# How to prepare for a sales meeting with AI

> What a sales meeting brief should hold, where homemade prep breaks, how Day AI briefs every meeting, plus a copyable template, a section per attendee, timing, and first-meeting vs renewal examples.

Updated October 19, 2026

**The short answer.** A good meeting brief pulls account history, stakeholders, past commitments and open risks into one page before the call. To get one from AI, give the model the calendar invite, the CRM record, recent call transcripts, email threads and Slack discussion about the account, and ask for a fixed template: the objective, who is in the room and what each person has said before, open commitments both ways, deal stage and risks, suggested questions, and a source after every line. To get one before every meeting, the brief has to run on its own against each rep's calendar and land where they already look. That last part is where most homemade prep stops working.

This guide is about the document a rep reads before the call: what goes in it, where building it yourself breaks, how Day AI produces it, and a template and checks you can use either way. Writing back to the CRM after the call is covered in [How to update your CRM from call recordings](/guides/update-your-crm-from-call-recordings).

The prep itself is moving to AI. Andreessen Horowitz describes the shift this way: "Prep that used to happen case by case, if it happened at all, now happens every time as a matter of course" ([a16z](https://a16z.com/from-system-of-record-to-system-of-intelligence/)). Vantage Point lists "meeting prep and account research" among the jobs moving to the AI layer, where "assistants synthesize records and history instantly" ([Vantage Point](https://vantagepoint.io/blog/ai/crm-ai-engagement-layer-three-years)).

---

## What a strong meeting brief contains

Outreach's definition is a good baseline: "A strong meeting brief includes account background, stakeholder roles, deal stage, risks, goals, and recommended talking points to guide the conversation" ([Outreach](https://www.outreach.ai/ai-agents/meeting-prep-agent)). Gong's pre-meeting summary has a similar shape: meeting objectives, key topics to be discussed with prior discussion highlights and action items for you, suggested questions to ask, and general reminders ([Gong](https://help.gong.io/v1/docs/get-ready-for-meetings-with-ai-powered-meeting-prep)).

Each part needs an input, and no single system holds all of them:

| Brief section | What it answers | Input it needs |
|---|---|---|
| Account background | Who the company is, how long you've worked together, what changed lately | CRM, email |
| Stakeholder roles | Who is attending, their title, their part in the deal | Calendar invite, CRM contacts, past meetings |
| What each person has said | Needs, objections and competitors, in their words | Call transcripts, email |
| Commitments | What each side promised, and whether it happened | Calls, email |
| Deal stage and risks | Stage, amount, next step, silent stakeholders, slipped dates | CRM, calls, Slack |
| Goals and talking points | What a good outcome looks like, what to raise | The rep, informed by all of the above |
| Suggested questions | What's still unknown | Gaps in all of the above |

A brief built from the CRM alone fills the first and fifth rows. The rows that make a rep sound prepared, what each person said and what was promised, live in transcripts and email.

## How to build it yourself

Start by hand: paste the template below into your assistant with the material for one meeting. Once the output is good, schedule it:

1. **Read each rep's calendar** for the day through the Google Calendar API or Microsoft Graph.
2. **Keep external meetings.** Drop meetings where every attendee is on your own domain.
3. **Match each attendee** to a person, company and deal by email address. When there's no match, mark the attendee as a first contact.
4. **Gather history per attendee and per account:** CRM fields, recent transcripts from your call recorder, email threads with that domain, and Slack discussion about the account.
5. **Fill the template,** one model call per meeting, with the source rule in the prompt.
6. **Deliver it** on a fixed schedule to a fixed place.
7. **Log every run:** the meeting, the sources used and the brief produced, so you can trace a bad brief later.

