Govern by use case
Apply the NIST Govern, Map, Measure and Manage functions proportionately to the impact of each workflow.
AI uses for
Explore where AI can accelerate legal work, connect firm operations and support complex analysis—while professional judgment and responsibility remain human.
Small/medium general law firmPossible AI utilization
Showing
Mature, accessible assistance and bounded automation available with today’s mainstream tools.
Gather contact details, matter facts, documents, urgency, and other predefined intake fields.
A bounded assistant can gather structured facts while avoiding advice and offering human escalation.
Apply a controlled taxonomy, extract metadata, and identify duplicates or filing exceptions.
Automation is practical with access controls, audit logs, and an exception queue.
Retrieve relevant precedents, templates, checklists, and prior work from an access-controlled corpus.
Retrieval can be automated, while suitability and currency require review.
Tag, summarize, and route approved material into the firm's knowledge taxonomy.
Human spot checks and approval-state controls remain necessary.
Draft narratives, classify entries, and flag inconsistencies for billing review.
Billing accuracy, reasonableness, and client rules require review.
Propose times, resolve routine conflicts, prepare invitations, and update calendars.
Routine scheduling is highly automatable with permissions and confirmation rules.
Deployment context · United States
The technical use cases remain available, but deployment should be governed by sector rules, state law and a documented risk process.
This context does not currently change the Possible AI Utilization analysis or any experimental score.
Apply the NIST Govern, Map, Measure and Manage functions proportionately to the impact of each workflow.
Professional, health, employment, consumer and privacy duties still apply when AI assists the work.
Evaluate accuracy, harmful bias, security, privacy and human fallback using representative local inputs.
Observed utilization
Direct adoption observations are available, but they measure different constructs—tool use, use frequency, and tool type. They cannot yet be normalized into a task-coverage score comparable with Possible.
01
Direct law-firm evidence
United States and nine European countries
02
Direct law-firm evidence
United States and nine European countries
03
Broader legal-profession evidence
Legal professionals surveyed for the 2025 Legal Trends Report; item geography not separately published
04
Broader legal-profession evidence
Legal professionals surveyed for the 2025 Legal Trends Report; item geography not separately published
Observed utilization · United States evidence · US
Broader official sector proxy · not direct industry adoption
These are country-level broad-sector observations. Industry links are transparent mappings, not claims about every business in that industry.
Businesses reporting AI use in any business function during the previous two weeks
Period2026-07-13 to 2026-07-26
Industry mappingBroader than the selected industry: this sector also contains other professional, scientific and technical services.
EvidenceGrade A at the source-sector level · High confidence. The mapped AIBI industry is a broader-sector proxy.
Utilization gap
The opportunity analysis is task-level, while observed reports measure tool use, frequency, or broad-sector adoption. Those constructs cannot be subtracted honestly, so no numeric gap is shown.
Not comparable yet
About the experimental index
It combines task importance with current AI capability, reliability, integration, risk and human-oversight constraints for a representative law firm. It does not mean 52.3% of legal work can be automated, it is not country-specific, and it is not used as the observed-utilization measure. Read the methodology →
Explore every function
Scores describe bounded practical use, not permission to automate professional responsibility.
