Where we are
The recap went to Mira Chen inside the hour and she corrected one line. Then three things arrived: a price question forwarded from Priya Raman, the CFO, who was not on the call; a note from Jonas Weber in IT saying their existing suite "already includes a note-taker"; and silence on the date you proposed. Tonight is stage 4 of the deal loop, Advance. The deliverable is a rehearsed, defensible answer to your three worst objections, and a correct diagnosis of what each one actually is. b7 turns the surviving deal into a proposal.
Rehearsal without learning to bluff 7 min live
This is the one place where AI does something a rep genuinely could not get before. You can run the CFO call ten times on a Tuesday evening, against a buyer who interrupts, who has a budget reality, who does not care about your roadmap. Practice used to require a colleague with a free hour and the willingness to be difficult. Now it costs four minutes. And it will make you worse if you use it wrong, because rehearsal optimises for fluency, and fluency is not what your buyer is short of. Gartner predicts that by 2030, 75% of B2B buyers will prefer sales experiences that prioritise human interaction over AI. They are not starved of smooth answers. They are starved of true ones.
LiveThe rail: one follow-up you did not script4 min▶
Here is the failure, concretely. You rehearse the security objection eight times. By the eighth run your answer is smooth: retention controls, access model, SOC 2, happy to send the documentation. Then Jonas asks the ninth question, the one you never scripted: "Where is the data actually stored, and can we keep it in-region?" And you say something. Something confident, because you have been confident for eight rounds. And it is wrong, and Jonas checks it, and now everything else you said is under review too.
So the rail is a single sentence, and it applies to every rehearsed answer you will ever build:
- Make the simulated buyer push twice. A model that accepts your first answer is flattering you. The prompt in the build-along explicitly instructs it to push back twice and to get more specific each time.
- Then have a human ask one. A colleague will ask the question a model will not, because they know where your product is thin. This is the step people skip and it is the whole exercise.
- The good outcome is often "I do not know". "I do not know where it is stored by default. I will find out today and send you the answer with the docs" is a defensible answer. It costs you nothing except the illusion of expertise, and it is what a buyer who has been sold to before is actually listening for.
- Never rehearse a claim you cannot source. If the answer contains a number, know where the number came from. The Pacific Freight figure is yours to use because it is real: their Tuas ramp took action items closed within a week from 41% to 88% in one quarter. Any number you cannot trace that way does not go in the answer.
Self-studyWhy simulated buyers are easy to fool, and how to make them harder3 min read▶
A model playing a buyer defaults to being agreeable, well-informed and interested. Real buyers are none of those things reliably. Four fixes that make the rehearsal worth the time:
- Name the person and give them a constraint. "Priya Raman, CFO, Northwind Logistics, has a headcount freeze until the Q3 close and was not on the discovery call" produces a completely different conversation from "act as a CFO".
- Give them a reason not to care. Real objections come from people with something better to do. Tell the model the buyer thinks this is a nice-to-have and has twenty minutes.
- Ban the resolution. Without instruction, a simulated buyer says "that makes sense" and books the meeting, because it has read a million sales transcripts that end that way. Tell it explicitly that it does not agree at the end of this call.
- Do not let it invent facts about the account. Give it what you know about Northwind - the Jurong lease, the three open ops reqs, Jonas in IT, Priya's freeze - and tell it to ask you rather than invent anything else. Otherwise you rehearse against a fictional company and get fluent about nothing.
And when it is over, remember what you were practising for. Rehearsal is a way to find your own answer under pressure. It is not a way to acquire an answer you do not have.
Most objections are a missing stakeholder 6 min live
Reps lose deals by answering objections well. The answer is fluent, the buyer nods, and nothing moves, because the thing that was said was not the thing that was true. There are only four kinds. Price is usually value that has not been framed in their numbers, or the wrong seat entirely. Timing is usually no trigger, which sends you back to Target in the deal loop. Authority is a person who is not in the room and has not heard any of this. Trust is something you did earlier. Only the first one is answered by talking, and the third one is the biggest branch by a distance.
