When large language models entered the legal field, many expected a flood of successful LegalTech products. Yet most have not taken off. Why?
Bloomberg’s 2025 report offers a clue: many lawyers prefer general-purpose tools like ChatGPT or Copilot over niche products. Flexibility matters more than narrow workflows. Traditional “boxed” products feel rigid; LLMs feel adaptable.
This reflects an outdated mindset. Old automation required single-purpose products. LLMs, however, are engines of reasoning. Forcing them into product boxes only disappoints. Lawyers revert to general-purpose models — not because they are perfect, but because they can be adapted on the fly.
Still, working directly with raw models is inefficient. The missing piece is integration. Like SAP in enterprise IT, the future lies in integrators: specialists who know law, AI, and IT, and can tailor workflows. Today, such integration can be done in hours, not months.
I have seen small firms achieve real value this way — automating reports, generating templates, applying review rule — none of which off-the-shelf products could deliver.
The failure of LegalTech startups is simple: they keep selling “products” while the market needs integration.
Sources
- Bloomberg Law Report: Artificial Intelligence – The Impact on the Legal Industry (2025)