Five lessons, about an hour. A meeting that produces no written record is a meeting that happened to nobody. This module is the discipline that turns an hour of talk into an agenda that made it shorter, notes worth keeping, actions with names and dates on them, and the follow-through that gets them actually done.
Adaeze has a call booked with MedSupply — the equipment supplier from module 4's handover file — to renegotiate a contract that renews automatically in three weeks. Thirty minutes on the calendar. No agenda. Left to itself, this call will run forty-five minutes, cover the pricing question twice because nobody wrote down that it got answered the first time, and end with "let's circle back" on the one thing that actually needed a decision.
An agenda is not a formality you add to look organised. It is the single intervention most likely to make a meeting shorter, because it forces the decision about what the meeting is for to happen before the meeting, when there is time to think, instead of during it, when there is not.
A meeting without an agenda is a conversation. A meeting with one is a piece of work with a start and an end.
A working agenda is short enough to read in the time it takes to sit down. It has four parts, in this order:
For the MedSupply renewal call, the objective read: decide whether to accept the proposed 8% increase, counter, or move to the alternative supplier scoped in June. Three topics, each timed — MedSupply presents the increase, we compare it against the June quote, then decide — and one decision needed by the end: accept at 8%, counter at a number, or give 30-day notice today, since the window closes in three weeks. The full template, ready to fill in for your own meetings, is in the worksheet pack above.
Notice what this does before the call even starts. MedSupply now knows they need to justify the number, not just state it. Adaeze knows the meeting cannot end vaguely — one of three named outcomes has to happen. And you, taking notes, already know the shape of what you are listening for.
An agenda sent an hour before a call is barely better than no agenda — nobody has time to look at the pre-read or think about the decision in advance, so the meeting still does that thinking live, which is exactly what the agenda was meant to prevent. Twenty-four hours is the useful minimum. For anything with a pre-read longer than a page, give it two working days.
Sometimes the other side calls the meeting and won't produce an agenda — a client's accountant, say, who "just wants to chat through a few things." You can still impose structure from your side: send a one-line note beforehand — "Ahead of tomorrow, is there anything specific you'd like Adaeze to have looked at, or a decision you're hoping to reach?" — which usually surfaces the real objective even when the other party hadn't named it to themselves.
A good agenda gets referenced mid-meeting — someone says "we're on the pricing topic, let's park the delivery-schedule question for its own slot." A bad one gets sent and never mentioned again, because it was written to satisfy a checklist rather than to actually run the meeting. If nobody refers back to it, the objective and decisions sections probably weren't specific enough.
Take a real meeting you have coming up this week — yours, a client's, anyone's. Write the objective sentence first. If you can't write it in one sentence without the word "discuss," that's the sign the meeting itself needs rescoping before the agenda does.
The MedSupply call happens. Thirty-eight minutes of talk: pleasantries, a story about a shipping delay in April, the actual pricing negotiation, a tangent about a trade show, and finally the decision. A transcript of all of it would run to nine pages. Nobody will ever read a nine-page account of this call — not Adaeze, not you, not future-you trying to remember what was agreed six months from now.
Meeting notes are not a record of what was said. They are a record of what changes because of what was said.
Before writing anything down during a meeting, run it through one question: if this sentence didn't exist, would anything Adaeze does next week be different? If yes, capture it. If no — however interesting, however true — leave it out. The shipping-delay story is genuinely interesting. It changes nothing about what happens next. It does not belong in the notes.
Good meeting notes sort everything into four buckets. Nothing else earns a line:
"So yeah, we had that issue back in April with the shipment, took like three weeks longer than expected, real headache honestly. Anyway on pricing — so we're looking at, I think it's an 8% increase this year, mostly freight costs, everyone's dealing with that. Bola mentioned she saw a trade show in Lagos next month, might be worth attending. Adaeze pushed back a bit, asked if there was room to move — MedSupply said they'd need to check with their finance team but could probably do 5% if we signed today rather than at renewal. Adaeze said yes to that. Oh, and someone needs to check whether the current contract has an early-termination clause, wasn't sure. Also delivery schedule for Q4 still needs sorting, that's a separate conversation apparently."
Decision: Accepted 5% renewal increase (countered down from 8%), effective immediately rather than at the automatic renewal date.
Action: You — confirm the 5% in writing with MedSupply and update the contract file, by end of day.