If you work in Claude Code, these steps fit in a skill with your CRM, recorder and email connected over MCP; [How to give Claude Code memory of every customer conversation](/guides/give-claude-code-customer-memory) covers that setup. If your calls are in Gong and your CRM is HubSpot or Salesforce, [Gong calls into HubSpot or Salesforce](/guides/gong-calls-into-hubspot-or-salesforce) covers getting the transcripts across. Step 3 is where a context graph helps most; [How to build your own context graph](/guides/build-a-context-graph) covers how that works.

---

## Where DIY meeting prep breaks

A homemade brief works for one rep with one inbox. These are the places it breaks as more of the team relies on it.

### A prompt over the CRM alone misses the conversation

The CRM holds stage, amount and whatever a rep typed in. What the customer said on the last call and what your team promised in email aren't there. A brief built only on CRM fields reads like a record summary and misses the lines that matter in the meeting.

### The context is scattered across tools

Calls sit in the recorder, email in each person's inbox, discussion in Slack, and deal fields in the CRM. Each brief joins them again, and each join can fail: a recorder export that runs nightly misses this morning's call, and an inbox the script can't read leaves out a teammate's thread. Nothing tells the rep what's missing.

### Your teammates' context is out of reach, or out of control

The history that makes a renewal brief useful is rarely all in one person's inbox. The account executive ran the pricing thread, the customer success manager recorded the quarterly review, and a solutions engineer answered the security questions. A prompt you run over your own Gmail and your own calls sees none of that.

Getting it means reading other people's mail and recordings: everyone's OAuth tokens, or a domain-wide service account that can read every mailbox. Then you need rules for what each brief may quote, and you have to apply them every time. Most homemade setups skip that step. Pooling everyone's mail into one store with one key fixes coverage and creates a different problem: every brief can now quote any thread, including a manager's compensation email or a thread nobody meant to share.

### Briefs are generic when they aren't tied to each attendee

Matching attendees to people is harder than it looks. People use personal addresses, change jobs, and show up in the recorder under a nickname or a speaker label. When matching fails, the brief falls back to the account level and can't say who said what. When it matches the wrong person, one customer's history lands in another customer's brief.

### No source trail, no trust

If the brief doesn't cite its sources, the rep can't check a line, and after one wrong line in front of a customer they stop reading briefs. When a prompt drifts or a field changes, briefs get worse without failing, and without a log of runs and sources nobody can tell why.

### Prep that depends on someone remembering to run it

A prompt in a chat window produces a brief when a rep remembers to open it, which is the busiest week of the quarter least often. A script on someone's laptop stops when the laptop is closed or the token expires. Prep that happens "when there's time" is the case-by-case prep the a16z quote describes, not prep that happens every time.

None of this means you can't build it. Teams do. It means that after the first good brief, most of the work is plumbing, permissions, scheduling and checking, and the prompt is the small part.

---

## How Day AI handles it

Day AI provides the context graph and agent control plane to quickly and securely upgrade your company brain. Teams also describe it as an AI-native CRM: it's what comes after the CRM, running on top of Salesforce or HubSpot or in place of them.

For meeting prep, that means the brief is built on a self-updating customer record from every meeting, email and call, including what your colleagues have shared, by governed agents that prep every meeting, with every fact traced to source conversations and each person's sharing choices applied.

### A briefing every weekday morning, without anyone running it

When you set up an agent, Day AI gives it a **Morning Briefing** skill. It runs at 9am on weekdays in your time zone and arrives by email. By default it:

- **Lists today's meetings** where you're on the invite, and leaves out any you declined. Because invites include people who never replied, it presents them as meetings on your calendar, not confirmed attendance.
- **Preps each external meeting.** For meetings with people outside your company, it researches the attendees, their companies, and the whole workspace's recent interactions with them, not only yours: recent meetings, emails, Slack messages, CRM notes and open opportunities.
- **Lists internal meetings briefly,** without deep prep.
- **Ends with your new action items:** those assigned to you in the last 24 hours.
- **Adds what you care about.** It reads what your agent has saved about your interests and priorities, and surfaces the last 24 hours of activity that matches.

