Skip to main content
There’s a quiet pattern across most knowledge work. Meetings happen and get captured. The capture creates a growing archive — every customer call, every 1:1, every project meeting. Users invest in the capture; few invest in the retrieval. The result is an archive that’s mostly write-only. You put meetings in; you rarely pull them out. Meanwhile, the next meeting is starting in 15 minutes. You’re rushing between calls. You have a vague sense of the topic and the participants. The calendar description is sparse or empty. You walk in and spend the first three minutes reconstructing context: “Where were we last time?” “Did we ever resolve that thing about the integration?” “What did you commit to follow up on?” The meeting starts cold; momentum is rebuilt from scratch every time. This workflow inverts the direction of every other workflow in this guide series. Instead of capturing what happened in a meeting, it uses your captured archive to prepare you before the next one. The captured corpus stops being a write-only backstop and becomes an active operating substrate for every meeting you take. This guide is a specific instance of the workflows pattern, applied prospectively. It’s the systematic, scheduled-before-every-meeting use of the ad-hoc Q&A capability — the same query layer, with templates and a habit around it.

What works today and what doesn’t

Worth being precise about up front, because this workflow is more cross-meeting-dependent than any other in the series. Chat reads your calendar and your past meetings together. With your calendar connected, Chat looks up to two weeks ahead in Google Calendar or Outlook and pairs each upcoming event with related past meetings. Ask “Help me prepare for tomorrow using my past meetings” — or click the Prep for tomorrow suggestion on the Chat home page — and Chat walks through tomorrow’s calendar alongside what was said before, with source chips back to the calls it drew from. Multi-call briefings are one question. “What’s open across our last three calls with this customer?” is a plain question in Chat. For a renewal or a long relationship, say “all my meetings with Acme” or name the period. Every answer ends with a Coverage line — the date range searched and how many meetings were reviewed — so you know how complete the briefing is before you walk in. Meeting chat preps the meeting in front of you. Open the meeting in Earmark, click Chat in its header, and choose Help me prepare for this meeting. The chat window moves and resizes, so it can sit beside your call. Projects focus the corpus. If a customer, report, or meeting series has its own project, ask from the project page. Chat looks only at that project’s meetings, including ones teammates shared, and follows the project’s Instructions. The morning sweep takes a few minutes. One question covers the whole day’s calendar. Scan the results, then run a fuller template (below) on the two or three meetings that matter most. If you want to run your own analysis — a custom pass across every customer call in two years, say — exporting local transcripts and running an external agent is still an option. The rest of this guide is built around these mechanics: what to ask, which template to use, and the habit that makes it stick.

Foundation

Light setup compared to other workflows.

Which meetings warrant prep

Not every meeting needs a briefing. The meetings that actually benefit: The default: any meeting with a recurring counterparty where context from prior meetings would meaningfully change how the meeting goes. About 30–60% of most knowledge workers’ meetings.

When to prep

Briefings have a useful window. Too early and you’ll forget by meeting time; too late and you’re scrambling. The sweet spot: 5–30 minutes before the meeting. Practical patterns:
  • Between meetings. If you have back-to-back meetings, run the briefing as the previous one ends. One question; one minute to read; into the next meeting prepped.
  • Morning sweep. For meetings happening later in the day, do a sweep first thing in the morning. Briefings ready to glance at right before each meeting.
  • End-of-day prep for next morning. For the first one or two meetings of the next day, prep at end of today. The Prep for tomorrow suggestion on the Chat home page is built for this.
  • On demand. When meetings move or new ones drop in, run the briefing in the moment.

Organize the corpus for retrieval

The briefing queries your captured archive. Better organization produces sharper briefings.
  • Name meetings consistently. “Customer X — Weekly check-in” beats “Meeting.” Cmd+K uses titles to find relevant prior calls.
  • Group each relationship in a project. One project per customer, account, or meeting series. Add meetings at any stage — even before they happen — from the project chip next to the meeting title, or tick several on the Meetings page and choose Add to project. Prep asked from the project page then draws only on that relationship.
  • Keep identifiers consistent. Don’t switch between “Acme” and “Acme Corp” and “Acme, Inc.” — the model handles small variations but consistency helps.
The corpus doesn’t need to be perfect; it needs to be navigable.

