Paralegals and Legal Assistants
23-2011.00Median wage $62,890/yr392,880 employed (US)Rank #185 of 923 scored · top 20% by substitution
Assist lawyers by investigating facts, preparing legal documents, or researching legal precedent. Conduct research to support a legal proceeding, to formulate a defense, or to initiate legal action.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution40
Exposure38
Augmentation68
Substitution — the headline: capability discounted by cost, barriers and adoption.
Exposure — technical capability alone, regardless of whether anyone deploys it.
Augmentation — how much AI assists without replacing. High here + moderate substitution = a changing job, not a disappearing one.
Tasks on the substitution scale
12 rated tasks, binned by substitution score.
Position among all scored occupations
Distribution of 923 occupation scores; the marker is this occupation.
Tasks with substitution ≥ 70
8%
Run 1.0.0-draft.1 · computed 2026-08-05 · rater panel: claude-sonnet-5, claude-haiku-4-5-20251001 · intervals span rater disagreement.
Why this score
The five weighted dimensions of the composite, averaged across this occupation's tasks (importance-weighted, panel mean). Exact weights and formulas: /api/v1/methodology.
Task automatabilityw 35%39
panel mean rating 2.6/5 → substitution pressure 39/100
Technical feasibility todayw 20%37
panel mean rating 2.5/5 → substitution pressure 37/100
Cost vs. human wagew 15%41
panel mean rating 2.6/5 → substitution pressure 41/100
Adoption barriersw 20%inverted — strong barriers lower the score41
panel mean rating 3.3/5 (barrier strength) → substitution pressure 41/100
Sector adoption velocityw 10%41
panel mean rating 2.6/5 → substitution pressure 41/100
Task breakdown (12 tasks)
Substitution pressure per task, weighted by O*NET importance in the composite. Expand a task for the full rater audit trail — every rating, every model, every rationale.
Gather and analyze research data, such as statutes, decisions, and legal articles, codes, and documents.
77CI 75–79 · exposure 75 · augmentation 100 · importance 3.9/5 · click for rater detail
Gather and analyze research data, such as statutes, decisions, and legal articles, codes, and documents.
77| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Legal services and law firms are highly digitized sectors with rapid AI adoption; major legal tech vendors, BigLaw, and corporate legal departments are actively deploying AI research assistants in production. Measured displacement and pilot-to-production pipelines indicate strong, fast adoption momentum in the profession. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Legal services, a professional/information sector, has seen rapid uptake of AI research tools over the past two years, with major firms and vendors integrating them into standard workflows. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | AI dramatically amplifies paralegal productivity by instantly surfacing relevant precedents, statutes, and summaries, allowing humans to focus on analysis and strategy rather than manual research. This is a canonical example of AI augmentation—the human attorney or paralegal remains in control and makes final judgments, but the AI partner greatly accelerates the research phase. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI dramatically accelerates gathering and synthesizing statutes, cases, and secondary sources, letting paralegals produce first-pass research far faster while still validating and refining outputs. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current AI systems (LLMs, semantic search, document analysis tools) can automatically gather, summarize, and organize statutes, legal decisions, and documents with high speed and quality. Paralegals spend significant time on these research and synthesis tasks, which are predominantly text-based and rule-driven—AI can achieve well over 50% time savings with minimal human oversight on routine research. |
| Task automatability | claude-sonnet-5 | 4/5 | Legal research and analysis over statutes, case law, and secondary sources is largely text-based retrieval and synthesis, a task current LLM-based legal research tools handle with substantial time savings, though complex jurisdictional nuance and final validation still require human review.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | No licensing requirement mandates human paralegals perform research; firms and lawyers retain discretion to use AI tools. However, attorney oversight and professional responsibility create mild friction (lawyers must verify results), and some client contracts may prefer human paralegal involvement, but these are not hard blockers. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement bars paralegals from using AI tools, but attorneys bear supervisory and ethical responsibility for research accuracy, creating moderate liability-driven caution around unverified AI outputs. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI inference and document processing cost is orders of magnitude cheaper than paralegals' loaded hourly rates ($50–$80+/hour). A month of AI-assisted research costs less than a few billable hours of paralegal time, making the economic case strongly favorable. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | AI legal research subscriptions cost a small fraction of billable paralegal hours for equivalent research volume, though oversight and verification time offsets some savings. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Mature products (LexisNexis AI, Westlaw AI-Assisted Research, legal document analysis platforms) are deployed in law firms and legal departments today, performing statute lookup, case law retrieval, and document summarization in production. Error rates on straightforward retrieval and summarization are low, though complex legal interpretation still requires human validation. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Deployed products like Westlaw's CoCounsel, Lexis+ AI, and Harvey are used in production at law firms for statutory/case research with citation retrieval, though hallucination risks require verification workflows. |
Keep and monitor legal volumes to ensure that the law library is up-to-date.
