Matter-scoped AI vs research-scoped AI: how to tell which one you actually need

What matter-scoped AI actually does

Matter-scoped AI lives inside your practice management stack. It is given access to your firm’s own matters, documents, correspondence, and internal notes, and it answers questions about those. Typical tasks: draft a client update email in the firm’s voice; summarise a long matter file for a supervising partner; review a vendor contract inside the matter and flag clauses against your firm’s playbook; draft a chronology from the stored documents; generate a timeline of events from the matter’s email history.

What matter-scoped AI is not designed to do: reason over the open universe of Australian case law. If you ask a matter- scoped assistant “is the duty of care in Donoghue v Stevenson still good law after the Civil Liability Acts?” it either has to reach outside the matter (which many matter-scoped products refuse by design) or pass the question to a bolt-on research layer that has its own product boundaries and its own pricing.

What research-scoped AI actually does

Research-scoped AI lives outside practice management. It is given access to the public body of Australian case law and legislation, and it answers questions about those. Typical tasks: find the leading authority on a duty of care question in a specific Australian jurisdiction; show the treatment chain for that authority across every later case; generate an AGLC4 citation ready to paste into a submission; read a long judgment and produce a plain-English explanation anchored to specific paragraphs of the source text; compare two cases and surface the distinguishing fact pattern.

What research-scoped AI is not designed to do: draft a client email in your firm’s voice from a specific matter file, run your trust account, or answer questions about your own documents unless you hand them to it explicitly. It treats each research session as a public-law question, not as a private-matter question.

How to tell which one you need

Ask yourself two questions.

Question 1: whose data do you want the AI to read?

If the answer is “my firm’s matters, emails, and precedent library,” you need matter-scoped AI. If the answer is “the public body of Australian case law and legislation,” you need research-scoped AI. If the answer is both, you need two tools.

Question 2: what does the output look like?

If you want a drafted client email, a summarised matter file, a filled-in form, or an internal chronology, that is a matter-scoped output. If you want a cited authority, a treatment classification badge, an AGLC4 citation, or a shareable research packet that a supervising solicitor can review, that is a research-scoped output.

Why many firms use both

A typical Australian solo or small-firm workflow needs both. The matter-scoped assistant drafts the client correspondence and keeps the firm running. The research-scoped assistant answers the legal question that the matter turns on. The two surfaces live in different tools because they solve different problems and they have different data access rules. A matter-scoped assistant that opens the gates to strangers’ judgments starts to look like a research tool. A research- scoped assistant that reads your firm’s private matters starts to look like a practice management tool. Both directions dilute the product.

What you want from the matter-scoped tool: deep integration with your document automation, your time tracking, your billing, your client portal. What you want from the research- scoped tool: deep coverage of Australian case law and legislation, paragraph-level citation tracing, treatment classification badges, AGLC4 citation generation, and a shareable packet for collaboration.

How the pricing usually works

Matter-scoped assistants are usually sold as an add-on to a practice management subscription. The price depends on the underlying practice management plan and is often not listed publicly until you request a quote. Contracts are commonly multi-year. Research-scoped assistants are usually sold standalone, with a published monthly price and no multi-year contract. CaseSharp is free to start, with tier details published on the pricing page when we open.

Frequently asked questions

Can one product do both?

In practice, the products that try to do both end up privileging one surface. Practice management vendors that bolt on a research layer usually gate it behind the matter-scoped plan and ship a narrower research surface than a purpose-built research tool. Research tools that try to become practice management stacks end up with a shallow matter-management surface. Most serious firms run two tools that each do their job well.

Which one does CaseSharp sell?

CaseSharp is research-scoped. Sharp reasons over the public body of Australian case law and legislation, surfaces paragraph-level citation traces, generates treatment classification badges, and produces AGLC4 citations. CaseSharp does not run your trust account, send client emails in your voice, or manage your matters.

Will CaseSharp integrate with my practice management tool?

An MCP server is published so any AI coding or research tool that supports MCP can call CaseSharp’s research tools directly. Direct integration with individual practice management products is on the roadmap.

What about data residency?

CaseSharp runs on Australian infrastructure with full Australian data residency. No offshore processing of your queries or research history. This is a matter of policy, not a feature flag.