I Lost a $40K Deal Because of My English. Here's What I Changed.
I Lost a $40K Deal Because of My English. Here's What I Changed.
I want to tell you about a specific call. Not a composite, not a "story that represents many experiences." A specific call, on a specific afternoon, that changed what I spent the next eighteen months building.
The Call That Changed Everything
It was a discovery call with a VP of Sales at a mid-size SaaS company in Austin. I had been building a pipeline in the US market for about seven months. My English was good — or I thought it was good. I had been in London for two years before this. I could write emails in English that sounded native. I could present in English without visible difficulty.
This call was different. It wasn't a presentation. It was a conversation, and about twelve minutes in, the VP asked a question I wasn't expecting.
He said: "Look, we're also looking at [Competitor]. What would you say is the core reason you'd win that evaluation?"
I knew the answer. I had given a version of this answer dozens of times. In French, I could have said it in fifteen seconds without breaking stride. In English, in that moment, with the silence opening up after his question, the answer — which I knew — became inaccessible. Not partially. Completely.
What came out of me was something like: "So, I mean, there are several aspects that, you know, we think are quite important in terms of the differentiation, and basically what we would say is that our approach is..."
He waited. I could hear him waiting. I kept talking. I used more words to say less than I had said in any sales conversation in the previous year. By the time I arrived at a version of my actual point, the energy in the call had shifted. He was polite for the remaining eight minutes. The deal never moved again.
What Actually Happened
I spent a long time afterward analyzing that call because I knew it hadn't gone wrong for the reason most failed calls go wrong. It wasn't a product fit issue — the company had a real pain we addressed well. It wasn't pricing — we hadn't gotten to pricing. It wasn't competition — I had a good answer to the competitor question. It was the twelve-second gap between his question and me beginning to answer it in a way that sounded like I knew what I was talking about.
There was a specific moment when I said "proposition" instead of "proposal." I caught it as I said it but couldn't unsay it. There was a moment when I said "basically" three times in two sentences. There was the freeze — the three seconds of visible searching for language that signaled, to a VP who had done a hundred of these calls, that the person on the other end of the line was not in command of the conversation.
The VP of Sales in Austin was not evaluating my English. He was evaluating whether he would trust this person to sell him a $40,000 solution. And what he saw in those twelve seconds was someone who, under pressure, lost access to the language they needed. In his experience, that pattern doesn't get better under the greater pressure of an implementation or a renewal conversation.
He was probably right.
What I Tried First
The first thing I did was sign up for a Berlitz corporate English program. It was expensive — the equivalent of several hundred euros for a package of sessions — and what I got was a kind, qualified instructor who helped me work on pronunciation and wrote corrections on phrases I used incorrectly in our sessions.
It did not help with the call. In our Berlitz sessions, I was not under pressure. I was in a supportive environment where the worst that could happen was I said a word wrong and my instructor gently corrected me. The anxiety that locked my language in the Austin call did not appear. So I got better at English in conditions that did not exist on my calls. The improvement did not transfer.
The next thing I tried was YouTube. There are genuinely excellent channels on English for business professionals — I watched dozens of hours. I found videos specifically on sales English, on handling objections in English, on sounding more American in a business context. I took notes. I practiced phrases in my apartment.
The problem with practicing phrases in your apartment is that there is no pressure. You are not being evaluated. The stakes are zero. And when you get on an actual call and the stakes are real, all those apartment phrases are less available than you'd expect, because the retrieval conditions are completely different.
I tried practicing with friends who spoke English. This was more useful than the apartment practice — there was at least a person on the other side. But my friends knew I was practicing, knew I wasn't actually trying to close them on anything, and could not simulate the specific dynamic of a US buyer who is deciding whether to give you twenty more seconds or wrap up the call. My friends were kind. Kindness doesn't replicate call pressure.
I was spending time and money on things that felt like they should work but weren't addressing the actual problem.
What Finally Worked
The thing that finally worked was pressure.
Not high stakes pressure — the opposite of productive. Simulated pressure. The pressure of a realistic conversation where I didn't know what would come next, where the "buyer" wouldn't give me a pass on hesitation, where I would have to respond in real time to questions I hadn't anticipated.