61CI 23–100 · exposure 58 · augmentation 50 · importance 2.7/5 · click for rater detail
Keep and monitor legal volumes to ensure that the law library is up-to-date.
61| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Law libraries and legal departments have been slow to digitize and automate library management compared to other sectors; many still rely on manual volume maintenance. Adoption of AI for this specific task remains minimal, with most firms using traditional paralegals or library staff for collection upkeep. |
| Sector adoption velocity | claude-sonnet-5 | 5/5 | Legal services is a fast-adopting professional sector, and virtually all firms have already transitioned to digital legal research tools, phasing out physical library maintenance. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI systems could usefully assist by managing digital catalogs, identifying outdated publications, and flagging volumes needing replacement, allowing paralegals to prioritize physical inspection and ordering. This would enhance productivity in the organizational and planning aspects of library maintenance without removing human oversight. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Where paralegals still oversee physical or hybrid collections, digital alert systems and update services assist in tracking currency, though the task itself is largely obsolete. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Monitoring and updating physical law library volumes is primarily a manual, spatial task involving physical handling, verification of editions, and localized knowledge of organizational filing systems. While AI could assist with tracking metadata or flagging outdated publications via database analysis, the core task of physically inspecting, verifying, and organizing law volumes cannot be meaningfully automated to achieve 50% time savings end-to-end. |
| Task automatability | claude-sonnet-5 | 5/5 | Legal research databases (Westlaw, Lexis) already automate currency-checking and updating of legal materials, replacing physical volume maintenance almost entirely with automated updates and citators. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Law libraries operate within organizational workflows and may have custom cataloging standards, but there are no hard legal barriers preventing automation or AI assistance. Some friction exists around ensuring accuracy in legal citations and organizational preference for human oversight of collection integrity. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing or liability barrier prevents using digital tools for library maintenance; this is an administrative task with no legal requirement for human performance. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI could help with digital catalog maintenance and flagging outdated volumes via software, but the physical verification and shelving work must still be done by humans. The cost of any AI system for this narrow task would likely exceed the wages of a part-time library assistant performing manual checks. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Subscription-based digital legal databases cost far less than the paralegal hours needed to manually track and shelve print supplements and pocket parts. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product reliably performs the physical inspection, verification, and organization of law library collections at scale. The task requires embodied action (moving through shelves, handling books, comparing editions) that current AI systems cannot execute without robotics, which remains research-stage in law libraries. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Deployed legal research platforms have reliably automated update tracking, citator services (Shepard's/KeyCite), and version currency for decades in production use. |
Prepare, edit, or review legal documents, including legislation, briefs, pleadings, appeals, wills, contracts, and real estate closing statements.
57CI 49–65 · exposure 58 · augmentation 88 · importance 4.3/5 · click for rater detail
Prepare, edit, or review legal documents, including legislation, briefs, pleadings, appeals, wills, contracts, and real estate closing statements.
57| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Large law firms and corporate legal departments have begun pilot and limited production adoption of AI document automation; smaller firms and solo practitioners lag. Adoption is meaningful but constrained by regulatory conservatism, liability risk, and attorney gatekeeping. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Legal services is a fast-adopting professional-services sector with widespread pilot-to-production rollout of AI drafting and review tools at large and mid-size firms. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI drafting and review tools meaningfully accelerate paralegal productivity on document preparation by providing first drafts, flagging inconsistencies, and suggesting edits, with the paralegal and attorney retaining full control and judgment over the final output. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI dramatically speeds up drafting, editing, and first-pass review of legal documents while paralegals and attorneys retain final judgment and accountability. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can draft and edit routine legal documents (contracts, simple pleadings, closing statements) with moderate accuracy, reducing drafting time significantly. However, the task requires legal judgment, jurisdiction-specific compliance, and client-specific customization that current AI systems cannot reliably handle independently without expert review, limiting end-to-end automation to standardized documents. |
| Task automatability | claude-sonnet-5 | 4/5 | LLMs can draft and edit most standard legal documents (contracts, pleadings, closing statements) rapidly, saving substantial time, though final review and jurisdiction-specific nuance still require human oversight. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Legal documents often require attorney signature and malpractice liability attaches to document quality; bar regulations and ethical rules require licensed attorney oversight of substantive legal work, even when paralegals draft. These create hard barriers to full displacement. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement for paralegals themselves, but attorney sign-off, malpractice liability, and confidentiality rules create meaningful oversight friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI legal document tools cost a fraction of paralegal time per document generated or reviewed. At scale, inference plus integration overhead is substantially cheaper than paralegal labor ($25–50/hour), though final human review still absorbs some cost savings. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | AI drafting tools cost a small fraction of paralegal billable hours per document, though integration and mandatory human review add some cost back. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products like LexisNexis, Westlaw, and specialized legal AI tools exist and perform document drafting and review in production environments. However, material error rates persist (missed clauses, jurisdictional mismatches, client context failures), and deployment is narrowest on high-stakes matters like litigation or complex transactions. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Deployed legal-tech products (e.g., Harvey, CoCounsel, Clio Draft) reliably assist with drafting and review, but attorneys still supervise closely due to accuracy risks like hallucinated citations. |
Prepare affidavits or other documents, such as legal correspondence, and organize and maintain documents in paper or electronic filing system.
56CI 45–67 · exposure 62 · augmentation 88 · importance 4.4/5 · click for rater detail
Prepare affidavits or other documents, such as legal correspondence, and organize and maintain documents in paper or electronic filing system.
56| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Law firms are experimenting with document automation and e-filing systems, but adoption remains uneven. Large firms pilot AI-assisted drafting; smaller practices lag. Production displacement is emerging but not yet widespread. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Legal services is a professional services sector showing growing but still cautious AI adoption, with many firms piloting document automation and drafting tools rather than fully deploying them at scale. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI excels at augmenting paralegal productivity: template generation, citation checking, document organization, and consistency flagging significantly accelerate human-supervised work while keeping the paralegal in quality control and decision-making roles. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI drafting tools and automated document management systems substantially speed up affidavit preparation and filing organization while a paralegal remains responsible for accuracy and final review. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can automate significant portions of document preparation (drafting affidavits, correspondence templates, organizing files) but requires human review for accuracy, legal sufficiency, and case-specific tailoring. Estimated time savings approach 50% with setup, though quality verification remains necessary. |
| Task automatability | claude-sonnet-5 | 4/5 | Drafting affidavits and legal correspondence from templates and case facts, plus organizing files, are largely pattern-based text generation and document management tasks that LLMs and document automation tools handle well, though final review is still needed. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Affidavits and legal filings often require attorney signature and certification, creating legal/liability barriers. Court rules, jurisdictional variations, and malpractice exposure mean human review remains mandatory, limiting substitution. |
| Adoption barriers | claude-sonnet-5 | 3/5 | Affidavits often require notarization and attorney certification of accuracy, and paralegals typically work under attorney supervision, creating moderate liability and procedural barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | AI-driven document automation (LLM APIs, document assembly services) plus integration costs are approaching paralegal wage rates, but not yet substantially cheaper when oversight and error-correction labor are included. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | AI drafting and automated filing systems cost a small fraction of paralegal hourly billing rates for equivalent drafting and organizational throughput, though oversight and integration add some cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Document assembly and file management tools exist in production (contract automation, e-discovery platforms), but reliable end-to-end affidavit preparation and organizational compliance still require material human oversight. Error rates on legal specifics and filing requirements remain material. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Products like legal AI drafting assistants (e.g., Harvey, CoCounsel, Clio integrations) and generic document management/e-filing systems are deployed in law firms today for drafting and organizing documents, though accuracy on complex affidavits still requires attorney/paralegal review. |
Investigate facts and law of cases and search pertinent sources, such as public records and internet sources, to determine causes of action and to prepare cases.
