AI Implementation Roadmaps
Why Legal AI Adoption Stalls After Purchase and How to Restart It
By Anthony Villanueva ·
Many law firms have licensed AI tools that are not being meaningfully used. The adoption stall is a specific, diagnosable problem with a specific set of interventions. This guide covers the diagnosis and the restart plan.
Bloomberg Law's analysis of legal AI investment patterns in 2024 found that fewer than 10 percent of licensed AI tools in law firms had reached meaningful workflow integration. The other 90 percent were licensed, launched with some fanfare, and gradually relegated to the category of tools that attorneys know exist but rarely use. The adoption stall is not a technology failure — the tools are capable. It is an implementation failure, and it is recoverable.
Diagnostic using usage data
The diagnostic for an adoption stall begins with usage data. Most enterprise AI platforms provide usage analytics; if the firm does not have this data, requesting it from the vendor is the first step. Usage data reveals who is using the tool, how often, for what types of tasks, and where they tend to stop using it.
A platform with 90 percent seat adoption and 10 percent active usage has a different problem than one with 20 percent seat adoption — the former is a retention problem; the latter is an activation problem. The interventions are different.
Integration and activation failures
The most common cause of activation failure is that the tool was not integrated into the attorney's existing workflow. An attorney who has to navigate to a separate application, log in to a separate system, and reformulate their question in the tool's input format has too many friction points to sustain regular use.
Integration engineering — connecting the AI tool to the document management system, the research platform, or the practice management tool the attorney already uses daily — is the intervention that most consistently restarts adoption.
Causes of retention failure
Retention failure — attorneys who tried the tool and stopped — is typically caused by one of two things: the tool performed poorly on the use cases the attorney applied it to, or the attorney did not receive sufficient training to use the tool effectively for their specific work. These require different interventions: performance failure requires engineering work to calibrate the tool for the relevant use case; training failure requires a structured re-onboarding program with use-case-specific guidance.
The 8am Legal Industry Report consistently identifies training as the highest-leverage adoption intervention for tools that have already been licensed.
Frequently asked questions
What does usage data reveal about the potential problems with enterprise AI platform adoption?
The diagnostic for an adoption stall begins with usage data. Most enterprise AI platforms provide usage analytics; if the firm does not have this data, requesting it from the vendor is the first step. Usage data reveals who is using the tool, how often, for what types of tasks, and where they tend to stop using it.
A platform with 90 percent seat adoption and 10 percent active usage has a different problem than one with 20 percent seat adoption — the former is a retention problem; the latter is an activation problem. The interventions are different.
What is the most common cause of activation failure for legal AI tools?
The most common cause of activation failure is that the tool was not integrated into the attorney's existing workflow. An attorney who has to navigate to a separate application, log in to a separate system, and reformulate their question in the tool's input format has too many friction points to sustain regular use.
Integration engineering — connecting the AI tool to the document management system, the research platform, or the practice management tool the attorney already uses daily — is the intervention that most consistently restarts adoption.
What are the two primary causes for the retention failure of legal AI tools?
Retention failure — attorneys who tried the tool and stopped — is typically caused by one of two things: the tool performed poorly on the use cases the attorney applied it to, or the attorney did not receive sufficient training to use the tool effectively for their specific work. These require different interventions: performance failure requires engineering work to calibrate the tool for the relevant use case; training failure requires a structured re-onboarding program with use-case-specific guidance.
The 8am Legal Industry Report consistently identifies training as the highest-leverage adoption intervention for tools that have already been licensed.