How to Prep for Meetings Automatically Every Time

Most meetings start with about four minutes of reconstruction. Who is this person again. What did we agree last time. Did I ever send that thing. Everyone in the call is doing the same archaeology at the same time, and the first quarter of the meeting evaporates before anyone says anything new.

The information you need already exists. It is in the last email thread, the previous meeting's notes, the customer record, and the task you marked done in March. Nothing needs to be created. It needs to be gathered and put in front of you at the right moment, which is precisely the kind of work that should never involve a human.

This is the whole job, described so you can build it with whatever tools you already have.

How do you prep for meetings automatically?

Four stages, in order: trigger, gather, assemble, deliver. The calendar event is the trigger, because it is the only thing that reliably knows a meeting is coming and who will be in it. Gathering pulls the history from your email, notes, and customer records. Assembling forces that pile into a fixed template. Delivery puts the finished brief somewhere you will read it, at a time when reading it still changes something.

Skip the template stage and you get a data dump nobody reads. Skip the delivery stage and you get a beautiful brief sitting in a folder. Both failures are common, and both feel like the automation is working right up until the moment you need it.

The four stages of automatic prep

1. Trigger from the calendar — The event fires the process. Attendee email addresses are the key that unlocks everything else, which is why an event with no guests listed produces a thin brief.
2. Gather from your real sources — Recent email with each attendee, notes from the last meeting with them, their record in your CRM, open tasks either way, and any document attached to the event.
3. Assemble into one fixed shape — Same five sections every time. Familiar structure means you can read the brief in thirty seconds instead of parsing a new layout under time pressure.
4. Deliver at a useful moment — Prep-heavy meetings need the brief the evening before. Routine calls need it shortly beforehand. Both, if the meeting is genuinely important.

What belongs in an automatic meeting brief?

Five sections, and a hard limit of one screen. Who you are meeting and why they matter. What happened last time you spoke. What each side owes the other, including anything overdue. What documents are relevant. And the one outcome this meeting needs to produce. Anything past that is research you can do during the call if it comes up.

The section people leave out and then miss most is open commitments in both directions. Walking into a call having forgotten that you promised a revised proposal eleven days ago is the single most avoidable way to lose ground with a customer, and it is exactly the kind of thing a machine never forgets.

Where does the brief actually get its facts?

The quality of the brief is decided entirely by what it can read. A tool connected only to your meeting recordings can tell you what happened in the last call and nothing else. A tool connected to email, calendar, notes, and your customer records can tell you the whole story. Neither is smarter than the other, one just has better access.

That trade-off is visible across the products in this space. Fellow's pre-meeting brief, for example, summarizes the previous meeting and its follow-up points and is generated from that meeting's recording. You enable it under User settings, then Note Taker, toggle on the pre-meeting brief, and choose a delivery window anywhere from 15 minutes to 3 days ahead. It is available on their Business and Enterprise plans. That is a genuinely useful brief with a clearly drawn boundary: it knows your meetings, not your inbox.

So the first decision is not which tool. It is which sources you are willing to connect. If you only connect the calendar, expect a brief that lists attendees. If you connect email and notes as well, expect a brief that changes how the meeting goes.

Working across all of those sources at once is what a personal AI assistant is for. It reads the calendar, finds the thread with each attendee, checks what is still open between you, and writes the brief in the same five sections every morning, without you configuring a rule per meeting type.

When should the brief arrive?

Two moments, and they do different jobs. The night before, or first thing in the morning, is when a brief can still change your behaviour: you can send the thing you forgot, pull a number, or cancel a meeting that clearly has no purpose. Shortly before the call is when a brief works as a memory jog and nothing more.

For most people the right answer is a single morning digest covering the whole day, plus a short context note before each external call. One digest is a habit you can keep. Fifteen separate briefs are fifteen notifications you will start ignoring by Thursday.

If you want to see what that feels like before committing to a setup, the fastest route is to connect a calendar and an inbox to an AI Employee and ask for tomorrow's briefing. You can start with Sistava and have the first digest waiting the next morning.

How do you make briefs people actually read?

Length is the whole battle. A brief that takes longer to read than the prep it replaces has failed at its only job. Cap it at one screen, put the most surprising fact first, and end with a single line naming what you should do before the call starts.

• One screen, always. If it scrolls, something in it is not earning its place. Detail belongs behind a link, not in the brief.
• Lead with what changed. They replied at midnight, the deal moved stage, the invoice is overdue. Stable facts go lower down.
• Name the one action. Send the timeline before the call. A brief with no action is trivia about a person you are about to meet.
• Same shape every time. Familiar structure lets you skim in thirty seconds. Novelty in a brief is a cost, not a feature.
• Flag what it could not find. A brief that quietly omits a source it failed to read will make you confident about something wrong.

What automatic meeting prep will not do

It will not know what you are trying to achieve. A brief can tell you everything that has happened and still be silent on the thing that matters, because your objective usually lives in your head rather than in any system. Write one line of intent on the event and the brief becomes twice as useful, which is a small manual step worth keeping.

• It cannot read what it is not connected to. Context living in a personal notebook, a WhatsApp thread, or a colleague's memory stays invisible.
• It struggles with brand new contacts. First meetings have no history, so the brief becomes public research and can only be as good as what is published.
• It can be confidently wrong about people. Matching a name to the wrong company record produces a brief that reads perfectly and misleads you completely, so check the identity line first.
• It does not replace thinking about the meeting. Knowing the history is not the same as knowing what you want, and only one of those wins the call.
• Access has a cost. A prep assistant that reads your inbox and customer records needs the same care about permissions as any person you would give that access to.

Start with your external calls only. They are where preparation changes the outcome most, and where the embarrassment of forgetting something is expensive. Internal meetings can wait until the template has proved itself.

One habit makes every brief better and takes ten seconds. When you book a meeting, write a single sentence in the event description saying what it is for. That sentence is the only piece of the brief no automation can source for you, and it turns a summary of the past into a plan for the next thirty minutes.

Prep, notes, and follow-through are one loop rather than three separate purchases. Whatever you use for the brief should also be the thing that writes the recap, because the recap is next week's prep. That is the argument for handing the whole meeting lifecycle to a single AI Employee instead of three tools that each know a third of the story.

• Fellow Help Center: meeting briefs