Let new reps make their mistakes on you, not on buyers
Rehearse, score, and replay the objections your buyers really raise
This one is for the sales managers, the enablement leads, and anyone who has watched a promising new hire walk into their first live call and improvise an answer to an objection the rest of the team has heard two hundred times.
What: Using Breeze Assistant as a practice partner that plays your actual buyers: you hand it the objections you have already evidenced, from your battlecard or straight out of your closed-lost reasons, and it becomes the buyer who raises them. A new rep rehearses the conversation twenty times before having it once for real, gets scored against your own playbook rather than generic best practice, and replays the exact moment that cost them until they can handle it.
Prompt of the week:
Picture the second week of a new rep's job. They have absorbed a product deck, sat in on two calls, read the pricing page, and been told to shadow whoever was free. Then they get their own first conversation, and roughly four minutes in the prospect says something about budget, or about the incumbent supplier, or about needing to speak to someone else first, and the rep invents an answer on the spot. It is not a bad answer, necessarily. It is just the first time anyone has ever asked them, and they are finding out what they think while a real buyer listens.
That moment is expensive, and it is repeated across the industry constantly. The benchmarks put average business-to-business rep ramp somewhere between three and six months depending on who is counting, with one widely cited survey placing the median account executive at a little over four months and sales development reps at around 142 days before their first booked meeting. Ramp has stretched by roughly a third since 2020, as products got more complicated and managers got less time each to coach. Every month of half productivity is real money, and about a third of new sales development reps leave inside a year, often before they ever became fully productive at all.
Here is the part that should change how you onboard. When people compare the teams that ramp in ninety days with the ones that take a hundred and eighty, the difference is not product knowledge, and it is not motivation. It is practice volume in the first fortnight. A traditional programme gives a new rep perhaps five to ten mock calls in their first month, squeezed around a manager's diary. Teams that lean on simulated practice put reps through dozens of rehearsed conversations in the same period, so that by the time a real buyer pushes back, the rep has met that objection thirty times already. It also fixes the other well-known problem, that the large majority of sales training is forgotten within three months unless it gets applied.
The good news is that you almost certainly already know what the objections are. A new rep will hear broadly the same fifteen to twenty of them in their first month, and they are recorded in your closed-lost reasons and your call notes, which is exactly the raw material our battlecard prompt turns into a ranked list. So this prompt does not spend its time re-deriving that list. If you have a battlecard, you paste it in. If you do not, it takes a quick look at your lost reasons and pulls a shortlist. Either way, the objection set is the input, not the output.
What the prompt actually does is the part nobody has built: it turns that list into a person. Breeze becomes a composite buyer from your market, opens the call, raises those objections the way your buyers really raise them, and refuses to be talked round by the first plausible answer. Then it drops character, scores the attempt against your playbook, and replays the single exchange that went worst so the rep can try it again while it still stings. That distinction matters, because a battlecard tells a rep what to say and rehearsal is what makes them able to say it. It is text rather than voice, so it builds the thinking and the language rather than the delivery, which is exactly the part that improvising damages most.
Prompt structure
Paste this into Breeze Assistant and make sure CRM data access is enabled in your AI settings so Breeze can reference your closed-lost reasons, deal history, call and meeting notes, contacts, and the personas behind them:
Role: You are a sales coach and practice partner. You know that new
reps ramp faster through rehearsal than through reading, that the
objections worth rehearsing are the ones our own buyers actually
raise, and that your job is the practice rather than the research. You
play a buyer convincingly, you do not go easy, and you coach honestly
once the call is over.
Task: Run a practice call for a new rep. Take the objection set I give
you below, become a composite buyer who raises those objections, and
role-play the conversation in character. Then score the rep against
our playbook, replay the exchange that cost them most so they can try
it again, and tell them the one thing to work on.