Research assistance
Capability v2026.08 · effective 2026-08-23
Semantic/document search
Capability v2026.08 · effective 2026-08-23
45/100
Augment
Risk constraint (45)
Medium
Research assistance
Capability v2026.08 · effective 2026-08-23
Summarization
Capability v2026.08 · effective 2026-08-23
40/100
Assist
Human oversight constraint (40)
Medium
Information extraction
Capability v2026.08 · effective 2026-08-23
Classification
Capability v2026.08 · effective 2026-08-23
62/100
Partially automate
Human oversight constraint (62)
Medium
Document comparison
Capability v2026.08 · effective 2026-08-23
Summarization
Capability v2026.08 · effective 2026-08-23
62/100
Partially automate
Human oversight constraint (62)
Medium
Drafting
Capability v2026.08 · effective 2026-08-23
Information extraction
Capability v2026.08 · effective 2026-08-23
62/100
Partially automate
Human oversight constraint (62)
Medium
Drafting
Capability v2026.08 · effective 2026-08-23
Research assistance
Capability v2026.08 · effective 2026-08-23
40/100
Assist
Human oversight constraint (40)
Medium
Conversational/voice assistance
Capability v2026.08 · effective 2026-08-23
Information extraction
Capability v2026.08 · effective 2026-08-23
Scheduling/workflow automation
Capability v2026.08 · effective 2026-08-23
75/100
Partially automate
Human oversight requirements (65.1)
Medium
Information extraction
Capability v2026.08 · effective 2026-08-23
Semantic/document search
Capability v2026.08 · effective 2026-08-23
Scheduling/workflow automation
Capability v2026.08 · effective 2026-08-23
45/100
Partially automate
Risk constraint (45)
Medium
Email assistance
Capability v2026.08 · effective 2026-08-23
Summarization
Capability v2026.08 · effective 2026-08-23
62/100
Assist
Human oversight constraint (62)
Medium
Drafting
Capability v2026.08 · effective 2026-08-23
Summarization
Capability v2026.08 · effective 2026-08-23
40/100
Assist
Human oversight constraint (40)
Medium
Classification
Capability v2026.08 · effective 2026-08-23
Information extraction
Capability v2026.08 · effective 2026-08-23
Scheduling/workflow automation
Capability v2026.08 · effective 2026-08-23
78/100
Mostly automate
Risk constraint (78)
Medium
Information extraction
Capability v2026.08 · effective 2026-08-23
Scheduling/workflow automation
Capability v2026.08 · effective 2026-08-23
Agentic workflow execution
Capability v2026.08 · effective 2026-08-23
45/100
Partially automate
Risk constraint (45)
Low
Semantic/document search
Capability v2026.08 · effective 2026-08-23
Summarization
Capability v2026.08 · effective 2026-08-23
78/100
Partially automate
Risk constraint (78)
Medium
Classification
Capability v2026.08 · effective 2026-08-23
Summarization
Capability v2026.08 · effective 2026-08-23
Scheduling/workflow automation
Capability v2026.08 · effective 2026-08-23
82.5/100
Mostly automate
Professional and risk constraints (76.8)
Medium
Drafting
Capability v2026.08 · effective 2026-08-23
Classification
Capability v2026.08 · effective 2026-08-23
Data analysis
Capability v2026.08 · effective 2026-08-23
78/100
Partially automate
Risk constraint (78)
Medium
Data analysis
Capability v2026.08 · effective 2026-08-23
Summarization
Capability v2026.08 · effective 2026-08-23
74/100
Augment
Recommended AI role (74)
Medium
Scheduling/workflow automation
Capability v2026.08 · effective 2026-08-23
Conversational/voice assistance
Capability v2026.08 · effective 2026-08-23
73.2/100
Mostly automate
Integration complexity (55.5)
Medium
Scheduling/workflow automation
Capability v2026.08 · effective 2026-08-23
Information extraction
Capability v2026.08 · effective 2026-08-23
45/100
Partially automate
Risk constraint (45)
Medium
Drafting
Capability v2026.08 · effective 2026-08-23
Summarization
Capability v2026.08 · effective 2026-08-23
68/100
Assist
Recommended AI role (68)
Medium
Data analysis
Capability v2026.08 · effective 2026-08-23
Classification
Capability v2026.08 · effective 2026-08-23
74/100
Augment
Recommended AI role (74)
Medium
Research assistance
Capability v2026.08 · effective 2026-08-23
Data analysis
Capability v2026.08 · effective 2026-08-23
Summarization
Capability v2026.08 · effective 2026-08-23
40/100
Assist
Human oversight constraint (40)
Medium
No capability is mapped because this work is not considered practically appropriate for AI execution.
0/100
Not practically appropriate
No applicable AI capability
Unknown
Summarization
Capability v2026.08 · effective 2026-08-23
Data analysis
Capability v2026.08 · effective 2026-08-23
Drafting
Capability v2026.08 · effective 2026-08-23
40/100
Assist
Human oversight constraint (40)
Medium
No capability is mapped because this work is not considered practically appropriate for AI execution.
0/100
Not practically appropriate
No applicable AI capability
Unknown