LiveHow to tell which kind you are actually holding4 min▶
The words the buyer uses are close to useless as a classifier. "It is too expensive" is said by all four. Use these tests instead, in order:
| Test | If yes | What it means |
|---|---|---|
| Did the objection arrive after you sent something to someone you have never spoken to? | Authority | You are hearing a second-hand summary of a stranger's reaction. Get the stranger in the room. |
| Can she tell you what the current cost of the problem is, in her own numbers? | Price if no | You never framed the value in her units, so the price has nothing to sit against. |
| Is there a dated event forcing a decision this quarter? | Timing if no | There is no trigger. This is a Target problem, and talking will not create one. |
| Did she go quieter, shorter and more polite than she was on the first call? | Trust | Something changed. Usually something you said, promised, or over-claimed. |
Then the one question that resolves most of it, asked plainly: "Other than you, who has to be comfortable with this before it happens?" Ask it on the first call, not the fourth. Mira will answer it honestly, and the answer is usually two names you have not met.
Self-studyThe stall that is not an objection at all3 min read▶
Silence is the fifth thing, and it is not on the diagram because it is not an objection - it is an objection you were not told about. The date passed, Mira did not reply, and the deal sits in your pipeline looking alive. Three honest readings:
- She hit an internal wall and it is embarrassing. Priya said no, or IT added four weeks. Champions go quiet when they have lost an argument on your behalf. The move is to make it easy to say so: "If this has slipped behind other things, that is completely fine - tell me and I will come back in the new quarter."
- Nothing changed for her. No trigger, so no urgency, so no reply. Timing. Back to Target, and park it with a date and a reason.
- You are the fourth of eleven emails and none of them were interesting. Which is b3's problem, arriving late.
What you must not do is send the chase that pretends nothing happened. Two follow-ups, spaced, each carrying something new, then a clean close-out that leaves the door open. b7 covers the cadence properly and b9 covers what a stalled deal is allowed to do to your forecast, which is: nothing.
Competitors, bundles, and the built-in AI cards · read in any order
Five cards. Two of them are the objections you will meet in almost every deal from here, one is the rule about what a model may never say about another company, and one is a worked script for the exact situation you are in with Priya Raman: a price question from a person who was not on the call.
LiveHonest comparison, and the line a model must never cross4 min▶
You will be asked to compare yourself to two kinds of competitor: a big incumbent suite that bundles a note-taker, and a cheap point tool. Here is what you may say, and what you must never generate.
| You may say | You may never say |
|---|---|
| Published, checkable facts about their product, dated and sourced | Any weakness, incident or limitation you cannot point to a source for |
| What you do that they do not, stated as a capability rather than an insult | Anything about their security, finances or customer losses that you heard second-hand |
| Where they are genuinely a better fit, out loud | A comparison table a model produced about a competitor you did not verify line by line |
| What your own customers switched for, in their words | A claim about their roadmap. You do not know it and neither does the model |
The safe and genuinely more effective version: ask the model to find and cite what the competitor publicly claims, then you read those claims and decide what is true. Finding, not writing. Same rule as b3.
Live"We already have this - it is bundled"3 min▶
Jonas says their existing suite includes a note-taker. This is not a price objection and it is not a competitor objection. It is a question about whether the problem is real, wearing a procurement costume. Three moves, in order:
- Agree fast and specifically. "You do have one, and it transcribes well." Arguing with a true statement costs you the next four minutes and all of your credibility.
- Ask what it has changed. "How many of the action items from your Monday ops call get closed inside the week now?" If they know the number, you are in a real conversation. If nobody knows the number, you have just shown them the gap without saying a word about your product.
- Name the difference as a job, not a feature. Transcribing a meeting and closing the loop on what was decided are different jobs. Pacific Freight had recordings before Cadence too - their Tuas ramp went from 41% of action items closed within a week to 88% in one quarter, and the recordings were not the variable.
Then the honest half: if their bundled tool is genuinely doing that job and the number is good, say so and stop selling. You will lose the deal you were going to lose anyway, and you will keep a relationship that is worth more than the quarter.