Ownership

Each briefing has one owner: you, the person who’ll be in the meeting. There’s no shared ownership model because the briefing is for one person’s preparation. For coverage and handoff scenarios, the person being briefed is the consumer; the regular owner generates the briefing.

The four templates

Each template is a prompt you paste into Chat, or into a project’s chat, with the braces filled in. The Source material line tells Chat which past meetings to draw on; it finds them itself. Run one in the Composer instead and it works from that one meeting, which suits single-call briefings.

Template A — Account or customer briefing

For sales calls, CSM check-ins, customer review meetings, and any recurring conversation with an external party. The workhorse for customer-facing roles.
Three things in this template are load-bearing. “Open threads from prior calls” is the load-bearing section. It’s what makes you walk in knowing what to follow up on. Without it, you’d address the topics you remember; with it, you address the topics they remember too. Verbatim quotes are extracted. Reading what the customer actually said — not your interpretation — primes the meeting differently. You hear their voice; the meeting opens with their language. The “recommended opening” gives you a first sentence. “Last time you mentioned the integration timeline was tight — how is that going?” beats “How are things?” The recommended-opening prep is what makes the warmth feel intentional.

Template B — 1:1 / direct report briefing

For managers prepping for their next 1:1 with a report.
The “topics they raised across recent 1:1s” section is what makes the longitudinal pattern visible. A report who’s been raising the same concern in three consecutive 1:1s is signaling something; the briefing surfaces the pattern even if the manager hasn’t been tracking it explicitly.

Template C — Handoff / coverage briefing

For when you’re covering for a colleague (or being covered for). The audience is someone walking into a meeting cold; the briefing has to do the work of catching them up.
The “what to NOT commit to” section is what makes coverage briefings actually usable. The cover walks in armed to handle the meeting and clear about the limits of their authority — so they don’t accidentally commit to something the regular owner wouldn’t have. Who runs it depends on what Chat can see. Chat uses meetings you recorded plus meetings shared with you through a project. If the colleague has shared the relationship’s project with you, run Template C from that project yourself. If not, the colleague runs it and sends you the result.

Template D — Recurring meeting refresh

For ongoing internal meetings (weekly leadership sync, recurring project meeting, monthly cross-functional). The cadence is recurring; the briefing primes you on what’s recently happened.
The recurring-meeting briefing is what makes you the person who walks into the standing meeting with the most current state of the work. Most attendees walk in with their own slice; the briefing gives you the team’s slice.

Save them as private templates

1

Start with the template you'll use most

For sales and CS: Template A. For managers: Template B. For internal team leads: Template D. For anyone whose colleagues take PTO: Template C. Save the most-used first.
2

Save with Private visibility

Briefings are personal prep, not team standards. Private visibility is the right default. Templates can be exported and shared when standing up handoff scenarios. You can also ask Chat to find a saved template or turn a format into a new one; new templates stay private unless you ask to share them with the workspace. See Custom templates.
3

Practice once on a real meeting

Before relying on the workflow daily, run it once end-to-end on a real upcoming meeting: paste the template into Chat with the braces filled in, then read the Coverage line and hover a few source chips. If Chat reviewed fewer meetings than you expected, widen the date range or name the meetings, and note the phrasing for next time.

Running the workflow before a single meeting

1

Pick the template

Match to the meeting type. A for customer/account, B for 1:1s, C for coverage scenarios, D for recurring internal.
2

Ask Chat

Paste the template into Chat with the braces filled in — or into the project’s chat, if the relationship has a project. Chat finds the prior meetings itself and shows which ones it’s reading as it works. For long histories (renewal preparation across two years of customer history), say “all my meetings” or name the period.For a quick version, open the meeting, click Chat in its header, and choose Help me prepare for this meeting. For a single-call briefing, run the template in the Composer on the prior call.
3

Check the Coverage line

Did Chat review the meetings you expected? If the count looks low, widen the range or name the calls, and ask again. Hover the source chip on anything you plan to say out loud.
4

Scan, don't read

The briefing is a tool for priming, not a document for reading. Sixty seconds of scanning is usually enough. Headlines, bullets, verbatim quotes — close it, walk in.
5

Walk in armed

The meeting opens. Your first sentence references real prior context. The counterparty notices. The meeting starts warm.
For longer or high-stakes meetings, keep the briefing open on a second screen — not to read, but to reference if you blank on a specific item or commitment. Renewal calls and board sessions especially benefit from this.