Open question: Does the current contract have an early-termination clause? Needs checking before next renewal cycle.
Parking lot: Q4 delivery schedule — separate conversation, not this call. Lagos trade show next month — Bola to send details if relevant.
Notice the shipping-delay story and the pleasantries appear nowhere in the right-hand column. That is not an oversight — it is the whole discipline. Ninety seconds of the call mattered. The rest was the social fabric that makes people willing to have the conversation at all, valuable in the room, worthless on paper afterward.
Write the four buckets as headings before the meeting starts, and drop lines under them as things happen — you will not remember the exact wording of a decision made forty minutes ago if you wait until the call ends to start writing. Then clean the raw version into the sent-out version within the hour, while the context is still fresh enough to fix anything ambiguous.
Cleaning up notes means removing what doesn't matter and sharpening the wording of what does — it does not mean softening a decision that was actually made, or turning a firm "no" into "we'll consider it" because that reads more diplomatically. Notes that quietly rewrite what happened are worse than no notes, because everyone believes them.
Find notes from any meeting you've sat in recently — a class, a call, anything with a written record. Re-sort them into the four buckets. Whatever's left over that doesn't fit any of the four is very likely the part nobody needed written down.
"We should look into whether the insurance portal can be set up for automatic renewals." Someone said this in a meeting, everyone nodded, and it will never happen — because it is not an action. It is a thought wearing the costume of an action. Nobody owns it, nothing is due, and "look into" doesn't describe a task anyone could actually finish and check off.
Talk becomes work only when it has a verb, an owner, and a date. Remove any one of the three and it quietly reverts to a thought.
Every line in the actions bucket from lesson 5.2 should be checkable against this shape before it's written down:
Notice the fourth and fifth lines have owners who were not you or Adaeze. That is deliberate — an action list only for your own side's tasks is half a record. If MedSupply's Tunde owes a confirmation, write it down with his name on it, same as anyone else's. It is what makes the follow-up in lesson 5.5 possible at all: you cannot chase a debt nobody wrote down.
People often agree to do something without naming when. Don't quietly assign a date yourself and hope nobody minds — ask, in the room if you can, or in the follow-up note if you can't: "By when should we expect the early-termination check?" A date supplied afterward, in writing, that nobody objects to, functions the same as one agreed live. A date you invented and nobody confirmed does not — it just moves the ambiguity from "when" to "did we actually agree to that."
The follow-up note — decisions, actions with owners and dates, open questions — should go out within an hour of the meeting ending, while everyone still remembers agreeing to it. Close it with one line: "Flag anything here that doesn't match your understanding — otherwise I'll take this as agreed." That single sentence turns a passive record into an active one: silence becomes confirmation, not ambiguity.
Everything else in this module — good agendas, clean notes — exists to make this step possible. A meeting can have a perfect agenda and beautifully sorted notes and still produce nothing, if the actions inside it stay vague. This is the step where the meeting's value actually gets banked.
Take the "everything said" transcript from lesson 5.2's worked example. Find the two implied actions buried in it that never made it into the right-hand column — one is the trade show, one is the early-termination clause. Write both out in full verb-owner-date form.
An automatic transcription or note-taking tool can join a call, produce a rolling transcript, and hand you a first draft of a summary before you've even finished typing your own notes. Used well, it is a genuine time-saver. Used as a replacement for paying attention, it will eventually put a wrong number in front of a client and your name will be on the email that sent it.
A transcript is a first draft, never a final record. It listened to the audio. It did not understand the meeting.
Automatic transcription is genuinely reliable at catching the general shape of a conversation — who spoke roughly when, the rough sequence of topics, and a searchable text version of a call you might otherwise have no record of at all. For a low-stakes internal check-in, the raw transcript is often good enough on its own. The failure modes below matter most exactly where the stakes are highest: numbers, names, and who-said-what in a negotiation.
Actually said (MedSupply's Tunde): "We could probably do 5% if you sign today rather than waiting for the renewal date."
Auto-transcript: "We could probably do 5% Tunday if you can wait for it, renewal date."
What went wrong: "sign today" became "5% Tunday" — the tool heard a name-shaped sound and guessed a name, and inverted the meaning of "rather than waiting" into "if you can wait." Read on its own, the transcript now says roughly the opposite of what was agreed.