Live"We will just use the built-in AI"3 min▶
The cheap point tool version of the same objection, and the one where reps most often bluff. The honest answer starts by conceding the part that is true:
- Concede the transcription. It is close to a commodity. Do not build a case on being marginally better at it, because you probably are not and the buyer can test it in a day.
- Move the question to what happens after the meeting. Who owns the action item, what happens when nobody does it, and whether anyone can see the pattern across forty meetings. That is a different product, and it is a fair thing to say out loud.
- Answer the price gap directly. "It is more expensive, and here is what the difference buys" beats any attempt to make the gap disappear. Priya will respect the first version and distrust the second.
- Do not invent limitations. If you do not know whether the built-in tool supports something, say you do not know and find out. The buyer has probably already tried it.
This is exactly where rehearsal earns its keep and exactly where it turns into bluffing. Run this objection through the loop until your answer survives the follow-up "what specifically can yours do that ours cannot, that we would actually use?" If it does not survive, you do not have an answer yet, and knowing that is worth more than the fluent version.
LiveWorked script: Priya asks a price question Mira cannot answer4 min▶
The situation: Priya Raman, CFO, was not on the discovery call. She has forwarded one line through Mira: "Why is this three times the cost of what we already pay for the same thing?" Mira does not know the answer and is asking you for a reply she can paste.
Diagnose first. This looks like price. It is authority. The person who has to be comfortable is not in the room, has heard none of the discovery, and is being asked to evaluate a number with no context attached. Answering the price question by email is answering the wrong question well.
- It refuses to make Mira your proxy. Champions lose arguments they were not equipped to have, and then they go quiet. Protect her from that.
- It gives her one line she can forward anyway, because she may not get you the meeting and she needs something.
- It names the price honestly. More expensive, and here is what for. The price stays human - that is the trust line from b1, and it is the sentence in this deal you must write yourself.
- The ask is small and specific. Fifteen minutes with a named person, with a reason that is about Priya's interest and not yours.
Self-studyBuild your objection library from your own calls, not a battlecard4 min read▶
Vendor battlecards are written by people who have never been asked the question by your buyer. They contain the objections marketing expects and the answers marketing wishes were true. You have something better and you are already sitting on it: your recorded calls from b5.
- Pull the objections out of the last twenty calls. Search the transcripts for the moment the buyer pushes back. Write each one down in the buyer's exact words, not your paraphrase - the paraphrase is where the objection quietly becomes easier.
- Tag each one with which of the four kinds it really was, judged by what happened next, not by how it sounded. You will find that a third of the ones you filed as price were authority.
- Record what you actually said and whether it worked. Not the answer you wish you had given. The library is only useful if it is honest about the ones you fumbled.
- Keep the ones you lost. The objection that ended a deal is the highest-value entry in the file and the one every battlecard omits.
- Review it monthly with one colleague, and let them ask the unscripted follow-up on the three you feel most confident about. Confidence is where the bluffs live.
Twenty real objections in your buyers' own words beats any deck. And it feeds b8: if you are logging the objection type on the deal record, b9's forecast gets to know the difference between a deal that stalled on timing and one that stalled on a person nobody has met.
The roleplay drill ★ 22 min · everyone runs it
Bring your worst objection. Not the one you have a good answer for - the one you quietly hope does not come up. We run it through the loop four times, and the exit condition is the rail: it survives a follow-up you did not script. Pair up for step five.
Minute 0-3 · write the objection in the buyer's words. Verbatim, from a real call if you have one. If you find yourself tidying it up, you are already avoiding it.
Minute 3-6 · classify it before you answer. Price, timing, authority or trust. Use the four tests. Write down which one, and be ready to be wrong at minute 20.
Minute 6-10 · draft your answer in three lines and say it out loud. Out loud matters. Answers that read fine collapse audibly.
Minute 10-16 · run the simulated buyer below. It plays Priya Raman with a real budget constraint and it pushes back twice. Answer live, in the chat, in your own words. Do not paste a prepared answer.
