foregrounds a technical and business tradeoff: build deep, narrowly tuned models and product workflows that encode legal expertise, or risk being commoditized by platform providers like OpenAI/Microsoft if they choose to push Copilot/DocuGPT deeper into legal workflows. Investors are betting that domain depth, data, regulatory know‑how and service models will keep vertical startups viable — but the threat of infrastructure players entering remains the key uncertainty.