Over the past two years, LegalTech has seen an explosion of interest in generative AI. Startups raised funds, firms launched pilots, and consultants predicted efficiency gains across the industry. Yet, as both Bloomberg’s 2025 report and MIT’s State of AI in Business 2025 show, the reality has been far less inspiring. Most investments into GenAI solutions have failed to deliver measurable ROI. MIT’s data is striking: 95% of enterprise pilots stall before creating any business impact.
Why is this happening?
The Hype vs. the Numbers
AI hype runs far ahead of results. Reports themselves warn about “AI washing,” where companies overstate capabilities. But in the end, numbers do not lie. Each system has costs — development, adaptation, maintenance — and those costs are too often ignored.
Before deploying AI, a firm must weigh:
- Adaptation costs — restructuring workflows to fit the tool
- Maintenance costs — security, compliance, updates
- Payback horizon — when measurable value will actually appear
In practice, break-even points often stretch five years out, by which time the tool may already be obsolete.
Few firms can make these calculations honestly, and vendors have incentives to oversell. Sometimes the answer is simple: this does not make economic sense. Saying this openly preserves long-term trust.
AI in LegalTech is not doomed. But moving forward requires discipline: fewer bets on hype, more reliance on numbers. Calculate, assess, and only then decide. Otherwise, what looks like innovation today may become sunk cost tomorrow.
Sources
- Bloomberg Law Report: Artificial Intelligence – The Impact on the Legal Industry (2025)
- MIT Project NANDA: State of AI in Business 2025