Legal AI and operational AI are not competing. They are answering different questions.
A firm can own an excellent research assistant, a capable drafting tool, and a strong review product, and still lose several hours a week to not knowing what is happening across its own matters.
Last reviewed
This page describes categories of legal AI tooling rather than ranking named products, because capabilities change quickly and vary by plan. It is published by Makase and reflects the landscape as of August 2026. Evaluate any specific product against its own current official documentation.
Standalone legal AI tools work on what you hand them: a question, a document, a file set. Makase Law's intelligence layer, Tretis, works across the firm's connected matters to coordinate operations. Most firms benefit from both, and neither replaces attorney judgment or human review.
Six categories worth distinguishing
Legal AI is discussed as one market. It is really several, and the products in each are strong at different things. Understanding which category a tool belongs to prevents most disappointment.
Legal research AI
Answers legal questions against curated case law, statutes, and regulations, with citations. Strong on jurisdictional depth. Has no view of your firm's matters unless facts are pasted in.
Legal analysis and strategy AI
Analyzes facts, precedent, and outcomes to support litigation strategy or valuation. Useful at volume. Generally blind to matter status, client relationship, and firm economics.
Drafting AI
Produces first drafts of contracts, pleadings, memos, and correspondence from templates or prior work. Fast and consistent. Rarely knows the firm's live matter context or house style unless supplied.
Document review and summarization AI
Processes large document sets to extract facts, flag issues, and summarize. Valuable in discovery and diligence. Does not carry matter economics or coordination.
General-purpose assistants
Flexible, inexpensive, and available immediately. No verified citations, no access to firm information, and no built-in confidentiality safeguards designed for professional practice.
Operational intelligence in an operating system
Works across the firm's connected matters, tasks, communication, and billing to coordinate work. Not a substitute for specialist research or high-volume review.
The limiting factor for AI in a law firm is rarely the model. It is what the model is allowed to see.
Side by side
| Dimension | Makase Law | Standalone legal AI tools |
|---|---|---|
| What it sees | The firm's connected work: matters, stages, ownership, documents, communication, time, and billing activity. | Whatever is provided to it in a session: a document, a question, a transcript, or a set of files. |
| What it is for | Operational coordination. Knowing what is happening, what is stuck, what needs attention, and what should be prepared next. | Substantive legal tasks. Research, analysis, drafting, and review, each within a bounded input. |
| Typical output | Matter summaries, prioritized work, follow-up prompts, and drafts of routine communication for a person to review and send. | Answers with citations, drafted documents, extracted facts, or summaries of the material supplied. |
| Where it is strongest | Anything requiring knowledge of the firm's live work across many matters at once. | Anything requiring depth on a specific legal question or a specific body of documents. |
| Where it is weakest | It does not perform legal research, form legal arguments, or replace dedicated high-volume review tooling. | It cannot tell you which matters are stalled, who owns the next step, or what has gone unanswered for two weeks. |
| Human review | Designed so a person reviews and sends anything that reaches a client, with review built into the workflow. | Depends on the product and the firm's own discipline. Professional responsibility for output remains with the attorney in all cases. |
| How firms usually adopt it | As the system the firm runs on, configured around its practice during implementation. | As a tool alongside the system of record, adopted per person or per team. |
The firm's connected work: matters, stages, ownership, documents, communication, time, and billing activity.
Whatever is provided to it in a session: a document, a question, a transcript, or a set of files.
Operational coordination. Knowing what is happening, what is stuck, what needs attention, and what should be prepared next.
Substantive legal tasks. Research, analysis, drafting, and review, each within a bounded input.
Matter summaries, prioritized work, follow-up prompts, and drafts of routine communication for a person to review and send.
Answers with citations, drafted documents, extracted facts, or summaries of the material supplied.
Anything requiring knowledge of the firm's live work across many matters at once.
Anything requiring depth on a specific legal question or a specific body of documents.
It does not perform legal research, form legal arguments, or replace dedicated high-volume review tooling.
It cannot tell you which matters are stalled, who owns the next step, or what has gone unanswered for two weeks.
Designed so a person reviews and sends anything that reaches a client, with review built into the workflow.
Depends on the product and the firm's own discipline. Professional responsibility for output remains with the attorney in all cases.
As the system the firm runs on, configured around its practice during implementation.
As a tool alongside the system of record, adopted per person or per team.
How firms use them together
Research stays specialist
Keep the research product your attorneys trust. Makase does not compete with it.
Review stays specialist
High-volume discovery or diligence review belongs in a tool built for it.
Drafting where it earns its place
Use drafting AI where template volume justifies it, and connect the output to the matter.
Operations in one place
Matters, coordination, billing, and follow-up belong in the system the whole firm shares.
One record of the work
Whatever tools are used, the matter should end up with a single, complete history.
Review discipline throughout
Whatever produced the draft, a person reviews it before a client sees it.
Which approach to reach for
Reach for a standalone tool when
- The task is a legal question needing authority and citations
- A large document set must be reviewed or summarized quickly
- A specific document needs drafting from a template or precedent
- Depth on one substantive task matters more than firm context
- The need is individual rather than firm-wide
- A specialist product is already trusted by the attorneys using it
Reach for operational intelligence when
- Nobody can say what is happening across matters without asking
- Follow-up depends on someone remembering to send another email
- Billing is reconstructed rather than assembled continuously
- Coverage during absence is genuinely difficult
- Client updates take longer to prepare than to deliver
- The firm's constraint is coordination rather than legal capability
Questions about legal AI tools
- Do we still need a legal research tool if we use Makase?
- Very likely yes. Makase does not perform legal research and does not attempt to replace dedicated research products. Operational intelligence and legal research answer different questions, and most firms will want both.
- What is the difference between operational AI and legal AI?
- Legal AI works on legal substance: research, argument, drafting, and review of documents you provide. Operational AI works on how the firm runs: what is happening across matters, what needs attention, what has gone unanswered, and what routine communication should be prepared for review.
- Why can't a general AI assistant do this?
- A general assistant only knows what is typed into it. It cannot see the firm's matters, deadlines, ownership, billing status, or communication history, and it carries no built-in handling of client confidentiality. It can be useful for isolated drafting, but it cannot coordinate a practice.
- Is embedded AI less capable than specialist AI?
- On specialist tasks, often yes. A dedicated research or document-review product is usually stronger at the thing it was built for. Embedded operational AI is stronger at anything requiring knowledge of the firm's live work, because that is the only place that context exists.
- Can these tools be used together?
- Yes, and most firms do. A practical arrangement is a research tool for legal questions, a drafting or review tool where volume justifies it, and an operating system carrying matters, coordination, billing, and follow-up.
- How should we evaluate any AI tool for a law firm?
- Ask what information it can see, what it does with that information, what it produces, who reviews the output before it reaches a client, and how confidentiality and retention are handled. Those five questions separate most serious products from most demonstrations.
- Does Makase provide legal advice?
- No. Makase is software supporting the operational work surrounding a firm's legal services. It does not provide legal advice, practice law, or make substantive legal determinations, and attorneys remain responsible for professional judgment and review.
Continue reading
- AI operating system for law firms
What the term means, and how to evaluate it.
- Makase Law compared with Clio
Two platforms with embedded AI, compared honestly.
- Choosing a law firm technology stack
Deciding what to keep, connect, or consolidate.
- Makase Law
The connected operating system for modern law firms.
Keep your legal AI. Fix the coordination.
We will show you what operational intelligence looks like across your firm's actual matters.