How to analyse your own negotiations and stop repeating mistakes

How to analyse your own negotiations and stop repeating mistakes

During a conversation you are working out what to say next and miss half of what the client said — not through inattention but because you are busy. A review of the recording finds it, and your own mistakes along with it. Below is a real review with two of mine.

During a conversation you are working out what to say next and miss half of what the client said — not through inattention but because you are busy. A review of the recording finds it, and your own mistakes along with it. Below is a real review with two of mine.

---

The call happened, and what now is unclear

A familiar feeling: you talked, it seemed fine, the client did not say no. You put the phone down and cannot say anything more definite than "seemed okay".

What worked? Where did you lose them? What did they actually want with that odd question in the middle? Two days later you remember half of it, and the decisions have to be made now: what price to set, what to put in the proposal, what to press on.

And crucially — you will run the next call exactly the same way. Experience does not accumulate by itself. It accumulates only where there is a review, and almost nobody can review themselves honestly: the brain remembers the good moments and deletes the awkward ones.

So I do not try to remember my calls. I hand them over for review.

What I do straight after a call

I record the call, upload the audio to the sales assistant and write one sentence: "here is the audio of a call with this client, analyse what could be improved and how I should carry the conversation on."

What follows is what it produced for the conversation with Zakhar Fomin, owner of a dental clinic. Published as is, including my mistakes. A review where everything went perfectly is useful to nobody.

Why your own memory is no good for a review

The root cause: during a conversation you are busy having it. Attention goes on listening, answering and not falling silent. There is no resource left for observing the structure — and that is normal, it works that way for everybody.

The second reason is less pleasant: you hear what you are ready to hear. The client drops a sentence that is actually the main lever of the sale — and you miss it because at that moment you were working out what to say next.

And the third: your own mistakes are invisible from inside. Not because you are deceiving yourself but because you do not know what to look for. To notice "here I blurred the focus" you have to be holding in mind what the focus should have been.

A machine has none of that. It listens to the whole recording, evenly, with no nerves and no ego.

What the assistant pulled out of the client's words

This is the most valuable part of the review, and it is not about me but about the client. Zakhar handed over his own markers of lost revenue — and I did not register them in the conversation.

The fear of "stupid bots" on expensive patients. He said outright: leaving a high-value client to talk to a bot is a dubious business. His real pain is not the money for the build but that AI will damage the reputation and lose an expensive implant patient.

*How to answer:* this is not a linear answering machine, it is a business-class concierge. It recognises the status of an enquiry, qualifies tactfully and hands over to the owner's personal phone when it sees a hot, high-value request.

Reactivating the base by hand. Once a year Zakhar goes through the calendar manually, finds old patients and asks the doctors to ring them. Or hires a remote administrator who works an hour a day.

*How to answer:* show how the bot scans the base itself, sees somebody has not been for six months and sends an invitation with the option to take a slot in one tap. That is pure payback and a cost line removed.

Dead social media. One patient every two months from Instagram. They paid a specialist who burned budget because the content was not tied to a funnel.

*How to answer:* content is not about pretty pictures but about targeted traffic that lands straight in the funnel and sells.

Three pains, three ready blocks of a future proposal. I registered none of them in the conversation as a lever.

What it said about me

Strengths

Weaknesses

First: mid-call I drifted into technical questions about assistants and blurred the focus away from business metrics.

Here I half agree with the review. I did it deliberately: it mattered to me to understand how Zakhar's patient journey runs from first contact to the chair, because my second step was to sell him a CRM, and that is a separate project worth several thousand dollars. But for this conversation the comment is correct: it was about something else, and I did blur the focus.

Second: I did not press the metaphor home when Zakhar raised his fear of bots.

That is a clean mistake and I have no objection. I should have said, right then, "this is not a replacement for a doctor, it is a smart concierge" — in his language, with no technical terms. I did not, and the fear stayed unaddressed to the end of the conversation.

Opportunities and threats

The opportunity: Zakhar is willing to pay well if he sees the system can work the existing base for repeat visits. That is his clearest lever of profit.

The threat, stated plainly: name a price off the top of your head tomorrow, with no link to payback and to where he saves, and the Controller will switch into Analyser mode and take time to think. And "think" with that type means weeks.