Nobody has to remember to run it. Skills work asynchronously: the agent does the research in the background, whether or not the rep has Day AI open, and sends the finished brief to their inbox or Slack.

### Your colleagues' shared emails and meetings, with permissions built in

A Day AI brief isn't limited to your own inbox. It draws on your email and meetings plus everything your colleagues have shared with the workspace: the account executive's pricing thread, the customer success manager's recorded quarterly review, the solutions engineer's security answers. For a renewal or a handoff, that's usually most of the history.

Sharing is set by each person, not by whoever builds the brief:

- **Meetings.** Each person sets a default for internal and external meetings, 1:1 and group, as **Invite only** (just them, anyone who also recorded it, and people they invite) or **Workspace**. The setting covers the recording's summary, notes and transcript together, and any single meeting can be changed.
- **Email.** Each person's email is **Private** (only them and their agent) or **Workspace** by default, with rules by domain, address or Gmail label on top. Sharing everything with a customer's domain while keeping a recruiting label private is one rule each.
- **Nothing is shared until they choose.** A new account starts with meetings Invite only and email Private. Contacts and companies are always shared.

Day AI applies those choices as data arrives. It decides who can see each email thread before any person or agent can read it, and when a message sits in several inboxes, a private rule from any of them keeps it private. For recorded meetings, the most private setting among attendees wins. Agents inherit their person's permissions, so each rep's brief includes what colleagues shared and that rep's own data, and nothing else. Each skill also has a **Private Data Access** setting: off, it uses only workspace-shared data; on, it also uses your own private email and recordings.

That's the difference from a homemade setup: nobody writes or maintains the permission rules, and nobody has to choose between briefs that miss the team's history and briefs that can quote anything. The [trust page](/trust) covers security and data handling.

### The web for what the record doesn't have

Day AI agents on every paid tier can search the web, in chat and inside skills. The agent runs a search, opens the most promising pages, and answers from what they actually say, citing the links it used. It can limit results to recent pages for news, and it reads a specific public page or PDF when you link to it.

For prep, that fills the gaps the customer record can't: a company's recent news, a new attendee's background, a press release or an earnings report the buyer mentioned. It matters most for a first meeting, when there's no call or email history yet. Pages behind a login or paywall can't be read, and the agent says so instead of guessing.

### Built on calls, email and Slack, not only CRM fields

The briefing reads the same graph every agent reads:

- **Meetings** from Day AI's notetaker, which writes each meeting into the graph when it ends, and from [Gong](/gong) and Granola, which connect with their full history and come in with their transcripts.
- **Email and calendar** from each person's own Gmail and Google Calendar. A new person's first sync takes one to six hours. Email comes from Google Workspace today; Outlook mail doesn't sync into the graph, and agents can reach Outlook, Outlook Calendar and Teams live through connectors in beta.
- **Slack** from the moment the Day AI app joins a channel.
- **Your CRM,** because Day AI supports connecting MCP servers. Agents read from any server you connect and write to it where the server allows. Connect the Salesforce or HubSpot MCP for two-way sync.

Each person, company and deal has one record with every source resolved to it, so the briefing starts from an attendee and follows the links to their company, deals and past conversations, instead of joining four tools on every run. [Customer memory](/customer-memory) shows how the record is built.

### Every line can point to its source

Answers cite the call clip or the message they came from, and every value keeps its history: who changed it, person or agent, when, why, and what it was derived from. That's the source trail rule described below, built into the data. If you want the citation after every line of the briefing, put that in its instructions.

### Build the brief you want with your agent

The Morning Briefing is a default, not the only option. Describe the brief you want to your agent in plain words and it sets it up as a skill: what goes in, in what order, when it runs and where it lands. For example:

> "Every weekday at 7am, brief me on each external meeting today. For every attendee, give me their role, when we last spoke, what they've said, and what changed. Then list open commitments both ways, deal stage and risks, three questions, and the call or email behind every line. Send it to Slack."