I started by building something rough: a simple flow that put me in a conversation with an AI configured to play a skeptical US buyer. I set it up so the AI would not accept vague answers, would push back on hedging language, and would end the conversation if I spent more than three seconds searching for words without recovering.
The first twenty sessions were uncomfortable. I froze. I said "basically" six times in five minutes in one session. I used "proposition" instead of "proposal" again — and caught myself, in real time, which was the first sign of progress. But I kept running sessions. Every morning. Thirty minutes before I did anything else.
By the fifth week, something changed. When the simulated buyer asked the competitor question — the question that had ended the Austin deal — I responded within two seconds with a direct, clear answer. Not because I had memorized the answer. Because I had answered it so many times under simulated pressure that the retrieval path was automatic. The anxiety still fired. I just no longer needed to not be anxious to access the language.
That's the distinction that matters. Not becoming less anxious. Becoming functional while anxious.
Why I Built DojoSales
After those first sessions, I had one clear insight: the thing that was making me better was repetition under pressure with feedback. Not language classes. Not workshops. Not YouTube. The repetition was building a neural path from the anxiety to the language that was fast enough that the anxiety couldn't block it.
I also had a second insight: I wasn't the only person with this problem. Every French sales rep I knew who was working a US market had a version of the Austin story. Some of them had lost smaller deals. Some had lost bigger ones. The common thread was not that they didn't know English — they all did. It was that they couldn't access it consistently under the specific conditions of a live sales call with a US buyer.
The training infrastructure that existed was either too generic (language platforms built for travelers, not sellers), too expensive (Berlitz, human coaches), or missing the language layer entirely (AI sales platforms built for native English speakers). There was no product built specifically for the non-native sales rep who needed call practice, at scale, with feedback on both the sales mechanics and the language patterns.
That gap became DojoSales. AI buyer personas calibrated to US and UK sales contexts, difficulty levels that range from friendly to hostile, and a feedback engine that does not give generic scores — it points to specific moments in the transcript, the exact phrase that leaked uncertainty, the timestamp where the freeze happened, the sentence that should have been a direct statement and wasn't.
The price is $39 a month. Berlitz cost me more than that per session.
The First User Who Told Me "I Didn't Freeze Today"
About three months after the first users started using DojoSales, I got a message from a user in Paris. She was an AE at a Series B company selling into US mid-market. She had been using DojoSales for four weeks.
Her message was one sentence: "Je n'ai pas freezé aujourd'hui."
I didn't freeze today.
She had just gotten off a call with a VP in Seattle. An unexpected technical question had come up. In her previous calls, a question she hadn't prepared for would produce the same cascade that happened to me in Austin — the visible search, the filler density spike, the loss of command. In this call, she had paused, said "that's a good question — let me make sure I'm answering the right thing," and responded clearly.
The pause was a technique she had practiced fifty times in DojoSales sessions. The bridging phrase was one she'd used repeatedly in simulations until it was automatic. She had not eliminated the anxiety — she had trained it out of her response.
That message was the validation I needed that the problem I had experienced was not mine alone, and that the thing I built to solve it was working.
What You Can Do Right Now
If you are a non-native English speaker selling into US or UK markets, and you have a version of the Austin story — a call you know went wrong not because of product fit but because of language under pressure — here is the shortest path to a different outcome.
Stop the language classes. Not permanently — but recognize that they are not addressing the variable that matters. The variable is your language access under call pressure, not your language knowledge in a classroom.
Start practicing under simulated pressure. This does not require DojoSales specifically. It requires a realistic simulation with someone — human or AI — who will not give you a pass on hesitation, who will respond to your actual performance rather than your best-case performance, and who will do this at the volume you need (which is more than twice a week).
If you want the fastest path: DojoSales is 30 minutes a day, AI buyer personas calibrated to US market conditions, feedback at the transcript level, $39 a month. It is what I wish had existed when I lost the Austin deal.
The Austin VP was not wrong to disengage. I was not ready. Two months of the right practice, and I would have been.
For the full framework on sales English skills, see The Complete Guide to Sales English.