Frequently Asked Questions About Legora Agentic Law Transformation Framework
22 answers covering everything from basics to advanced usage.
// Basics
What is the Agentic Operating System (AOS) in Legora's framework?
The AOS is a single connected system with six engineered layers that facilitates the flow of information, communication, and execution of legal work. The layers are: Foundation Models, Agentic Harness, Data and Integrations, Context and Knowledge, Legal Skills and Capabilities, and Enterprise Security and Governance. Unlike point solutions for narrow problems, the AOS achieves compound effects by having all layers work together under a single roof.
What are the three phases of Agentic Law transformation?
Phase 1 is Leverage — AI handles high-volume, low-complexity work and lawyers get capacity back. Phase 2 is Differentiation — service models adjust so a team of five competes like a team of 50, winning on service delivery innovation rather than headcount. Phase 3 is Reinvention — AI-native legal teams rethink processes from the ground up, productise knowledge, and create entirely new business models. Most legal organisations are currently in Phase 1.
What is the difference between an agentic harness and a raw LLM?
A raw LLM is stateless — it has no memory, no guardrails, and no ability to carry a task through multi-step execution. The agentic harness gives the LLM its spine: memory across interactions, guardrails for safe operation, tool routing to connect with external systems, and the ability to execute retrieval, checking, flagging, drafting, and reviewing in sequence. Without a harness, an LLM cannot perform the sustained, multi-step legal work that Agentic Law requires.
What inputs do I need to start applying the Legora Agentic Law Framework?
You need five required inputs: your legal team type (in-house, law firm, or legal services), current AI maturity (what tools and workflows are automated), team size and structure (headcount, jurisdictions, stakeholders served), primary pain points (high-volume/low-complexity work, bottlenecks, service gaps), and strategic ambition (efficiency, differentiation, or full reinvention). Optional but valuable inputs include your existing tech stack (DMS, signing platforms, data sources) and regulatory or jurisdictional scope.
Why does the Legora framework say 'Legal AI is dead'?
The framework declares Legal AI dead not because AI in legal is over, but because first-generation Legal AI — which made lawyers faster at the same work — is insufficient as a destination. Legal AI made workflows smarter and quicker but did not fundamentally change the structure of legal work. Agentic Law represents a generational leap that changes what a lawyer can be, not just how fast they work. Legal AI is now a foundation to build on, not an endpoint to aim for.
// How To
How do I write Skills for the Agentic Operating System?
Write Skills as if you are briefing a new team member in plain English. Each Skill should capture how your team structures its work product, firm-specific language and style guides, practice-area workflows, risk tolerance rules and escalation triggers, and client-specific preferences. Start with a library of out-of-the-box Skills for common legal work, then customise them to your team's standards. Involve practicing lawyers directly in Skills authorship — they understand the domain better than anyone.
How do I set up Monitors for regulatory horizon scanning?
Select the jurisdictions and topics relevant to your team (e.g. EU antitrust, UK FCA regulations, employment law). The Monitor tracks regulatory changes in real time across official sources, updated hourly. The agent then stitches regulatory changes together with your existing policies and documents to produce gap assessments, compliance briefs, or policy updates — all cited to original sources. Format outputs for board briefings or client distribution to maximise value communication.
How do I design human oversight checkpoints for agentic legal workflows?
For any long-running agentic task, follow this pattern: set the objective clearly, have the agent produce a Plan before executing, review and approve or modify the plan, allow the agent to flag uncertainty and ask clarifying questions mid-task rather than guessing, and have the agent return completed work product to a List for human review. Never design a workflow where the agent runs to completion without checkpoints — the plan approval moment is critical for setting intent.
How do I use Lists in the Legora Agentic Operating System?
Lists are the connective tissue between agents, lawyers, and stakeholders. Use them for closing checklists, due diligence trackers, case chronologies, and in-house intake. Each list item is an atomic unit — an object with structured metadata including status, assignee, parties, jurisdiction, source citation, and dates. The agent can populate, update, and flag items autonomously while humans review, annotate, and reassign. Lists replace Excel, Word, and Post-its because they carry executable structure, not just text.
How should I prioritise which legal workflows to automate first with the Legora framework?
Prioritise by three criteria: volume (high-frequency tasks), complexity (start with low-to-medium complexity), and pain (where time is most wasted). Common high-value starting points include unstructured data room organisation, closing checklist generation, due diligence report drafting, and regulatory monitoring. These represent Phase 1 Leverage use cases that deliver immediate capacity back to lawyers while building organisational confidence in agentic execution before progressing to more complex Differentiation and Reinvention use cases.
// Troubleshooting
My legal team deployed AI tools but nothing feels different — what went wrong?
You are likely stuck in the Legal AI trap — applying AI to existing structures and workflows, making lawyers faster at the same work without fundamentally changing how work gets done. The Legora framework diagnoses this as being stuck in early Phase 1 without a path forward. The fix requires rethinking service delivery models, not just adding speed. Map your stack against the six AOS layers, identify gaps, and design Skills that encode your team's unique knowledge rather than simply automating existing document workflows.
