Will AI Steal My Job? · Role analysis

Paralegal

O*NET 23-2011.00 ESCO: Paralegals and legal assistants

High exposure

Paralegals assist solicitors and lawyers with research, document drafting, case preparation, and client communication. They handle the high-volume, process-driven legal work that keeps a firm running — from document review and bundling to legal research, client onboarding, and case administration.

Section 01

Task Map

Task	AI impact	Why
Conduct legal research on case law and statutes	🔴 High exposure	Legal AI research tools (Harvey, Lexis+ AI, Thomson Reuters CoCounsel) now perform this task faster and often more accurately than junior paralegals. This is one of the most disrupted legal tasks.
Review and summarise large document sets	🔴 High exposure	Document review — reading hundreds of contracts to extract key terms or flag issues — is the classic AI legal use case. eDiscovery and contract review AI tools are already deployed at scale.
Draft standard legal documents and correspondence	🔴 High exposure	Template-based drafting of NDAs, consent letters, and standard agreements is heavily automatable. AI tools produce first drafts that require relatively little editing for routine work.
Manage case files and court deadlines	🟡 Changing	Case management systems increasingly automate deadline tracking and document management, but someone must still oversee the workflow and catch exceptions.
Prepare court bundles and hearing files	🟡 Changing	Document bundling software automates much of this, but organising complex litigation files, understanding what the judge needs, and adapting last-minute still requires a knowledgeable human.
Communicate with clients on case progress	🟡 Changing	Routine updates can be automated, but a client call to explain a court delay, answer a worried question, or manage expectations is a human client-service skill.
Conduct land registry and company searches	🔴 High exposure	Automated search ordering and results processing is already standard in conveyancing and corporate work. Much of this is already digital workflow, not manual search.
Support compliance and AML checks	🔴 High exposure	AML screening, ID verification, and sanctions checking are largely automated in modern legal firms. Human oversight of flagged cases is still needed, but volume processing is automated.

Section 02

What Stays Human

Complex judgment calls within cases When a document set throws up something unexpected, when a client situation doesn't fit the standard workflow, or when an issue requires genuine legal judgment — not just task completion — a thinking, trained paralegal adds value that automation cannot.
Client-facing communication Many clients form their view of a firm through their paralegal contacts. The reassuring, knowledgeable human voice that answers a worried call or explains a process clearly is a client-service skill that matters to client retention.
Oversight of AI outputs AI-generated legal documents and research contain errors that can have serious consequences. An experienced paralegal who knows the case, knows the law, and can spot where AI has gone wrong is performing critical quality-control work.

Section 03

What to Do Next

Qualify as a solicitor or CILEx lawyer. The Solicitors Qualifying Examination (SQE) route and CILEx are more accessible than ever. Paralegal experience counts — and qualification gives you the professional accountability and practice rights that are the most protected part of legal work.
Become expert in your firm's legal technology tools. The paralegal who can train others on AI research or document review tools, who understands their limitations and can design quality-checking workflows, is taking on a new role that sits above routine automation rather than beneath it.
Specialise in complex, high-judgment areas. Paralegals in insolvency, complex litigation, immigration, or family law — where every case is genuinely different and client complexity is high — are doing work that requires accumulated expertise and contextual judgment. Volume commercial work is more vulnerable to automation than complex caseloads.

Sources: O*NET Online (onetonline.org) · ESCO (esco.ec.europa.eu) · All task data cross-referenced against O*NET occupation profiles. This analysis uses task-level exposure, not occupation-level prediction.