What you can shape:

- **The structure.** Put the template below in a Page and attach it to the skill under **Attached Context**, and the agent reads it every time the skill runs. Attach your discovery questions, qualification criteria or renewal playbook the same way.
- **The schedule.** Daily, weekdays or weekly at the time you pick, so a brief can land first thing or the evening before. Executive and Super agents also get an every-4-hours cadence, for a midday refresh.
- **The research.** Tell it to search the web for each company's recent news, or to look up attendees you've never met.
- **The delivery.** Email or Slack. The skill does the work on its own and sends you the result.
- **On demand.** Every skill has a slash command, so a rep can run it from chat right before a call. Asking "prep me for my 2pm" works too, because the agent picks the matching skill when the request fits what it's for. With the Day AI app in a Slack channel, skills also run from Slack slash commands.
- **The default too.** To adjust the Morning Briefing itself, tell your agent what to change, such as "put open commitments first for every external meeting", and it updates the instructions. You can also move the delivery time or switch delivery to Slack.

Scheduled and event-triggered skills count toward each agent's automated-skill limit: 2 on Turbo, 5 on Professional, 10 on Executive and Super. Every skill keeps a history of its runs. When a brief isn't what you want, you tell the agent, it shows you the skill definition, you change it, and every later run uses the change.

One limit to plan around: skills run on a schedule, on events such as a recording being ready or an email landing, or when someone invokes them. None of those is tied to a meeting's start time, so for a last look right before a call, the rep runs the skill or asks the agent.

### It works with what you've built

If you prep from Claude Code, keep your repo and skills. The graph is available over MCP at `https://day.ai/api/mcp`, and [day.ai/mcp](/mcp) covers setup for Claude, Claude Code and Cursor. MCP access starts at the Turbo tier, some tools are reserved for higher tiers, and there's no usage-based pricing. In a workspace with at least one paid seat, teammates on free seats fill the graph with their meetings, email and calendar. Free seats don't connect over MCP. Plans are at [day.ai/pricing](/pricing).

After the call, the same record can draft the follow-up. That's a separate job from the brief.

---

## A meeting brief template you can copy

Use the same headings every time, so reps know where to look and you can tell when a section comes back empty. This works pasted into any assistant, and as a Page attached to a Day AI skill.

```markdown
# {Meeting title}, {date and time}

## Objective
One sentence: what a good outcome of this meeting looks like.

## Who's in the room
For each external attendee:
- **{Name}, {title}**, {role in the deal}. {First meeting, or last spoke on <date>.}
  - Said before: "{quote or close paraphrase}" ({source})
  - Cares about: {priority} ({source})
Our side: {names and roles}

## Open commitments
- We owe them: {item}, promised {date} ({source}). Status: {sent / not sent}.
- They owe us: {item}, promised {date} ({source}). Status: {received / not received}.

## Deal stage and risks
- Stage: {stage}, amount {amount}, next step {next step} ({CRM field})
- Risk: {risk} ({source})

## Suggested questions
1. {question that closes a gap above}
2. {question}
3. {question}

## Sources
- {Call, date}
- {Email thread, subject, date}
- {CRM record}
```

To generate it by hand, paste the template into your assistant with the material and three instructions: use only the material below, cite the source after every line, and write "not in the material" where a section has no answer. The last instruction keeps the model from filling gaps with plausible guesses.

## Write one section per attendee

A brief that describes "the account" treats a five-person meeting as one voice. The CFO who asked about payment terms and the engineer who asked about SSO need different answers, and the rep needs to know which person said what.

Outreach describes doing this with "conversation intelligence with speaker separation to identify each stakeholder individually, capturing their unique pain points, objections, and priorities from past interactions" ([Outreach](https://www.outreach.ai/ai-agents/meeting-prep-agent)). That needs transcripts with speakers mapped to names, then a lookup of everything each attendee said across past calls and threads.