Why does the agentic agent produce poor results on complex legal tasks?
The most common cause is skipping the plan approval checkpoint. When you send an agent to complete a long-running task without first reviewing its proposed plan, you fail to set the right intent from the outset. Other causes include missing context layers (no access to precedents, clause libraries, or matter history), using a raw foundation model without an agentic harness, or relying on horizontal AI tools rather than a vertical legal system with domain-specific data and workflows.
Our lawyers resist AI adoption — how does the Legora framework address change management?
The framework treats change management as equally important to technology deployment. It prescribes the Legal Engineer role — former lawyers who are deep AI experts and understand legal workflows, risk tolerance, and organisational culture. Deploy as many Legal Engineers as software engineers. They serve as embedded partners who translate between technology capability and legal practice reality. The framework explicitly warns against treating technology deployment as sufficient without proportional investment in change management support.
How do I prevent building point solutions instead of a connected agentic system?
The Legora framework explicitly warns against building individual tools for narrow problems because they cannot achieve the compound effect that comes from all layers working together under a single roof. Instead, map every technology investment against the six AOS layers and ensure each tool integrates into the unified architecture. Every data source, workflow, and agent should share context, knowledge, and governance. Point solutions create siloes; the AOS creates compounding value across all legal work.
// Comparisons
How does the Legora Agentic Law Framework compare to deploying Microsoft Copilot or generic AI assistants for legal teams?
Microsoft Copilot and generic AI assistants are horizontal platforms — built for everyone, therefore built for no one in the legal domain. The Legora framework requires a full vertical legal stack with six specific layers including legal-specific data integrations, domain context, custom Skills, ethical walls, matter-centricity, and audit trails. Horizontal AI with legal plugins lacks the agentic harness, legal knowledge architecture, and governance infrastructure needed for autonomous multi-step legal execution with proper oversight.
How does the Legora framework compare to traditional legal project management approaches?
Traditional legal project management organises work manually using spreadsheets, emails, and static checklists. The Legora framework replaces this with Lists as executable, structured objects where agents autonomously populate, update, and flag items while humans review and reassign. The framework also adds agentic execution within each task — agents can draft, research, review, and complete work products. Traditional approaches manage workflow; the AOS manages workflow and executes the underlying work autonomously with human checkpoints.
How is Knowledge Productisation different from creating legal document templates?
Templates are static documents that require manual completion and expert judgment to use correctly. Knowledge Productisation encodes the same expertise into executable Skills and automated workflows within the AOS — making it scalable and deployable on demand. A productised knowledge system can handle automatic intake, triage, and respond to legal queries autonomously, applying the team's standards and risk tolerance rules. Templates capture form; Knowledge Productisation captures reasoning, judgment, and process.
// Advanced
Can the Legora framework work for small law firms with fewer than 20 lawyers?
Yes — in fact, the framework's Phase 2 (Differentiation) is specifically designed for this scenario. A team of five can compete like a team of 50 by deploying agentic execution for high-volume work, productising knowledge into Skills, and offering managed services like regulatory monitoring that were previously too cost-prohibitive. Small firms benefit disproportionately because they gain the most competitive leverage from service delivery innovation. Start with Phase 1 use cases like data room organisation or closing checklists and expand.
What does Phase 3 Reinvention actually look like for a legal team?
In Phase 3, legal teams are AI-native — processes are designed from scratch rather than adapted from pre-AI workflows. Knowledge is productised: static guides become AI-powered workflows, handbooks become interactive Skills, and legal expertise is deployable 24/7 at scale. Automatic intake, triage, and query response operate without manual intervention. New business models emerge: subscription-based legal services, managed regulatory monitoring, and client-facing portals replace hourly billing for routine matters. This phase is still new territory for most organisations.
How do sub-agents work in complex legal transactions like due diligence?
In a complex due diligence, a main orchestrating agent launches specialised sub-agents to work in parallel across distinct workstreams — corporate, IP, employment, real estate, data protection, and tax. Each sub-agent completes its portion of the analysis and returns its work product to the main agent. This enables work that would take six hours sequentially to complete in parallel. The main agent synthesises the findings, and the combined output is surfaced in a List for lawyer review with structured metadata per workstream.
What does proper work packaging look like for agentic legal output?
The framework emphasises that static Excel files and plain Word documents fail to communicate value. Proper work packaging includes interactive deliverables, structured portals, visual outputs like corporate structure diagrams, risk heat maps, and radar charts. Regulatory updates should be formatted for board briefings. Due diligence outputs should include structured findings with metadata. The perception of value matters — stakeholders who receive a well-packaged interactive deliverable understand the quality of work better than those receiving a flat document.
Can in-house legal teams use the Legora framework differently than law firms?
Yes. In-house teams typically focus on compliance monitoring at scale (using Monitors across multiple jurisdictions), intake triage and query routing for internal stakeholders, and consistent AI governance across departments. Law firms focus more on client-facing service delivery innovation, managed services, and productising legal knowledge for external consumption. Both follow the same three-phase progression, but their Phase 2 and Phase 3 outcomes differ — in-house teams optimise internal service delivery while firms create new revenue models.