Context:
- Company: [COMPANY NAME]
- Industry: [INDUSTRY]
- What we sell, in one line: [PRODUCT OR SERVICE]
- Typical deal size and sales cycle: [so the pressure is realistic]
- The rep practising: [name, role, and how new they are]
- The objection set to rehearse (choose one):
[PASTE the objections table from our battlecard, or a list of the
objections we lose to, or write "none, pull a shortlist from our
closed-lost reasons"]
- The buyer type to practise against: [job title, segment, or
"pick the one we lose to most often"]
- The call stage to rehearse: [first discovery call / demo /
pricing conversation / late-stage push back]
- Difficulty: [WARM, receptive but busy / SCEPTICAL, needs
convincing / HARD, short on time and loyal to the incumbent]
- Our playbook or methodology, if we have one:
[name it, or "none documented"]
Run the following:
1. SETTLE THE OBJECTION SET (keep this brief)
- If I pasted a battlecard or a list, use it as given. Do not
re-derive or re-rank it
- If I did not, take a quick look at our closed-lost reasons and
propose a shortlist of the five objections most worth
rehearsing. Do not produce a full analysis, we only need enough
to practise against
- Confirm the three or four you will actually raise in this call,
then move straight on. The rehearsal is the point, not the list
- If our recorded data is too thin to find any real pattern, say
so plainly rather than substituting generic objections
2. BUILD THE BUYER
- Create a composite buyer for the type I named: their role, what
they are measured on, the pressures they are under, what they
care about, and how they tend to push back
- Base this on patterns across our real contacts and deals, but
never simulate a specific named individual or use a real
person's identity
- Give them a plausible current situation: what they already use,
what prompted this conversation, and what would genuinely worry
them
3. RUN THE ROLE-PLAY
Open the conversation in character and wait for the rep to
respond. Then follow these rules:
- Stay fully in character. Do not coach, hint, or comment on the
rep's performance while the call is running
- Raise one objection at a time, drawn from the ranked set, in an
order that feels natural for this stage of conversation
- Do not fold at the first reasonable answer. A real buyer pushes
back twice. Concede ground only when genuinely persuaded
- Behave like the difficulty level I chose: distracted and brief
if hard, engaged but unconvinced if sceptical
- End the call the way a real one ends: agreeing a next step,
asking for something, or politely declining. Do not end on a
summary
4. SCORE THE ATTEMPT
Once the call has ended, drop character and coach:
- Score the rep against what good looks like in our playbook: did
they acknowledge the objection before answering it, ask a
diagnostic question rather than pitching, use a relevant proof
point, and secure a concrete next step?
- Quote back the moment that went best and the moment that cost
them most, using their actual words
- Be direct. Encouraging is fine, flattering is not: a rep who is
told they handled it well when they did not will lose the real
call
5. REPLAY THE MOMENT
Do not move on from the exchange that cost them most. Run it back:
- Quote the buyer line that caused the trouble, go straight back
into character, and put it to them again
- Let them try it as many times as they want. Do not accept a
weaker answer than the one before just to move things along
- When they get it right, say what specifically made that version
land, so they can repeat the reasoning rather than the sentence
6. ESCALATE AND VARY
- Recommend the next difficulty level and the next objection
category to rehearse
- Mix the objections between sessions so the rep builds judgement
rather than memorising a script
7. TRACK PROGRESS
- Keep a short running record: which objections this rep now
handles well, and which keep catching them out
- Name the pattern their manager should coach in person
Constraints:
- Every objection must come from our recorded data, not from a
generic business-to-business objection list. Practising against
invented objections teaches the wrong reflexes
- Never break character during the role-play. Coaching belongs in the
scoring phase only
- Never simulate a real named person. Build composite buyers from
patterns across accounts
- Score against our playbook where we have one. Where we do not, say
which standard you are applying so it can be challenged
- Do not invent objections, deal outcomes, or quotes from calls. If
something you need is not visible from the current context, state:
"SIGNAL MISSING: [what needs checking manually]"
- Keep the coaching to one primary improvement per session. A rep
given six corrections will apply none of them
Output format:
### I. WHAT WE ARE REHEARSING
{The three or four objections you will raise in this call, listed in
one line each. No analysis}
### II. THE BUYER YOU ARE ABOUT TO MEET
{The composite buyer: role, pressures, situation, and how they push
back. Two short paragraphs, no more}
### III. THE CALL
{Open in character, in their voice, and stop. Wait for the rep}
### IV. SCORECARD
| What Good Looks Like | What You Did | Score | How to Fix It |
### V. REPLAY
{Back in character: the line that caused the trouble, put to them
again. Then stop and wait}
### VI. THE ONE THING TO WORK ON
{A single specific improvement, with the better wording to try}
### VII. NEXT SESSION
{The objection category and difficulty to rehearse next, and the
pattern for their manager to coach in person}
Why this prompt works, and how to adapt it
Onboarding tends to be built around transferring information, on the reasonable-sounding theory that a rep who knows enough will sell. The trouble is that selling is a performance, and nobody gets good at a performance by reading about it. What separates a fast ramp from a slow one is the number of times a rep has already been through the moment before the moment matters, and that is a thing you can manufacture cheaply once you accept that the practice partner does not have to be a busy human being. This prompt manufactures it, and grounds it in the only objections that count, which are the ones your own buyers keep raising.
A few things to note about how it is constructed:
It consumes your evidence rather than rebuilding it. Any AI can generate a list of plausible business-to-business objections, and rehearsing against those trains a rep for a market they do not sell into. But you probably do not need Breeze to find your objections a second time, because if you have built a battlecard from your closed-lost reasons then the ranked list already exists. The prompt is deliberately told to take it as given and move on, which keeps the session short and puts the effort where the value is.
It is instructed not to fold. The most common failure of an AI role-play is that it caves gratefully at the first half-decent answer, which feels lovely and teaches false confidence. Real buyers push back a second time, then a third, and the rep who has only ever met the polite version of an objection is unprepared for the real one. Telling the model to concede only when genuinely persuaded is what makes the practice worth having.
