How to prepare for a sales call with AI

Most pre-call briefs die the same way: a 12-page PDF lands in the inbox 40 minutes before the meeting, the rep opens the first two sections, skims a company overview they already know, and walks into the call with the same three talking points they used last week.

The research was fine. The delivery was the problem. A brief that takes longer to read than the call is worth preparing for is not a brief. It is a homework assignment the rep will skip.

This post is about what to cut, what must stay, and how to tell the difference before the next invite hits the calendar.

The brief is a reading problem, not a research problem

Sales teams already know how to research. LinkedIn, the CRM, the last call recording, the proposal thread. The failure mode is not missing information. It is burying the next 45 minutes under everything that was ever true about the account.

A useful brief answers three questions a rep will ask in the elevator:

  1. Who is in the room, and what do they care about?

  2. What changed since we last spoke?

  3. What should I ask or push for today?

Everything else is archive. Archives are valuable. They do not belong in the document someone is supposed to finish before a 2pm call.

If a section does not change what the rep says in the first ten minutes, cut it. Company founding year, generic industry trends, a competitor landscape that has not moved since the last meeting: all of it can live on the deal record. None of it earns a place in a pre-call brief.

What reps actually open

Watch what happens when a brief is short enough to finish. Reps open the sections that map to risk in the room:

  • Attendees they have not met, or have not met recently

  • Open risks that could surface on this call

  • Recommended questions or talk tracks tied to those people

  • Where the deal stands against the last meeting, in one screen

They skip the rest. That is not laziness. That is a rational response to a clock.

Ruby ships every pre-meeting brief in a fixed shape for this reason. The headers are always the same, so a rep recognises the document before they read a word of it: About The Attendees, Where We Stand, Last Meeting, Meeting Recommendations, Deal Recap, Account Recap. First meetings swap Where We Stand and Last Meeting for Top Use Cases. The email version stays under 300 words, with a link back to the full note on the HubSpot deal. The shape is the product. Length is a design choice, not an accident of how much research the model produced.

The cut list

Use this as a filter on any brief your team still builds by hand, or as a checklist against whatever tool writes them for you.

Cut: company encyclopedias. A paragraph on what the account sells is enough when the rep has already had a first call. A full firmographic dump is for the AE who has never touched the account. If both audiences need the same brief, you have two audiences and one document. Split them.

Cut: recycled battle cards. Competitor content that has not changed since the last meeting does not need to be pasted in again. Link it. Name the competitor only when someone on this invite raised them, or when the deal stage makes a head-to-head likely today.

Cut: every stakeholder who is not on the invite. Influence maps are deal-level assets. The brief should narrow to the people who accepted the calendar hold. Recommendations that ignore who is actually in the room get ignored in return.

Cut: MEDDPICC scores with no implication. A health score without a sentence on what it means for this call is decoration. If pain is strong and the economic buyer is still missing, say that, and say what question unlocks it today.

Cut: action items the buyer already closed. Old next steps that cleared last week create false urgency. The brief should carry open loops only.

Keep: one row per attendee. Name, role, last touch, what they care about in this deal. Tables beat paragraphs here. A rep scanning for "who is the CFO" should not have to parse prose.

Keep: what changed. Two to five lines on what moved since the last meeting: a new risk, a silent stakeholder, a proposal still outstanding, a compelling event that got stronger or weaker. Change is the only reason to reopen the deal before the call.

Keep: three meeting recommendations, max. Not a brainstorm. Three pushes or questions filtered to the people in the room, each with a why. If you cannot defend the third, you only had two.

Keep: the deal state in one block. Current stage, open risks, MEDDPICC hot spots, next best action. Enough that a manager reading over the rep's shoulder knows what "good" looks like for this call.

A worked example of the cut

Before: 1,800 words. Company overview (220), full org chart (400), competitive landscape (350), last three meeting summaries pasted in full (500), a twelve-item recommendation list (330).

After: ~450 words.

  • Attendee table: three rows (champion, economic buyer proxy, technical validator)

  • Where we stand: four lines (proposal out 11 days, EB still absent, renewal language from last call fails the three-part compelling-event test, technical validator asked for security questionnaire)

  • Meeting recommendations:

    1. Get a named date and consequence from the champion on what happens if the renewal lapses with no decision

    2. Ask who owns the security questionnaire on their side, and whether procurement is already looped in

    3. Do not re-pitch product. Open with the open risk: EB has not been in a room with you

  • Deal recap: health score, top risk, one-line hypothesis for the call

Same research underneath. Different job for the document.

How to run the audit on your next five briefs

Pick five upcoming calls. For each brief your team currently uses, time how long a cold reader needs to find: (1) who is new on the invite, (2) the single biggest risk for this call, (3) the first question the rep should ask.

If any of those three takes more than 60 seconds, the brief is too long or badly ordered. Fix order before you fix research. Put attendees and risks above history. Put recommendations above account lore.

Then apply a hard cap. A pre-call email under 300 words. A deal note that a rep can finish in four minutes. Anything past that belongs in the CRM as reference, not in the brief as required reading.

Ruby writes the HubSpot pre-meeting note up to 24 hours before an eligible call, and keeps the email short on purpose. The discipline is the same whether a human or a model does the writing: the brief exists to change the next conversation, not to prove the research was thorough.

Frequently Asked Questions

How long should a pre-call brief be?

Short enough that a rep finishes it before the call without treating it as a project. A practical target is an email under 300 words plus a deal note a cold reader can finish in about four minutes. If the brief is longer than the prep time the meeting deserves, it will not get read.

What belongs in a pre-call brief vs the CRM?

The brief carries what changes what the rep says today: attendees in the room, what moved since last time, open risks, and a small set of recommendations. The CRM holds the archive: full history, org maps, battle cards, and closed loops. Linking to the archive beats pasting it.

Should every meeting get a brief?

Brief the meetings where a prospect is in the room. Internal-only meetings do not need one. First meetings still do: research from scratch, labelled as a working hypothesis rather than confirmed deal positioning.

How is this different from a call recording summary?

A recording summary looks backward at what was said. A pre-call brief looks forward at who will be there and what to do about it. Both matter. Only one changes the opening of the next call.

How does Ruby deliver pre-call briefs?

Ruby checks the calendar for eligible meetings in the next 24 hours and writes a Ruby Pre-Meeting Brief to the linked HubSpot deal as a note. An email version under 300 words can go out with a link back to the deal; reps can switch the email off and still get the note on the record.