Each attendee section answers the same four questions, in this order:

1. **Who are they in this deal?** Name, title, and role: economic buyer, champion, technical evaluator, end user, or unknown.
2. **When did we last hear from them?** The date and the call or thread, or "first meeting".
3. **What have they said?** One to three lines in their words, each with a source. Prefer objections and asks over pleasantries.
4. **What changed?** New to the invite, missing from recent calls, new title, or added by someone else.

An example, for a fictional account:

> **Priya Shah, VP Finance**, economic buyer. Last spoke on the 3 September pricing call.
> - Said before: "We can't sign a multi-year term until we see the security review." (Call, 3 Sep)
> - Asked for a payment schedule split by quarter. (Email, "Order form questions", 5 Sep)
> - Not on the last two calls. Joining again after three weeks away. (Calendar)
>
> **Marco Ruiz, IT Director**, technical evaluator. First meeting.
> - Added to the invite yesterday by Priya. (Calendar)
> - No prior calls or emails with us. (Search of calls and email)
> - Leads identity and access at the company, per their public team page. (Web, company site)

Marco is the reason to write sections per person. A new name on the invite, added by the buyer the day before, is often the most important line in the brief, and "what changed" is the question that catches it.

## The source trail rule

Every line in the brief points to the call or email it came from: a call and its date, an email subject, a CRM field. No source, no line.

The rule does three things:

- **The rep can check before repeating.** A price, a date or a promise said back to a customer has to be right. With a source, checking takes seconds.
- **Gaps become visible.** A section with no sources is a section with no evidence, and the brief says so plainly.
- **You can debug bad briefs.** When a line is wrong, the source shows whether the model misread a good input or got a bad one.

Put sources inline, after each line, and list them again at the end, so a rep can open last week's call directly from the brief.

## When and where the brief should land

A brief has to arrive early enough to read and late enough to include yesterday's call.

**When.** Gong shows its pre-meeting prep on a customer call that hasn't started or is in progress, marked "ON AIR" from 15 minutes before start, and its mobile app sends "a push notification 30 minutes before each external meeting" ([Gong](https://help.gong.io/v1/docs/get-ready-for-meetings-with-ai-powered-meeting-prep)). Those are short windows, good for a final read. Day AI's default Morning Briefing takes the other approach: every meeting for the day, in one message, at 9am, and a brief skill you build yourself can run on another schedule or on demand right before a call. Many reps want both: the day's meetings first thing, then a last look shortly before each call.

**Where.** Deliver it where the rep already looks:

- **Inbox:** a morning email listing the day's briefs. This is where Day AI's Morning Briefing lands by default.
- **Slack:** a message with the brief or a link to it. Day AI skills can deliver to Slack instead of email.
- **Calendar:** in a homemade setup, the brief or a link to it in the event description, so it opens with the meeting.

Whatever you choose, use the same place every time. A brief a rep has to hunt for doesn't get read.

## The 5-minute self-check

Before the call, read the brief and ask one question: **could I walk in without opening another tab?**

If not, the brief is missing something. Check these in five minutes:

- I know the objective in one sentence.
- I know every external attendee's role, and which of them are new.
- For each returning attendee, I know what they last said and where.
- I know what we promised them and whether we did it, and what they promised us.
- I know the deal stage, the next step and the biggest risk.
- I have at least two questions I can't answer from the brief.
- Every line I plan to repeat has a source I could open.

Each unchecked box points to a gap in the inputs, not the prompt. A missing quote means the transcript wasn't in the material. A missing commitment usually means the email thread wasn't.