Building the briefing habit

The workflow’s value depends entirely on consistency. A briefing run before half your meetings is dramatically less useful than briefings run before all of them, because the asymmetric advantage erodes when the practice is inconsistent. The trigger. For most users, the trigger is the calendar. Every recurring meeting on tomorrow’s calendar gets a briefing today; every meeting today gets a briefing in the 5–30 minute window before it. The morning sweep. A useful pattern for managers and executives with packed days: first thing, ask Chat to “help me prepare for today’s meetings using my past meetings.” It pairs every event on the day’s calendar with related past meetings in one pass. Run a full template on the two or three that matter most. Briefings sit ready; you scan each one in the minute before the meeting starts. The morning sweep prevents the failure mode of “I’ll prep right before the meeting” — which collapses when meetings run long and you’re walking in cold to the next one. The between-meeting rhythm. For users in back-to-back meetings, the briefing becomes part of the transition: meeting ends, ask Chat to “help me prepare for my next meeting,” scan the answer while walking to the next room, meeting starts. The 60-second transition is enough. The “skip prep” check. For meetings you’d normally skip prep for, ask: would 30 seconds of preparation make this meeting better? If yes, run the briefing. The exception is highly repetitive meetings where the prior context is in your head — daily standups with the same team, twice-weekly status checks. The compounding effect. A user running this workflow for a quarter accumulates noticeable effects: relationships feel deeper because they reference specifics from prior interactions; commitments carry forward reliably; meetings move faster because the cold-start cost is eliminated. By a year in, the user is operating at a different level than peers who run the same meetings without the workflow.

Coverage and handoff scenarios

The single most underrated application of this workflow. Coverage scenarios — when someone covers for a colleague, when a new team member onboards, when an account transitions — are usually disasters because context evaporates. Covering for a colleague on PTO. Day before PTO starts, the colleague generates Template C briefings for each meeting their cover will attend. Briefings get shared with the cover (Slack DM, email, shared doc). Cover scans each briefing 5 minutes before its meeting; walks in informed; counterparty barely notices the handoff. The investment from the colleague going on PTO: maybe 15 minutes generating briefings. The payoff for customer experience: substantial. Share the project, not only the briefing. If the colleague keeps the account’s meetings in a project, they can Share it with the cover as a Collaborator before leaving. The cover can then ask Chat in that project the questions the briefing didn’t anticipate — “what did we promise them about the SSO timeline?” — and get answers with short excerpts as sources. The cover sees each meeting’s title, who recorded it, and when, but can’t open the colleague’s meetings, and pins stay private. The same move works for account transitions and onboarding. Keep 1:1s and other sensitive meetings out of shared projects. Onboarding a new team member into ongoing situations. Every new team member inherits a portfolio of ongoing relationships. Historically the onboarding is a series of “let me bring you up to speed” conversations. With this workflow, briefings get generated for each ongoing relationship; the new team member reads them in their first few days; they walk into their first meetings already informed. This is one of the highest-leverage uses of the workflow — new hires who would historically be useful at month four become useful at week three. Account transitions. When an account moves from one CSM/AE to another, the briefing is the handoff. The outgoing rep generates Template A; the incoming rep reads it; the relationship continues with the customer feeling continuity. Add personal context the model can’t extract — “Lisa is hard to read but always engaged when we talk about data; Marcus is the real decision-maker even though Lisa is the contact”. Manager transitions. When a report’s manager changes — reorg, departure, promotion — the new manager inherits the relationship cold. The new manager reading the 1:1 briefings can pick up real threads on the first 1:1 — “I know you’ve been raising the integration timeline issue; how is that going?” The transition feels continuous rather than reset. Returning from your own leave. Apply the workflow to yourself when returning from extended PTO: what’s happened in the meeting series while I was out, what threads should I pick back up, what did the cover commit to that I’m now inheriting. A 30-minute “catch up on my own life” briefing session at the end of a vacation prevents the “I’m back but I have no idea what’s going on” first week.

