By Jatin Modi, founder of Renaissance Group · Published 28 September 2026 · Updated 29 October 2026
What should you check before buying an AI tool that prepares sales meeting briefs?
Check that every line in the brief links to the record it came from, that the tool builds from your own CRM and email history, that it says so when an account has little history, and that the seller edits the page before the meeting. A blind test on ten past meetings checks all four in one afternoon. Seven preparation checks come first, because a brief can only summarise what your systems hold.
Put attendee names, their companies and a purpose line on every customer meeting invite.
Log meeting outcomes in Salesforce or HubSpot in the same fields every time, so the history block reads cleanly.
Link open orders and quotes to the account record.
Merge duplicate accounts and contacts, since a brief for a duplicate shows half the history.
Agree what sits at the top of the page for each meeting type: first call, renewal, escalation.
Name who edits the brief and who else reads it, for example the sales engineer or the executive sponsor.
Pick one measure: the share of customer meetings where the seller opened the brief before the call.
How much seller time goes to preparation?
Salesforce surveyed 4,050 sales professionals in 22 countries in August and September 2025. The average seller spends 40% of their time selling, and sellers expect AI agents to cut prospect research time by 34% once fully implemented (source: Salesforce State of Sales, 3 February 2026). Salesforce sells CRM and AI agents, so treat the figures as one vendor's survey.
An older XANT Labs survey of more than 200 sales reps put researching target accounts and contacts at 12% of the working week (source: InsideSales, 10 November 2017). The sample is small and the date is old, so read it as a rough indication of the size of the task.
The 34% in the Salesforce survey (source above) is an expectation that sellers reported. Your own meetings give the measured figure, and the blind test below produces it.
What goes on a one-page meeting brief?
Five blocks fill the page. The table shows where each block draws from and what separates a weak version from a useful one.
| Block | Drawn from | Weak version | Useful version |
|---|---|---|---|
| People in the room | Calendar invite and attendee profiles | A list of names and job titles | Each person with company, role, and the last time your team spoke with them |
| History | Salesforce records and email threads | The CRM fields copied out | What changed since the last contact, with dates and links to the threads |
| Open orders | Your order records | A total | Each open order with its status and owner |
| News | Company news from named sources | Headlines from any source | Dated items from named sources, each with a link and one line on why it matters to this meeting |
| What to bring | All of the above | A generic talking point | One suggestion tied to a specific line in the history or the news |
What goes wrong with AI meeting briefs?
| Failure | Control to ask for |
|---|---|
| The profile belongs to a different person with the same name | The matched person shown with company and a source link, so the seller sees why the match was made |
| The history block repeats CRM fields and tells the seller nothing new | A list of events since the last contact, each with a date |
| News items come from unrelated companies or old articles | Every item dated, linked and attributed to a named source |
| An account with little history gets padded text | An explicit line stating that no history is on record, with attendees and news moved up |
| The brief arrives after the meeting starts | Delivery fixed to the day before, attached to the invite |
| A private internal note appears on a page that other attendees read | A permission rule on which sources a shared brief may draw from |
What should you ask any vendor of AI meeting prep?
| Question | A good answer shows | Why it matters |
|---|---|---|
| Does every line show where it came from? | A source link per line that opens the record or article | The seller verifies a claim in one click before saying it to a customer |
| What does it write for an account with little history? | A statement that history is thin, and a page built from attendees and news | Padding invented to fill a template misleads the seller |
| Where does the brief appear? | On the calendar invite or in the tool the seller already opens | A brief in a separate app goes unread |
| Who can edit it before the meeting? | The seller, with the edited page kept as the version everyone reads | The seller answers for what is said in the room |
| Which systems does it read, and under whose permissions? | A named list: calendar, CRM, email, news; your tenant and your access rules | A brief that reads more than the seller may see creates a data problem |
| How does it choose between two contacts with the same name? | Matching on company and email domain, with the match shown | A wrong profile in front of a customer costs credibility |
| How current is the news block, and from which sources? | A source list and a date on each item | Stale news is the most visible error on the page |
| Where does meeting content go, and for how long? | Your tenant and a stated retention rule | Invites and notes hold customer information |
How can a buyer test an AI meeting prep tool in one afternoon?