## Two briefs side by side: a renewal call and a first meeting

The template stays the same. What fills it changes with how much history exists.

| Section | First meeting | Renewal call |
|---|---|---|
| Objective | Qualify: is there a problem worth solving, and who decides | Renew and expand: confirm value, surface risk early |
| Who's in the room | Titles and public research; how the meeting was booked | What each person said over the last year; who has gone quiet; who's new |
| What they've said | Only the inbound form, the booking email or a referral note | Quotes from past calls and threads, by person |
| Open commitments | Usually none; note anything promised while booking | Every open promise both ways, with dates |
| Deal stage and risks | Stage is "new"; risk is fit and urgency | Usage, open support issues, a champion who left, a competitor named |
| Suggested questions | Discovery: current process, impact, timeline | Gaps: what changed since the last renewal, who else should be in the room |
| Sources | Few; the brief says so | Many; the brief cites the most recent first |

Here are both, filled in for fictional accounts.

**Renewal call brief**

```markdown
# Northwind Freight renewal review, Thu 10:00

## Objective
Confirm the renewal at the current seat count and agree on a date for the order form.

## Who's in the room
- **Dana Okafor, Head of Sales Ops**, champion. Last spoke 12 Sep.
  - Said before: "Renewal is fine, I just need the usage numbers for my CFO." (Call, 12 Sep)
- **Sam Lee, Procurement**, new to the invite. No prior calls or emails. (Calendar)
Our side: account manager, solutions engineer

## Open commitments
- We owe them: usage report by team, promised 12 Sep (Call, 12 Sep). Status: not sent.
- They owe us: list of seats to remove, promised 19 Sep (Email, "Seat review", 15 Sep). Status: not received.

## Deal stage and risks
- Stage: renewal, contract ends 31 Oct (CRM)
- Risk: procurement joined late, and our usage report is overdue (Calendar; Call, 12 Sep)

## Suggested questions
1. What does procurement need from us to sign by 31 Oct?
2. Which seats are being removed, and why?

## Sources
- Call, 12 Sep; Email, "Seat review", 15 Sep; CRM renewal record; Calendar invite
```

**First-meeting brief**

```markdown
# Intro call with Lakeside Dental Group, Thu 14:00

## Objective
Find out whether they have a problem we solve, who decides, and by when.

## Who's in the room
- **Ana Brooks, COO**. First meeting. Booked through the website form on 14 Oct.
  - Form said: "Our front desk spends hours on reminder calls." (Inbound form, 14 Oct)
  - Company announced two new locations last month. (Web, company news page)
- No other attendees.
Our side: account executive

## Open commitments
- None. History is thin: one form submission, no calls or emails. (Search of calls and email)

## Deal stage and risks
- Stage: new (CRM)
- Risk: unknown budget and decision process.

## Suggested questions
1. How do reminders work today, and who does them?
2. What happens when a patient misses an appointment?
3. Who else would be involved in choosing a tool?

## Sources
- Inbound form, 14 Oct; company news page; CRM lead record; Calendar invite
```

The renewal brief is mostly evidence: quotes, dates and overdue promises. The first-meeting brief is mostly questions, and it says plainly that history is thin. That's the right behavior. Gong, for example, generates its pre-meeting summary and Ask anything "only when there's prior tracked activity with the account, such as calls or emails" ([Gong](https://help.gong.io/v1/docs/get-ready-for-meetings-with-ai-powered-meeting-prep)). Whatever produces your briefs, make sure a meeting with no history still gets a brief, built on how the meeting was booked, what's public about the company and attendee, and discovery questions. In Day AI, the agent can run that public research itself with web search, citing the pages it read the same way a renewal brief cites calls.

## Frequently asked questions

### What's the best way to prepare for a sales meeting with AI?

Give the AI your own history with the account and ask for a one-page brief. The inputs are the calendar invite, the CRM record, the last few call transcripts and email threads, and any Slack discussion about the account. Ask for the objective, who is in the room and what each person has said before, open commitments on both sides, deal stage and risks, and suggested questions, with a source after every line. Then run one test before the call: could you walk in without opening another tab?