What this workflow doesn’t do (and how to work around it)

The honest summary: this workflow is built on cross-meeting Q&A, and Chat now handles most of it directly — multi-call customer briefings, 1:1 briefings drawing on a report’s last six meetings, recurring-series refreshes, and calendar-aware prep for the next two weeks. The limits worth knowing:
  • Chat only sees what you can see. Meetings you recorded, plus meetings shared with you through projects. For coverage briefings on a colleague’s relationship, the colleague shares the project or runs the briefing.
  • Older history needs to be asked for. Say “all my meetings with Acme,” or the period you mean. Check the Coverage line to confirm.
  • Calendar lookahead needs a connected calendar, and reaches up to two weeks ahead. See Calendar.
  • Temporary meetings aren’t saved, so they can’t feed a later briefing.
  • Anything that didn’t happen in a meeting isn’t there. Slack threads, emails, and hallway conversations only show up if someone said them out loud on a recorded call.
Finding a specific prior meeting by title or keyword is still a job for Cmd+K. If you want to run your own analysis (a custom pass across a multi-year account, say), exporting local transcripts and running an external agent is still an option.

Common pitfalls

  • Skipping the briefing on busy days. The exact days when briefings are most valuable are the days when you’re rushed enough to skip them. The 60-second cost is worth paying.
  • Trying to read briefings in full. They’re scan documents. Headlines, bullets, verbatim quotes; close it; walk in.
  • Treating the briefing as a script. The briefing primes you; it doesn’t script the meeting.
  • Generating briefings too early. A briefing generated two days before the meeting may be slightly stale. Generate close to the meeting.
  • Pasting raw briefing content into the meeting context. Briefings often contain candid notes — “watch for: customer was frustrated about pricing.” Don’t paste these as-is into the meeting room or share with the counterparty.
  • Assuming the briefing is complete. The briefing reflects what was captured. If important context lived in Slack threads or emails Earmark didn’t see, the briefing won’t include it. And read the Coverage line: a briefing built from two meetings when you’ve had eight is missing most of the story.
  • Generating briefings for every meeting. Some meetings don’t benefit. Standups, all-hands, and first-time calls don’t need them.
  • Inventing context. The model can fabricate plausible-sounding prior context that didn’t actually happen. The source check is your defense — if a claim has a source chip, hover it and read the excerpt; if the briefing asserts a pattern without specifics or sources, scrutinize.
  • Not using briefings during handoffs and coverage. The single largest missed opportunity. Coverage briefings should be standard practice.
  • Letting briefings replace relationship-building. The briefing primes you; it doesn’t substitute for being attentive in the meeting itself.
  • Reading the briefing during the meeting. Visible reading reads as inattention. Read before; reference its content naturally.

Where to go next

  • Workflows — the general shape this is an instance of
  • Chat — calendar-aware prep, meeting chat, sources, and the Coverage line
  • Projects — grouping a relationship’s meetings, and sharing them for coverage and handoffs
  • Ad-hoc Q&A — the underlying capability this workflow systematizes; covers asking across meetings and verifying answers
  • Command menu — for finding a specific prior meeting by title or keyword
  • Local transcripts — if you want to run your own analysis
  • People and team meetings workflow — for the 1:1 capture artifacts that feed Template B briefings
  • Sales calls workflow — for the deal context that feeds Template A briefings on customer accounts
  • Customer onboarding and training workflow — for implementation contexts where briefings draw on session histories
  • Custom templates — visibility, sharing, and edit permissions
  • Composer — for single-call briefings run on one prior meeting
  • Before a meeting — the per-meeting workflow this prep workflow precedes