Coaching is banned during the call. If the buyer keeps stepping out of character to offer helpful pointers, the rep is no longer having a conversation, they are being tutored, and the pressure that makes rehearsal useful disappears. Splitting the session cleanly into the call and then the debrief mirrors how real coaching works, and it means the rep has to find their own way through the awkward pause rather than being rescued from it.
The replay is where the learning actually lands. A scorecard tells a rep what went wrong, which is information, not skill. Putting the difficult line straight back to them, in character, while the discomfort is still fresh, is the bit that changes what they will say on Thursday. It is also why the prompt is told not to accept a weaker answer than the previous attempt just to move the session along, and to explain what made the successful version work, so the rep carries the reasoning rather than a memorised sentence. Scoring is judged against your playbook where you have one, so the feedback matches what their manager would say anyway.
Difficulty escalates so nobody memorises a script. A rep who runs the same scenario ten times learns one answer very well and stays fragile everywhere else. By varying the objection order and stepping the buyer from receptive to sceptical to genuinely difficult, the sessions build the thing that actually transfers to live calls, which is judgement about which objection they are facing rather than a rehearsed paragraph.
“SIGNAL MISSING” protects the realism. This is the one prompt in the series where a fabricated detail does more than mislead a report: it trains a person. If Breeze cannot find enough recorded losses to know what your buyers object to, the honest move is to say the data is too thin, because a rep who spends a fortnight rehearsing invented objections has been made confident about the wrong conversation. Where the record runs out, the flag says so, and someone can go and fill the gap properly.
Adapting it for your portal:
Brand new to sales entirely? If the rep has never carried a quota, do not open on the hard version. Add: “This rep is new to sales, not just new to us. Start with a warm, receptive buyer and one objection only, and tell me when they are ready to move up a level.” The session becomes a confidence builder that still uses real material.
Onboarding a cohort at once? If several people started together, add: “I am onboarding [N] reps together. Use the same objection set for all of them, and after each session tell me which objections the group as a whole is weakest on so I can teach that once rather than five times.” You get individual practice and a cohort-level coaching agenda.
Losing to one particular competitor? This is where the battlecard you already built earns a second use: “Here is our battlecard for [competitor]. Play a buyer who currently uses them and is happy enough, raise their objections in that order, and push hard when I claim we are different.” The card stops being a reference document and becomes the script your rep rehearses against.
Selling into committees? If your deals involve several people, add: “Our deals involve multiple stakeholders. Play the buyer who was not on the original call, the one who arrives late with concerns nobody prepared for, and test whether the rep can handle an objection second hand.” It rehearses the stall that single-threaded reps never see coming.
Experienced hire, new product? If they can sell but do not know your world yet, add: “This rep is experienced but new to our product and market. Skip basic discovery technique and focus entirely on the objections specific to our category, our competitors, and our pricing model.” The practice targets the genuine gap instead of insulting them.
Want to see progress over time? Save the scorecards and re-run the session a fortnight later with: “Compare against the sessions from [DATE]. Which objections does this rep now handle confidently, which are still catching them out, and what should we rehearse this week.” That turns scattered practice into a visible ramp curve you can show a manager.
Beyond the prompt:
A practice partner is only useful if reps actually use it, and if what happens in the sessions travels into live calls. A few habits make that likely rather than accidental.
Put the sessions in the first fortnight, before the first real call. This is the whole point, and it is the thing most onboarding gets backwards: rehearsal usually arrives as remediation after a rep has lost something, when it is worth far more as preparation before they have anything to lose. Book the practice into week one and week two deliberately, the way you would book product training, rather than leaving it as something they might get round to.
Ask them to answer as though someone were listening. The temptation in a text-based rehearsal is to draft, edit and polish, which produces a beautiful answer nobody could ever say out loud under pressure. Reps get far more from typing their first instinct, fast and unedited, and better still from reading it aloud as they go, because the objection they will face on Thursday will not wait while they compose a paragraph.
Keep the scorecards, and read them sideways. One session tells you how a rep did on a Tuesday. Six sessions tell you which category they consistently fumble, and that pattern is a genuinely useful coaching agenda, arrived at without a manager having to sit through every rehearsal. It also turns a vague sense that someone is coming along nicely into something you can actually point at.
Then stay involved yourself, because the machine cannot replace you and should not try. The clearest finding in the onboarding research is that manager involvement is the variable that most separates fast-ramping teams from slow ones, and reps who have a good first ninety days are dramatically more likely to still be there a year later. Use the sessions to arrive at your coaching conversations already knowing what to work on, not to avoid having them.
Nobody remembers the objection they read about. They remember the one they fumbled, went away and thought about, and got right the second time. The kindest thing you can do for a new rep is to give them somewhere to fumble it that does not cost them a deal, and to make sure the thing they are fumbling is the objection your buyers will genuinely put to them, rather than a plausible invention. Get that right and a first live call stops being an examination and starts being a conversation they have had before.
Written by Marek Ertinger - Technical Consultant