What you send: The version from your own live notes, cross-checked against the transcript only for anything you weren't certain you'd heard right — never the raw transcript, pasted in.
Use the tool for what it's actually good at — a searchable backup, a way to catch something you missed live, a first pass at a rough summary you'll edit rather than send. Never paste its output directly into a client-facing note without reading every number, name, and disputed point against your own live notes first. Your live notes from lesson 5.2, imperfect and human, are the source of truth for anything with money or a decision attached. The transcript is a second opinion, not a substitute.
If you use a transcription tool on client calls, say so plainly and get consent before recording — this is a baseline courtesy, and in many contexts a legal requirement, not optional politeness. And be honest with Adaeze about what the tool guarantees: a rough draft, checked by you, not a verbatim legal record. Overselling its reliability is how a wrong number ends up believed simply because "the transcript said so."
If you have access to any transcription tool, record two minutes of yourself reading a made-up sentence with a number, a percentage, and an unusual name in it. Read the transcript back. Count the errors. This takes eight minutes and will change how much you trust the next one you see.
MedSupply's Tunde owed written confirmation of the 5% by Thursday. It's now Monday. You need that confirmation — without it, the contract renewal in three weeks is unprotected — but you also need Tunde to keep answering your emails for the next two years of this supplier relationship. Chase too softly and it never arrives. Chase too hard and you become the sender whose name makes people's stomach sink, which costs you every future ask.
The goal of a follow-up is never to make someone feel bad for being late. It's to make responding easier than continuing to ignore it.
The mistake most people make is picking one tone and repeating it — three increasingly polite emails, or three increasingly terse ones. Neither works, because each message should do a different job. Use a ladder that gets more specific and more visible at each step, never more emotional:
Over eight days chasing MedSupply's Tunde for that written 5% confirmation: Monday, a friendly one-liner assuming it slipped his mind. Wednesday, no reply — a specific ask naming the 24th deadline. Friday, still nothing — Adaeze gets cc'd, widening who sees it. Monday again, still silence — that stops being a message to Tunde at all, and becomes a note to Adaeze: what's outstanding, how long it's been, and a recommendation to call rather than email, with the early-termination notice as a fallback. Four fill-in-the-blank versions of this ladder are in the worksheet pack above.
Every chase should be visible in the record — sent from the account, dated, saved. If a relationship later sours and someone claims "nobody told us the deadline," a paper trail of four increasingly specific, entirely reasonable messages is the difference between that being their problem and it becoming yours. This is also, quietly, self-protection: you did your job at every step, and it is documented.
Step four exists for a reason: at a certain point, the decision about how hard to push, or whether to accept a missed deadline, stops being an assistant's judgment call and becomes the business owner's. The skill is not chasing forever — it's recognising the exact point where continuing to escalate on your own authority would be overstepping, and handing the decision back with enough context that Adaeze can make it in thirty seconds rather than needing to reconstruct the whole story.
This applies just as much to a client who owes Adaeze a signed form, or a colleague who owes you something for a shared project. The tone shifts slightly with the relationship, but the shape doesn't: assume good faith, get specific, widen visibility, then hand it up if it still isn't moving.
Think of something you're currently waiting on from someone — a reply, a document, a decision. Draft your own four-step ladder for it now, before you need step two. Having steps two through four written in advance is what stops step one's reminder from turning testy out of frustration.
You are given a rough 35-minute auto-generated transcript of a real call — the kind riddled with the exact errors lesson 5.4 named: a mangled supplier name spelled three different ways, one percentage that's plainly wrong given the rest of the conversation, two speakers merged into one during a stretch of overlapping talk, and a six-second silent gap where the transcript simply skips from one sentence to an unrelated one with no marker that anything was dropped. There was no agenda — this was a call booked last-minute.
Whether you can take a genuinely unreliable source — a machine transcript with real errors baked in — and produce a record someone could act on without re-listening to the call themselves. Every module's drill has tested judgement under a different kind of mess: an inbox, a calendar, a filing system, and now a conversation nobody wrote down properly. The tool changes. The discipline of sorting signal from noise doesn't.
You should now have: a four-part agenda template, a set of meeting notes sorted into the four buckets, an action list in verb-owner-date form, and a four-step chase ladder you've used at least once for real. Module 6 turns to research — turning a vague ask into a question with a stopping point, and building the client's trust that what you hand back is actually right.