44CI 39–50 · exposure 50 · augmentation 75 · importance 4.1/5 · click for rater detail
Investigate facts and law of cases and search pertinent sources, such as public records and internet sources, to determine causes of action and to prepare cases.
44| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Legal services remain traditional and risk-averse; while law firms have piloted AI research tools, production adoption of automated case investigation remains limited and defensive, concentrated in large firms. Digitization lags other professional services, and regulatory caution slows velocity. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Legal services are a professional/information sector with growing AI tool adoption (research assistants, e-discovery), but full production-scale integration into fact investigation workflows is still emerging rather than deep and fast. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly augments paralegal productivity on legal research, document collection, and fact organization by handling bulk searching, filtering results, and flagging relevant sources, allowing the paralegal to focus on synthesis and strategy. This is one of the stronger augmentation use cases in legal work today. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI substantially speeds up searching case law, statutes, and public records, letting paralegals focus on judgment-intensive synthesis and strategy while the human remains in the loop. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can automate significant portions of legal research, document retrieval, and fact-finding through legal databases and public records search, but the task requires judgment to determine relevance, synthesize findings into causes of action, and contextualize within specific client circumstances. This likely achieves 40-60% time savings on research components, but full end-to-end case preparation with legal strategy remains hybrid. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can accelerate legal research, fact pattern analysis, and public record searches, but synthesizing a coherent case theory and verifying accuracy still requires substantial human oversight, so it falls short of full automation at equal quality across the whole task. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Legal work faces material barriers: attorney liability and malpractice risk for errors, ethical rules requiring attorney supervision of legal work, bar regulations on unauthorized practice of law, and client expectations for human review. These create strong organizational and regulatory friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement mandates a human paralegal specifically, but attorneys retain ultimate responsibility for case strategy and factual accuracy under professional liability rules, creating moderate oversight friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI research tools cost hundreds to thousands monthly in subscriptions, plus integration and paralegal review time to validate outputs; junior paralegals performing routine legal research cost less per hour in many markets, though AI may reduce per-task cost moderately. Full cost advantage is not yet established. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI tools cut research time significantly but still require paralegal/attorney verification and subscription costs for specialized legal databases, making the net cost savings moderate rather than order-of-magnitude. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products like LexisNexis AI-Assisted Research, ROSS Intelligence, and general LLMs with legal training exist in production, but they operate within narrow scope (document review, research suggestions) and require human oversight to catch errors, misapplied precedent, and gaps. Reliably preparing an entire case remains outside deployed AI capability. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Legal research products (e.g., Westlaw Edge, Lexis+ AI, CoCounsel) are deployed in production, but hallucination risk and narrow reliability on fact investigation (as opposed to pure legal research) mean they don't yet perform the full task reliably without close review. |
File pleadings with court clerks.
39CI 28–50 · exposure 38 · augmentation 75 · importance 3.9/5 · click for rater detail
File pleadings with court clerks.
39| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Law firms have adopted e-filing platforms and document automation tools, but adoption of autonomous AI filing agents remains limited and cautious. Pilots exist, but production deployment of fully automated filing without attorney oversight is rare due to liability concerns. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Legal services broadly are mid-tier adopters of AI and e-filing automation; large firms use automated filing tools, but widespread AI-driven autonomous filing across all courts and jurisdictions remains inconsistent. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly assists by auto-populating fields, checking compliance rules, formatting documents, and flagging jurisdiction-specific requirements. Paralegals can review and submit AI-prepared filings far faster than creating them manually, creating substantial productivity gains while maintaining human control. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools can pre-check formatting, deadlines, and required fields, and populate e-filing forms, substantially speeding paralegal workflow while a human still confirms and submits. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Filing pleadings requires understanding court-specific requirements, formatting compliance, and procedural rules that vary by jurisdiction. While AI can generate compliant documents, the verification step, court portal navigation, and handling of rejections or special filing requirements still require significant human oversight, limiting time savings below the 50% threshold. |
| Task automatability | claude-sonnet-5 | 3/5 | E-filing systems and AI-assisted document preparation can automate much of the mechanical filing process, but final submission, verifying court-specific formatting rules, and handling clerk rejections still require human oversight.rr |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Court systems have strict rules about who may file (often requiring attorney authorization), procedural requirements that vary by jurisdiction, and liability for incorrect filings rests on the law firm. These regulatory and accountability structures create meaningful friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement mandates a human file the paperwork, but court rules, jurisdiction-specific formatting, and liability for missed deadlines create meaningful procedural friction that discourages full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current AI-assisted e-filing still requires significant paralegal review and manual court portal interaction. The setup costs, licensing for court access, and ongoing oversight mean the total cost per filing remains comparable to or slightly above traditional human-assisted filing for complex cases. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | E-filing portals reduce clerical cost significantly, but paralegal review of compliance with local rules and correction of rejected filings still requires paid human time, keeping costs roughly comparable to partially automated workflows. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some legal tech platforms offer e-filing integration, but they typically require human paralegals to verify documents, select correct filing categories, and handle court-specific workflows. No mature product autonomously handles the full filing task across diverse jurisdictions without material human intervention and error correction. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Electronic court filing (ECF/CM-ECF) systems are mature and widely deployed, but AI-driven autonomous filing (auto-formatting, deadline verification, error correction) is still narrow and firm-specific rather than universal. |
Prepare for trial by performing tasks such as organizing exhibits.
31CI 30–32 · exposure 25 · augmentation 63 · importance 4.0/5 · click for rater detail
Prepare for trial by performing tasks such as organizing exhibits.
31| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Law firms are adopting e-discovery and document management tools, but adoption of AI-driven exhibit organization specifically is still in pilot phases; most firms rely on paralegal-led preparation because trial strategy integration remains heavily human-driven. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Legal services are adopting AI at a middling pace—e-discovery and document review tools are common, but exhibit and trial prep workflows remain less digitized and slower to adopt full automation. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can meaningfully assist by suggesting categorization schemes, flagging missing exhibits, and generating indices or cross-references, raising paralegal productivity on the organizational mechanics; however, the paralegal or attorney must remain in the loop for strategic decisions about presentation. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools significantly help with searching, tagging, summarizing, and cross-referencing documents for exhibits, meaningfully speeding up the paralegal's preparation work even though the human remains essential. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Organizing exhibits involves physical arrangement, sorting, and logical structuring that current AI can partially support (e.g., indexing, tagging, document management), but the full end-to-end task—especially the judgment-driven organization optimized for trial strategy—requires human oversight and cannot achieve 50% time savings reliably without significant domain-specific setup. |
| Task automatability | claude-sonnet-5 | 2/5 | Physical and logistical organization of exhibits, binders, and courtroom materials involves handling, tabbing, formatting per court rules, and coordination that current AI cannot fully execute end-to-end.the digital sub-parts (indexing, labeling) can be assisted but not autonomously completed. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Attorneys typically review and sign off on exhibit organization before trial, and some jurisdictions have rules about chain of custody and exhibit handling; this creates moderate friction, though not a hard legal barrier to automation of the underlying organizational task. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement specifically for exhibit prep, but attorney work-product oversight, court procedural rules, and error-cost sensitivity in trial settings create moderate friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | While AI tools (e.g., document assembly, tagging services) have become cheaper, the overhead of integration, oversight, and inevitable human correction for trial-critical exhibit organization means all-in costs remain comparable to or exceed typical paralegal wages for this work. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI tools can reduce time on document tagging/organization but human paralegals still must review, physically prepare materials, and ensure compliance with court-specific rules, keeping costs comparable rather than dramatically lower. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI document management and e-discovery tools exist and can automate some indexing and categorization, but exhibit organization for trial typically requires human judgment about presentation strategy, physical logistics, and attorney feedback; no mature product reliably performs the complete task end-to-end in production at scale. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some litigation support software assists with exhibit tagging and Bates numbering, but no deployed product reliably manages full trial exhibit preparation without substantial human oversight. |
Arbitrate disputes between parties and assist in the real estate closing process, such as by reviewing title searches.