### How do I set up automatic meeting prep briefs for my reps?

Run the same brief on a schedule. A job reads each rep's calendar, keeps meetings with outside attendees, matches each attendee to a person, company and deal, gathers their calls, email and CRM fields, writes the brief from a fixed template, and delivers it where the rep already looks: their inbox, Slack, or the calendar event. Gong, for comparison, shows its pre-meeting summary on a customer call from 15 minutes before it starts, and its mobile app sends a push notification 30 minutes before each external meeting. In Day AI, each agent is set up with a Morning Briefing skill that covers the day's meetings every weekday morning without anyone running it, and reps can describe their own brief to their agent to get it as a scheduled or on-demand skill.

### How do I use a context graph for meeting prep?

Start the brief from each attendee's email address. In a context graph, that address resolves to one person, already linked to their company, open deals, and every meeting and thread they were part of, with a source on every fact. The brief can then quote what each attendee said, list what was promised and point to the call or email behind each line, without joining CRM, recorder and email results on every run.

### Does Day AI prepare meeting briefs automatically?

Yes. When you set up an agent, Day AI gives it a Morning Briefing skill that runs at 9am on weekdays in your time zone and arrives by email. It lists the day's meetings you're invited to, leaves out ones you declined, and for meetings with people outside your company it researches the attendees, their companies and the whole workspace's recent meetings, email, Slack messages, CRM notes and open opportunities with them. It ends with action items assigned to you in the last 24 hours. You can change its instructions, delivery time and channel (email or Slack) by telling your agent. Or describe the brief you want and your agent sets it up as its own skill, with your template attached as a Page, on the schedule you choose or run on demand with a slash command before a call. Agents on paid tiers can also search the web, so a brief can include a company's recent news or a new attendee's background, and the skill works in the background and sends you the finished brief by email or Slack.

### Can a meeting brief include my teammates' emails and calls?

Only if the system behind it handles permissions. A prompt over your own inbox can't see a colleague's pricing thread or recorded quarterly review, and pooling everyone's mail under one key lets every brief quote any thread. In Day AI, each person chooses what to share with the workspace: meetings as Invite only or Workspace by type, email as Private or Workspace with rules by domain, address or Gmail label, and nothing is shared until they choose. Visibility is decided as data arrives, and agents inherit their person's permissions, so each rep's brief includes what colleagues shared plus that rep's own data, and nothing else.

### What should a sales meeting brief include?

An objective for the meeting; who is in the room and what each person has said before; open commitments in both directions; deal stage and risks; suggested questions; and the sources behind every line. Outreach describes a strong brief as including account background, stakeholder roles, deal stage, risks, goals, and recommended talking points. Keep it to one page.

### When should a meeting brief arrive?

Early enough to read and close enough to the call to be current. Gong makes its pre-meeting summary available from 15 minutes before a customer call and pushes a mobile notification 30 minutes before each external meeting. Many teams send a morning digest of the day's meetings, then a short reminder before each call, delivered to the inbox, Slack or the calendar event.

### How is a first-meeting brief different from a renewal brief?

A renewal brief is built from history: what the customer said on past calls, what was promised, who has gone quiet. A first-meeting brief has little or no history, so it leans on public research about the company and each attendee, how the meeting was booked, and questions to qualify the deal. Gong notes that its summary is generated only when there is prior tracked activity with the account, such as calls or emails, so plan a separate first-meeting template. An agent with web search can do that research itself and cite the pages it used.

### What is Day AI?

Day AI is the customer memory and agent platform for go-to-market teams. Even the best CRM only holds a fraction of what actually happened with your customers. Day AI is what comes after CRM. It builds a self-updating customer record from every meeting, email, and call, and runs governed agents that create finished work on top of it, with every fact and every action traced to source conversations. It does everything teams use a CRM for, without the data entry, and does what a CRM never could. Agents prospect, prep every meeting, draft follow-ups, keep the pipeline clean, and much more, giving every rep added selling capacity that used to require hiring additional people. It runs as a layer on top of Salesforce or HubSpot or in place of them.

Canonical: https://day.ai/guides/ai-meeting-prep