Minute 16-20 · the human follow-up. Hand your answer to the person next to you and ask them for one question you did not script. This is the step that finds the bluff.
Minute 20-22 · rewrite and re-classify. Keep only what survived. Then look again at what kind of objection it really was - most rooms find at least one price that turns out to be authority.
You are Priya Raman, CFO at Northwind Logistics, a 3PL in Singapore with about 900 staff. Facts you know and must not contradict: - Northwind signed a new Jurong DC lease five weeks ago. - Three warehouse-ops reqs are open and unfilled. - You have a headcount freeze until the Q3 close. - You were NOT on the discovery call. You have heard about this from Mira Chen, your VP Ops, in about three sentences. - You already pay for a suite that includes a note-taker. How to play it: - You have 15 minutes and you think this is a nice-to-have. - Push back at least TWICE on whatever I say, and get more specific each time - the second push should be the question I would least like to be asked. - Do NOT agree, and do NOT book a next step at the end. - Ask me about anything you do not know. Invent nothing about Northwind beyond the facts above. - Stay in character. No summaries, no coaching, no "great point". Open by asking me why this costs three times what you pay today.
What is the best rebuttal to a price objection? Give me a few options I can use.
The second prompt is the most common AI-in-sales request there is, and it is the one that quietly makes reps worse. It produces four fluent rebuttals with no account, no constraint and no diagnosis, so you arrive at the call with an answer looking for an objection. Meanwhile the real question was never about price: Priya was not on the discovery call, she has a freeze until the Q3 close, and no sentence delivered through Mira fixes either of those. Gartner's 75%-by-2030 prediction is the same point from the buyer's side. They are not short of polished answers. They are short of someone who tells them the truth about which problem this actually is.
Try it yourself - this week ◐ 40 min total
- Write your three worst objections in the buyer's own words, pulled from real recorded calls where you can. Verbatim, untidied.
- Rehearse all three against a simulated buyer using the persona prompt: a named person, a stated budget reality, and instructions to push back twice and never agree.
- Then have a colleague ask you the one follow-up you did not script, on each of the three. Note which answers survived and which collapsed. The collapse is the useful result.
- Write down which of the four kinds each objection really was - price, timing, authority or trust - judged by what actually happened next in the deal, not by how it sounded. Count how many turned out to be authority.
- For every one that was authority, name the person who is missing and write the one line that gets them into the room.
- Start the objection library: twenty objections from your own calls, each tagged with its kind and what you actually said. Keep the ones you lost.
- Answer the course question: rehearsal gave you back the practice hours you never had. Did you spend them getting more fluent, or getting more honest?
Official sources covered
This session covers the buyer-preference research that sets the rail on rehearsal, plus the working craft of objection diagnosis and honest competitive positioning. Claims about named competitors are a legal question as well as a sales one - the rule on this page is conservative on purpose. This page covers:
Three questions before you go 🎯 ◐ 90 seconds
1 · You have rehearsed the security objection eight times and your answer is smooth. When is it actually ready?
A simulated buyer that accepts your answer is flattering you, and eight smooth runs is exactly how a rep ends up confidently wrong about where the data is stored. The rail is one unscripted follow-up. If the answer does not survive, the answer is the problem, not the question - and "I do not know, I will find out today" is a defensible answer.
2 · Mira forwards you one line from Priya, the CFO who was not on the discovery call: "why is this three times what we pay now?" What kind of objection is it?
It arrived second-hand from someone you have never spoken to, which is the clearest authority signal there is. Answering the price question by email answers the wrong question well and makes your champion defend a number she did not set. The move is to get Priya in the room, while still giving Mira one honest line she can forward if you do not.
3 · A buyer asks how you compare to a competitor's bundled note-taker. What must you never do?
A model asked for a competitor's weaknesses will answer fluently and invent some of it. Invented weaknesses about a named company are a legal and trust problem, and the buyer often knows that product better than you do. Use AI to find and cite what a competitor publicly claims, then read it yourself and decide what is true. Finding, not writing - the same rule as b3.