30CI 25–35 · exposure 30 · augmentation 75 · importance 3.0/5 · click for rater detail
Arbitrate disputes between parties and assist in the real estate closing process, such as by reviewing title searches.
30| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Legal services sectors show moderate AI adoption for document review and discovery, but arbitration and closing tasks remain highly human-centric with slow automation. Most law firms use AI assistively rather than as replacement. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Legal services adopt AI unevenly; document review tools are gaining traction but dispute arbitration and closing processes remain largely manual with slow, cautious adoption due to liability concerns. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI excels at augmenting this task by rapidly identifying anomalies in title searches, flagging missing documents, and summarizing complex closing documents, substantially raising paralegal productivity while the human retains judgment and authority. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can significantly speed up title search review, flag discrepancies, and draft summaries, meaningfully augmenting the paralegal's efficiency even though the arbitration function remains human-led. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with reviewing title searches and document analysis, actual dispute arbitration requires judgment, negotiation, and authority that AI cannot exercise. AI can extract and summarize information from title documents but cannot legally arbitrate disputes or make binding decisions. |
| Task automatability | claude-sonnet-5 | 2/5 | Reviewing title searches involves document analysis AI can assist with, but arbitrating disputes requires judgment, negotiation, and interpersonal trust that current AI cannot perform end-to-end.dispersion.Overall the compound task resists full automation despite partial support for document review. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Legal liability, authorization requirements, and regulatory oversight create strong barriers—a licensed attorney typically must sign off on arbitration and closing decisions. Unauthorized practice of law statutes and malpractice liability restrict autonomous AI substitution. |
| Adoption barriers | claude-sonnet-5 | 3/5 | Real estate closings often require licensed professionals (attorneys/paralegals under supervision) and dispute arbitration involves liability and trust issues, creating moderate-to-strong barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI document review tools reduce costs on the analysis component, but the wage for a paralegal/legal assistant remains substantial and the AI integration still requires significant human review and decision-making, preventing decisive cost advantage. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI can cheaply assist with title document review, but the arbitration/negotiation portion still requires substantial human time and oversight, keeping overall cost savings modest. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Document review and title search analysis tools exist and perform reliably in production (e.g., contract analysis platforms), but no deployed AI system can independently arbitrate disputes or manage complex closing processes end-to-end without human oversight. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Products exist for title search review and document summarization but no deployed system reliably arbitrates disputes between parties; the dispute-resolution component is not addressed by production AI. |
Direct and coordinate law office activity, including delivery of subpoenas.
