Survey Shows 83% of Legal Teams Use AI, but Trust in Outputs Remains Low

Why It Matters

The survey’s stark contrast between near‑universal AI access and low trust levels signals a pivotal inflection point for LegalTech. Without confidence in output quality, firms cannot justify the investment needed to scale AI beyond pilot projects, limiting the sector’s ability to achieve the efficiency gains promised by generative AI. For in‑house teams, higher trust could translate into faster contract turnaround and reduced reliance on external counsel, reshaping procurement and risk management strategies. For law firms, the trust gap reinforces the tension between the billable‑hour model and technology‑driven efficiency. If firms cannot demonstrate ROI, they risk falling behind competitors that adopt AI‑enhanced workflows. The data therefore pressures both vendors and legal departments to prioritize accuracy, transparent validation, and measurable performance metrics to unlock the next wave of AI‑driven legal services.

Key Takeaways

• 83% of surveyed in‑house and law‑firm leaders now have broad AI access, up from 61% in 2025.
• Only 22.1% report high trust in AI outputs, yet high‑trust teams are 3x more likely to see positive ROI.
• 69.7% of AI‑generated outputs still need targeted edits or extensive rework.
• 54% of respondents use legal AI tools often, but billable‑hour constraints limit clear ROI for law firms.
• Contract review and summarisation are the leading use cases for legal AI today.

Pulse Analysis

The ALSP Factor survey underscores a classic technology adoption curve: access precedes trust, and trust precedes scale. The rapid jump to 83% access suggests that procurement and IT teams have cleared the initial barrier of tool selection and integration. What remains is the hard work of embedding AI into the substantive legal decision‑making process, where errors can have material risk. The 22.1% high‑trust figure is a wake‑up call for vendors; they must move beyond generic large‑language models to domain‑specific, fine‑tuned solutions that ingest proprietary case law, contracts, and regulatory data. Such vertical specialization can shrink the 69.7% rework rate, turning AI from a drafting assistant into a reliable co‑counsel.

Law firms face a structural dilemma. The billable‑hour model rewards time spent, not efficiency, creating a perverse incentive to under‑utilize tools that could cut billable hours. Some firms are experimenting with blended pricing or value‑based fees to align financial outcomes with AI‑driven productivity, but industry-wide adoption will likely require a broader shift in billing philosophy. In‑house legal departments, unshackled from hourly billing, are better positioned to showcase ROI, making them early adopters and reference customers for AI vendors.

Looking ahead, the upcoming Legal Innovators Europe conference could serve as a catalyst for the next phase: moving from access to defensible impact. If vendors can demonstrate measurable accuracy improvements—perhaps by publishing benchmark studies that show a reduction in the rework rate from 70% to under 40%—the trust gap could narrow quickly. Until then, the sector will continue to wrestle with the paradox of ubiquitous AI tools that remain under‑utilized due to lingering doubts about their reliability.