Dictating a Legal Research Plan in the First Hour of a Matter

The first hour of a new matter is the sharpest hour, and it is also the one that decays fastest. During the intake call, the client names the doctrine, the regulator, the case date, and the clause that matters, and ten minutes later the details are already evaporating under the day's other demands. Dictating a legal research plan from the intake call preserves that sharpness, because the plan captures the reasoning while it is still connected to the client's own words. The plan that gets written from memory on day three is the plan that gently flattens the details, and a flattened research plan starts a matter in the wrong direction without anyone noticing.

The decay has a mechanism. The intake call gives you three kinds of information: facts, which the client states; questions, which the client's situation raises; and signals, the words the client repeats, the clause the client keeps referencing, the forum the client worries about. The signals are the sharpest facts and the first to go, because they do not look like facts. A research plan dictated in the hour captures the signals as a separate list, labeled intuition, and the list is what makes the research assignment distinctive rather than generic. The intuition list is also what the plan's first review tests, since a plan that leads with the client's repeated words is a plan the partner can confirm in one glance.

The cost of decay is visible in the final memos. The associate who inherits a research assignment from a day-old plan produces a memo that surveys the doctrine competently and misses the issue the client actually cares about, because the plan did not say the issue mattered. The client reads the memo and asks the question that should have been the memo's thesis, and the matter spends a week angling back onto its real course. The intake-hour plan is the cheapest insurance against the week, and it costs one dictation session. The session also produces the plan's first edit flag, since the client's own words in the memo are the ones the plan must not translate away.

The pipeline is two documents from one call. First, dictate the matter memo while the call is still fresh: the client, the dispute or deal, the lawyer's own understanding of what is being asked, the documents received, and the open questions. The matter memo is the record, and it is written in the client's vocabulary, not the firm's. Second, dictate the research plan from the memory of the same call, and keep the plan in the research planner rather than buried in the matter memo, because the plan is a working document that will be edited, tested, and handed to an assignee.

The matter memo and the plan diverge in a deliberate way. The memo says what the matter is; the plan says what the team needs to learn before it can answer. The distinction forces the sharp question: if the memo and the plan read alike, the plan has not been drafted, it has been copied. The plan starts with the legal question in the client's own words, quoted, and then restates the question in legal terms. The distance between the two formulations is the research gap, and the gap is the assignment. The gap sentence also becomes the plan's first line, so the assignee reads the client's question before the case law, which is the order that keeps the research pointed.

Dictate the plan in the same session as the memo, because the second pass over the call's memory is where the plan's structure comes out. The first dictation produces the memo's chronology; the second produces the plan's priorities. The two-pass method also forces the conflict to the surface: a client description that does not fit any known doctrine is the research opportunity, and the plan should name it as a question rather than assume the doctrine. An intake session that ends with both documents costs twenty minutes, and it converts a call from a note into an assignment. The two-pass session also produces the follow-up list the client will not think of, since the questions the call raised are written down while the call's memory is still warm.

The plan's body is a list of research tasks, and each task gets three fields dictated aloud: the database, the keyword string, and the jurisdiction. Saying the database names the tool and the subscription; saying the keyword string forces the drafter to think like a searcher rather than a reader; and saying the jurisdiction narrows the task to the place where the answer lives. The three fields are what make the plan assignable, because an assignee can execute a task that names its tool and its search terms without improvising the approach. The three-field task is also the unit that the plan's progress gets reported against, which makes the plan the thing the weekly check reads instead of the memo's status.

Dictate the keyword strings as exact strings, not as descriptions. The difference matters: a plan that says research whether notice was proper is a plan the assignee will interpret ten ways, while a plan that says run this query with these terms is a plan that returns a comparable answer regardless of who runs it. Write the string as you would type it, and say the operators aloud: the term and quote and the phrase. The exact string is also the testable part of the plan, because the morning-after pass will run it and check the result. The exact string is also the part of the plan the client can see, since a client who reads the keyword string understands the approach better than a client who reads the doctrine.

Jurisdiction deserves its own sentence in each task. A plan that asks for the standard of review without naming the court is a plan that drifts, because the assignee will research the nearest authority rather than the controlling one. Say the court, the circuit, or the state, and when the matter may move, say both and mark the question. Plan the backup as well: the database that fails, the jurisdiction that lacks authority, the string that returns nothing. A research plan with a fallback path is a plan the assignee can follow without a second call, and following without a call is the plan's real test.

Every research task gets a deadline, dictated with the task and not appended later. The deadline is not a courtesy; it is the thing that connects the research plan to the matter's calendar. Say the court date, the filing window, and the partner review in the same breath as the task, so the plan reads as a schedule rather than a wish list. A research plan without dates is a plan that sits in the planner until the court date forces a compressed week, and compressed research is the kind that cites a case for the first time in the draft. The deadline is also what the plan uses to prioritize.

The order of the dates matters as much as their presence. Dictate the matter's anchor date first, the dispositive motion deadline or the filing date, and then work the research tasks backward from it, so the plan shows which tasks produce the memo the dispositive cut-off depends on and which can wait for the second phase. The backward order is the difference between a schedule and a timeline, because it names the dependency: the motion memo is built on the research, the research is built on the keyword test, and the keyword test is built on the intake call. A plan that has no dependency order is a plan that will be executed in the order the assignee happens to prefer.

The partner review gets its own date in the same structure. Dictate the review as a slot in the plan, the day the memo is expected, and name the reviewer and the review's format, at the desk or in the email thread, so the assignee knows what the delivery looks like. The review date is the plan's real deadline, because the court date only matters if the memo reaches the reviewer in time. When the plan's dates are stated and the review slot is set, the plan becomes the matter's schedule, and the schedule is the thing the team can hold each other to.

The morning-after pass is the plan's quality check, and it runs before the assignee starts, not after the memo comes back. Open the plan, take the first keyword string, run it in the named database, and look at the results. If the string returns the cases the team expected, the plan is sound. If it returns an unrelated haystack, the string needs fixing, and the fix belongs in the plan now, while low heat is on it. The pass takes twenty minutes, and it is the only way to know that the plan the assignee will follow points somewhere real.

The pass catches three failure classes. The too-broad string that floods the results with noise; the too-narrow string that returns nothing and will send the assignee down a speculative path; and the jurisdiction mismatch, where the string is right but the query does not include the court the matter will live in. Each failure gets fixed by testing a variant on the spot, and the variant that works gets dictated into the plan as a replacement. The plan that has survived a test is a plan the team can stop investigating and start executing. The tested plan also survives the assignee's own questions, because the tested string is the one the assignee can defend without re-running it.

Close the pass with the assignment handoff: the plan saved, the assignee named, the first check-in set, and the summary stated aloud so the plan's assumptions are on the record. The handoff is the reason the morning-after pass belongs to the drafting attorney and not to the assignee, because the person who heard the intake call is the only person who can judge whether the search results are the right ones. A research plan built in the first hour and tested the next morning is the plan that keeps the matter's direction on the client's actual question. The handoff line is also the plan's close, and a plan that closes with the assignee named and the check-in set is a plan that the matter's file can quote a month later.