Glossary.

Plain-language definitions for the terms that come up around document automation and AI legal drafting.

Document automation: Software that assembles a document from a template built around rules: based on how you answer a series of questions, the template's logic determines which clauses, text, and calculations to include — and in what form — then inserts them into the right places. Sometimes called document assembly or rule-based drafting. Commonly used for contracts and other legal documents that follow a repeatable structure but vary in their details, such as Lawlift's Document Automation.
No-code legal templates: A template built by defining logic — conditions, calculations, variables — through a visual interface instead of writing code. No-code template editors let lawyers encode their own drafting logic themselves, without a developer, as in Lawlift's template editor.
Legal AI: AI applied to legal work — drafting, editing, reviewing, and analyzing legal documents. In drafting specifically, the more reliable approaches keep the AI within a firm's own approved templates and clauses (see Functional AI), so output stays predictable rather than generic — the basis of Functional Intelligence.
AI contract drafting: Using generative AI to write, or help write, a contract or other legal document — from a prompt, from an existing template, or by adapting a similar past document. Some tools generate free-form text; others, like Functional Intelligence, constrain the AI to a firm's own approved language.
AI contract review (redlining): Using AI to read a contract and flag risks, missing clauses, or deviations from a playbook — typically producing suggested edits as tracked changes for a person to accept or reject, rather than rewriting the document autonomously. See how Functional Intelligence handles this.
Tracked changes: A document-editing feature — native to Microsoft Word and central to legal drafting — that records every insertion, deletion, and formatting change as a visible, attributable edit, so a reviewer can see exactly what changed and accept or reject each one individually.
Functional AI: A way of building software that combines deterministic logic with AI, using each for what it does best: fixed rules and functions handle what should be exact and repeatable, while AI handles the parts that need to understand language and meaning — with agentic AI added where autonomous steps add value. The aim is reliability and flexibility at once, rather than trading one for the other as a rules-only system or a plain language model would.
Functional Intelligence: Lawlift's application of Functional AI to legal document work: specialised functions — drafting, review, translation, anonymization, comparison, formatting — that run inside Microsoft Word, each within your own templates, rules, and standards. See Functional Intelligence.
Document anonymization: Automatically detecting and removing or replacing names, addresses, monetary figures, and other identifying details in a document — typically before sharing it externally, using it as a work sample, or processing it for analysis.
Legal AI Word add-in: An AI drafting or review tool delivered as a plug-in inside Microsoft Word, rather than a separate web application — letting lawyers draft, edit, and review documents with AI assistance without leaving the word processor they already use, like Lawlift's Word add-in.
Deterministic automation vs. agentic AI: Two different approaches to producing legal documents. Deterministic automation follows fixed rules defined in advance — the same inputs always produce the same output, which makes it predictable and auditable. Agentic AI generates or adapts text dynamically based on context — more flexible, less predictable. Many legal teams need both, which is why Lawlift builds both.
EU data residency: Where a provider's servers — and, for AI tools, where the AI model itself runs — are physically located, and which country's or region's data protection law therefore applies. For EU-based legal teams handling sensitive data, keeping both application hosting and AI processing within the EU is often a compliance requirement, not just a preference. See how Lawlift handles this.