25CI 20–30 · exposure 20 · augmentation 50 · importance 3.3/5 · click for rater detail
Direct and coordinate law office activity, including delivery of subpoenas.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Law offices are adopting AI for research and document drafting, but coordination and service-of-process functions remain largely manual due to regulatory requirements and the need for human judgment in complex office management. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Legal support and office administration functions have seen slow AI adoption compared to document review or research, with physical coordination tasks especially resistant to automation. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can augment paralegals by automating calendar management, flagging compliance deadlines, and drafting routine coordination emails, moderately improving productivity on administrative portions of the role. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI tools (case management software, scheduling assistants) can help track deadlines, generate subpoena documents, and organize office workflows, providing moderate productivity gains while humans handle coordination and delivery logistics. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can help draft routine documents and manage calendars, the coordination of multiple office functions and the legal/procedural requirements of subpoena delivery require human judgment, local knowledge, and accountability that current AI cannot reliably handle end-to-end. |
| Task automatability | claude-sonnet-5 | 2/5 | This task combines physical coordination (subpoena delivery, often via process servers), scheduling, and office management—largely non-digital coordination work that current AI cannot execute end-to-end, though scheduling/tracking sub-components could be assisted. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Subpoena delivery often requires verification of service, compliance with court rules, and legal accountability; many jurisdictions require human paralegals or attorneys to sign off on service documentation, creating licensing and liability barriers. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No strict licensing requirement for this specific task, but legal service of process has procedural/jurisdictional rules and liability concerns that create moderate friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI assistance with administrative tasks (scheduling, document prep) is cheaper than human labor per unit, but the coordination role requires ongoing human oversight and decision-making that limits cost savings. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI could cheaply handle some tracking/documentation, but the core coordination and physical delivery logistics still require human labor or third-party services, keeping overall costs comparable to or only slightly below human costs. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI tools exist for document drafting and scheduling, but no deployed product reliably orchestrates full law office activity coordination or handles the procedural and compliance aspects of subpoena delivery at production scale. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product directs law office operations or manages physical document delivery; this remains a human administrative and managerial function. |
Appraise and inventory real and personal property for estate planning.
25CI 20–30 · exposure 20 · augmentation 50 · importance 2.9/5 · click for rater detail
Appraise and inventory real and personal property for estate planning.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Legal firms and estate planning practices are moderate adopters of technology; while some use AI for document assembly and research, property appraisal and inventory remain largely manual because of liability, regulatory requirements, and the need for local expertise and on-site work. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Legal services adopt AI unevenly and this task involves physical property assessment, a domain with low digitization and slow AI integration compared to pure document-based legal work. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can augment by automatically organizing property lists from documents, cross-referencing public records, and flagging items for human review, but the core judgment and valuation work still requires the paralegal or appraiser to lead the process. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help organize inventories, generate valuation reports from provided data, and draft estate documentation, meaningfully assisting paralegals even though core appraisal work remains human-driven. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can assist with document review and property classification, but appraising and inventorying requires on-site inspection, valuation expertise, and judgment about condition and market value that current AI cannot perform end-to-end with consistent quality across diverse property types. |
| Task automatability | claude-sonnet-5 | 2/5 | This requires physical inspection, valuation judgment, and often title/deed research plus coordination with appraisers, which current AI cannot execute end-to-end; AI can assist with documentation but not full appraisal and inventory work. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Estate planning appraisals often require licensed appraisers in many jurisdictions for high-value or complex assets; courts and probate proceedings typically expect credible human attestation of inventory and valuations, creating legal barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | While not always requiring a specific license for the paralegal role itself, formal appraisals often need certified appraisers, and estate documentation demands attorney oversight and accuracy for legal/tax purposes, creating moderate liability and procedural barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI document processing is cheap, but the core task—accurate appraisal—still requires licensed appraisers or paralegals with domain expertise whose time cost exceeds current AI automation savings, especially when accounting for liability and error correction. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Physical site visits, third-party appraiser coordination, and valuation judgment still require human labor, so AI only reduces a small portion of overhead costs rather than replacing the core cost driver. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI tools can help organize property lists and pull public records, no deployed product reliably performs complete property appraisal or estate inventory at the quality required for legal documentation without substantial human expert review and local knowledge. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs real property appraisal and personal property inventory for estate planning autonomously; this remains a human field/expert-driven process with AI at most supporting document organization. |
Meet with clients and other professionals to discuss details of cases.
12CI 7–16 · exposure 5 · augmentation 50 · importance 3.9/5 · click for rater detail
Meet with clients and other professionals to discuss details of cases.
12| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Legal services remain heavily relationship-driven and regulation-bound. While law firms adopt AI for document review and research, client-meeting automation adoption remains minimal in practice. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Legal services adopt AI for drafting and research but face-to-face client meetings remain a slow-changing, relationship-driven practice with minimal AI displacement. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist paralegals by preparing case summaries, suggesting discussion points, and documenting meetings post-hoc, improving preparation and follow-up efficiency without removing the human from the client interaction. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist via meeting transcription, note summarization, and prep of talking points or case summaries beforehand, improving efficiency without replacing the meeting itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Client meetings require interpersonal judgment, emotional intelligence, and real-time responsiveness to nuanced concerns. Current AI cannot reliably conduct autonomous client consultations or negotiations that establish trust and gather contextual details. |
| Task automatability | claude-sonnet-5 | 1/5 | This task is inherently a live interpersonal meeting requiring rapport, real-time judgment, and client trust; AI cannot conduct these meetings end-to-end today.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Legal practice requires licensed oversight and direct client communication. Client expectations, attorney-client privilege considerations, and professional ethical rules create strong barriers to full automation of client-facing meetings. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Client relationships, confidentiality, attorney-client privilege considerations, and professional trust create strong organizational and ethical barriers to replacing human interaction here. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI-assisted meeting preparation or follow-up has modest cost advantages, but the core task—meeting with clients—still requires human paralegals. Full substitution is not feasible, limiting cost savings to partial workflow automation. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Since AI cannot perform the meeting itself, there is no viable AI substitute cost to compare; human presence remains necessary, making AI not cheaper for this specific task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI can draft meeting summaries or prepare talking points, no deployed product reliably conducts client meetings end-to-end. Some early chatbots exist for intake, but they lack the judgment and relationship-building critical to legal practice. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product substitutes for a paralegal meeting with clients/professionals to discuss case details; at most AI transcribes or summarizes after the fact. |
Call upon witnesses to testify at hearings.
4CI 0–7 · exposure 0 · augmentation 25 · importance 3.0/5 · click for rater detail
Call upon witnesses to testify at hearings.
4| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | No meaningful adoption of AI for this task is occurring because the task is legally reserved to humans; even digitized subpoena filing still requires human authorization and legal judgment. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Legal services adopt AI for document review and drafting but procedural courtroom/hearing management functions see minimal AI adoption today. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could modestly assist with scheduling, contact list management, or drafting template language for witness notifications, but the core act of calling and compelling testimony remains irreducibly human. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can help organize witness lists, schedules, and reminders, offering modest logistical support, but does not meaningfully transform the act of calling witnesses at a hearing. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Calling upon witnesses to testify requires human authority, legal judgment about timing and strategy, and direct human-to-human communication that establishes legal standing and accountability. No AI system can substitute for the attorney or authorized legal representative who must formally summon witnesses. |
| Task automatability | claude-sonnet-5 | 1/5 | This task requires physically or procedurally coordinating and calling witnesses during live hearings, an interpersonal and procedural act that AI cannot perform end-to-end today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | This task is protected by strong legal barriers: only authorized legal representatives can issue subpoenas and summon witnesses; liability and enforceability require human accountability and signature authority. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Hearings typically require an authorized person (attorney, paralegal under supervision, or hearing officer) to manage proceedings, and legal/procedural rules govern how witnesses are called, creating strong institutional and procedural barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI provides no cost advantage here since the task fundamentally requires human legal authority and presence; human paralegals or attorneys must perform the substantive work. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this specific action, so cost comparison favors the human who must be present and authorized to conduct the hearing. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No production system exists that can independently call and compel witnesses to testify; this requires legal authority vested in a human agent and involves statutory compliance that AI cannot fulfill. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product manages live witness calling at hearings; this remains a human procedural function embedded in courtroom/administrative process. |
Related occupations — Legal
How to read this
A high substitution score does not mean this job disappears — it means a large share of its current tasks face replacement pressure, so the mix of tasks is likely to change. High augmentation alongside substitution typically means the occupation reorganizes around the protected tasks. Wide confidence intervals mean the rater panel disagreed: treat those scores as open questions, not verdicts.
What would change this score
New model capabilities (automatability, feasibility), falling inference costs (cost ratio), regulation and licensing shifts (barriers), and measured sector adoption (velocity) all re-enter at every index release. Each release is recomputed, versioned and kept queryable — scores are claims with a date on them, not permanent labels.