Take ten customer meetings from last quarter: five with long-standing accounts and five with thin history. Ask the tool to write a brief for each, using only information that existed the day before. Then ask the seller who ran each meeting to score every brief with four counts.
| Count | How to score it | What a weak result means |
|---|---|---|
| Correct and new to the seller | A true statement the seller did not already know and would have used | The tool adds little. This count shows the value. |
| Correct and already known | A true statement the seller already carried in their head | Expected on long-standing accounts. A brief made only of these saves no time. |
| Wrong or unsupported | A statement that its linked source does not back | The tool invents. Every line then needs a human fact-check. |
| Gap flagged | The tool says it has no history or no news | A high count on thin accounts shows honest behaviour. |
Any false statement about a named person with no source counts as disqualifying. Compare the time the seller spent preparing last quarter with the time the brief takes to read and edit. The test costs one afternoon and replaces a demo built on the vendor's own accounts.
How does Hastings run the meeting brief?
Hastings sells this job, so read this section as a vendor description. The calendar triggers the work. When a customer meeting sits in a seller's diary, Hastings reads the invite, the Salesforce history, the email threads, the attendee profiles, company news and open orders, writes the page the day before and attaches it to the meeting in Outlook or Gmail. A copy lands in SharePoint or Google Drive.
| Stage | What happens | Who |
|---|---|---|
| Inputs | Calendar invites, Salesforce, email threads, attendee profiles, company news and open orders, connected in setup | Connected in setup |
| Compile | One page per customer meeting with the five blocks and a suggestion for what to bring | Hastings agents |
| Human gate | The seller reads the page, edits it and decides what to say. Everyone joining the call reads the same page. | Your seller |
| Output | The page on the invite in Outlook or Gmail, with a copy in SharePoint or Google Drive, the day before | Delivered daily |
Week one starts with a working session, where Hastings builds a sample on your company from public data, so the first brief you read describes a company you know. Setup follows. According to the pricing page, Hastings connects your CRM, email, calendar and market sources, builds the knowledge base and configures the first jobs. The pilot calendar puts connection and build in weeks 1 to 2 and runs the jobs against agreed success measures in weeks 3 to 12.
Where does the meeting brief stop?
The brief reads what Salesforce and email hold. An account with a thin history gives it little to summarise, so news and attendee profiles carry more of the page. It prepares the meeting and ends at the door. The suggestion of what to bring is a starting point, and the seller decides what goes on the table. A person on your team approves everything that goes out.
What was agreed in the room is recorded by CRM kept current. The seller's day across all meetings is ordered by the Morning briefing, and changes at the customer's competitors come from Market and competitor watch. The managed AI operating system page explains how these jobs fit together.
Questions buyers ask
Is there an AI tool that writes a one-page brief before every sales meeting?
Yes. Hastings writes one page for each customer meeting the day before and attaches it to the calendar invite, with the attendees, the account history, open orders, news and a suggestion for what to propose. The seller edits the page before the meeting.
What should an AI meeting brief contain?
Five blocks: the people in the room, the history with the account, the open orders, recent news and a suggestion for what to bring. Each line should link to the record it came from.
What should you check before buying an AI meeting prep tool for a sales team?
Check that every line shows its source, that the tool reads your own CRM and email history, that it says so when an account has little history, that the seller can edit the page, and where meeting content is stored. Run the blind test on ten past meetings before you sign.
How does a meeting brief differ from a call summary?
A brief is written before the meeting from the history that exists. A call summary is written after the meeting from what was said. CRM kept current covers the second job.
When does the brief arrive, and who can read it?
The day before the meeting, attached to the invite in Outlook or Gmail with a copy in SharePoint or Google Drive. Everyone joining the call reads the same page, including a sales engineer who dials in for part of it.
Which systems does the brief read?
The calendar invite, Salesforce, email threads, attendee profiles, company news and your open orders.
Sources
Salesforce, State of Sales report (4,050 sales professionals, 22 countries, survey August to September 2025), announcement of 3 February 2026: salesforce.com.
XANT Labs, Time Management in Sales (more than 200 sales reps), 10 November 2017: insidesales.com.
Hastings pricing page for setup scope and the pilot calendar.