That warning determined how the proposal got built.

How this turned into money

A review is not an end in itself. The three pains the assistant pulled from Zakhar's words went into the proposal one per block: the fear for expensive patients closed by a slide about hot enquiries going to the owner, the manual reactivation by a slide about automatic scanning of the base, the dead social media by a block about content tied to the funnel.

The warning about a price off the top of the head worked too: the figure in the proposal did not stand as a separate line but next to a calculation of what an in-house team costs. The Controller saw not a rate card but a comparison — and did not go away to think.

Why it is worth showing your own mistakes

Worth saying separately, because it is about the approach rather than the tool.

A review where an expert with five years' experience publishes two of their own mistakes works harder than one where everything went perfectly. A perfect case helps nobody: it cannot be repeated and there is nothing to measure against. A mistake is a ready checklist.

And a second, more practical reason. The assistant sometimes pulls out things you would not notice even with substantial sales experience. Not because it is cleverer but because it listens to the recording without being in it.

What to do if the call cannot be recorded

It happens: the client objects, or the conversation was in messages, or you simply forgot.

Then straight after the call, while you remember, write out freely everything the person said about their situation: figures, complaints, staff names, sums, timescales. Do not structure it — just pour it out. Fifteen lines is enough.

Then give that text to the assistant with the same request. The review will be weaker than from audio: the intonation and pauses are lost, and it is often in those that you hear where somebody tensed up. But the three main pains come out of a summary like that perfectly well.

For messaging it is simpler still: it is already recorded. Screenshots of the exchange get submitted for review exactly like audio.

How to build this into your work

The procedure is simple and takes about fifteen minutes after a call.

1. Record your calls. Having told the other person — that is normal practice.

2. Upload the audio immediately to the assistant, asking for a review and for how to carry the conversation on.

3. Read the block about the client first, not the one about you. That is where the levers for the proposal are.

4. Then the block about you — and write out one mistake to hold in mind on the next call. One, not five.

The sales assistant is one of more than 140 AI assistants across 14 categories: sales, marketing, legal, HR, analytics, creative, media, engineering. It will also draw up the plan for the next call and play the client in a simulation if you want to run the conversation in advance.

Alongside is the ten-lesson sales module: objections, rejections, burnout, the four client types, shaping an offer around demand.

Registration is free and opens three days of full Basic access — the sales module, all the assistants, the prompt texts and six foundational express courses.

Start with your next conversation. Even if it goes badly — especially if it goes badly. Fifteen minutes of review after a poor call gives you more than five good ones you remember nothing about.

If you know somebody who says "seemed to go okay" after every call, forward this to them.

---

What to review a recording with

Three tools, covering different parts of the task.

MyMeet — turns an hour-long call into text. That is the first step: from there you work with a transcript that reads in ten minutes rather than an hour.

The sales assistant — takes the transcript and produces a psychographic portrait of the decision-maker: hidden fears, real motivators, pains. And separately analyses your side of the conversation.

The conversation analyst from the automations archive — for when there are many conversations and they run in messages. Text gets broken down into script errors and client pains; voice messages get transcribed automatically. The summary arrives in the form "the client doubts the price, the salesperson did not address the discount".

What the third tool adds beyond the review. It works continuously rather than on your request. You see the weak spots immediately rather than when you decide to revisit a recording.

And it sells as a service: deploying a conversation analyst into a sales team costs $360–1 200.

---

The series: one deal step by step

Eight pieces about one real deal — from a listing in a chat to the payment received. The client, Zakhar Fomin, owner of a dental clinic, gave permission to use his name and the figures.

---

What to read next

[A commercial proposal in ten minutes](/en/blog/proposal-in-ten-minutes) — what the review turns into.

[The fifteen-minute client call](/en/blog/fifteen-minute-sales-call) — the script for the conversation itself.

[Transcription and file analysis](/en/blog/transcription-and-file-analysis) — the technical side of working with recordings.

[An X-ray of your calls: auditing a sales department](/en/blog/sales-calls-audit) — the same skill sold to a company.

[AI API pricing in 2026](/en/blog/llm-api-pricing) — what a request costs and how to work out your own costs.