Most of the time lost in case-law research is lost before the first good case turns up: vague searches, too many results, and hours reading opinions that turn out to be about something else. The fix is less about which database you use and more about method. Here is a workflow that works in any research tool, free or paid.
Step 1: Frame the question before you search
Write the issue down in one or two sentences, as precisely as you can. A useful research question usually has four parts:
- The legal rule or element in dispute (not the whole claim).
- The jurisdiction whose law controls, and which courts bind the court you're in.
- The procedural posture — motion to dismiss, summary judgment, appeal after trial — because the standard of review changes what a case actually decided.
- The key facts that make your situation different from the general rule.
"Excessive force" is a topic. "Whether, in the Ninth Circuit, it was clearly established that a take-down of a passively resisting arrestee violates the Fourth Amendment, at summary judgment on qualified immunity" is a research question. The second one tells you what to search for and what to ignore.
Step 2: Start with secondary sources
Treatises, practice guides, law review articles, bar association publications and pattern jury instructions have done the first pass for you. They tell you the vocabulary courts use, the leading cases and the elements, and they often flag splits and recent changes. Even twenty minutes here can save hours of blind keyword searching.
If you don't have access to a commercial treatise, check a nearby law library or your state bar's member benefits, and look for pattern jury instructions published by your courts — their comments often cite the key authorities.
Step 3: Find one good case, then expand from it
The fastest route to ten on-point cases is usually one on-point case. Once you have it:
- Read its citations. The cases it relies on for the rule you care about are prime candidates.
- Look at who cites it. Later cases citing it — especially those that discuss the same proposition — show how the rule has been applied since. CourtListener and Google Scholar both show citing opinions for free; commercial citators add treatment signals.
- Use its words. Courts reuse phrases. The exact language a court uses to state the rule is often your best search string.
Step 4: Search precisely
Natural-language search is fine for getting oriented, but precise searches find the cases natural language misses. A few habits help:
- Use phrases for terms of art ("clearly established", "passive resistance").
- Use Boolean and proximity operators to require that two concepts appear near each other, instead of anywhere in a long opinion. CourtListener, for example, supports Boolean operators, phrase searches and proximity searching.
- Filter by court and date from the start. There is no point reading persuasive authority from another circuit before you know what your own circuit says.
- Search for the facts, not only the law. A handful of fact words ("leg sweep", "traffic stop", "handcuffed") can surface the cases closest to yours.
Keep a short research log: each search, where you ran it, and what it produced. It stops you from repeating yourself and makes the work easy to hand off or bill.
Step 5: Read strategically
You don't need to read every result start to finish on the first pass.
- Skim the syllabus or opening paragraphs to learn the posture and the question presented.
- Jump to the analysis of your issue and check whether the court actually decided it, or only mentioned it along the way (see our article on holdings and dicta).
- Note the disposition and the standard of review.
- Save a pinpoint quote for every case you keep, so you're not hunting for it later.
Two more habits pay off quickly. First, when a case looks promising, check its subsequent history before you invest time in it — a case that was reversed or limited is a detour. Second, keep a running list of rejected cases with a one-line reason ("different posture", "dicta only", "state-law claim"). When a partner or client asks whether you considered a case, you will have the answer.
Step 6: Know when to stop
Research expands to fill the time available. A practical stopping rule: stop when new searches keep returning cases you have already seen, you have checked binding authority first and then the best persuasive authority, and you have looked for contrary cases — not just helpful ones. Then verify everything you plan to cite: read the full opinion and confirm it is still good law.
Where AI fits — and where it doesn't
AI tools can speed up the first steps: planning searches, reading many opinions quickly and ranking which ones matter. They are not a substitute for reading the cases. Language models can produce case names and citations that don't exist, so any AI-assisted research should work only from opinions actually retrieved from a real database and should link you to every source.
That is how It's On Point works: you submit one research question, we search CourtListener, read the retrieved opinions, and report only real, linked cases, each with a relevance note and a verbatim quote that is checked against the opinion text.
For licensed attorneys. General information about legal research, not legal advice. Always read the full opinion and confirm that every authority is still good law before relying on it.
If you have a question you'd like a head start on, start a research request — $99 per topic. See how it works first if you'd like the details.