A battlecard is a document a rep is supposed to consult in the three seconds between a prospect saying "we already work with someone" and the rep answering. Nobody does that. Real-time AI coaching replaces the document with a one-line prompt, triggered by the objection itself, written from the answer that worked best on your own team's calls last week. Here is why the static version fails, what a live prompt should look like, and how your top performers keep it current without doing anything extra.

Why static battlecards fail on a live call

Three reasons, and none of them is laziness.

Cognitive load. A rep on a cold call is listening, deciding and talking at once. Opening a PDF and finding the right section is a fourth task. The pause it creates is audible, and in our experience a two-second silence after an objection is where most French prospects say "bon, envoyez-moi un mail" and hang up.

Decay. The card was written in January by enablement. A competitor changed pricing in March, your own packaging changed in May, and the objection your reps now hear most ("we're testing an AI SDR") did not exist when the card was written.

Genericity. Cards are written for the average objection. Reps hear specific ones, in a specific market. "On a déjà un prestataire" from a Lyon industrial SME is not "we have a vendor" from a San Francisco SaaS, and the answer that works is not the same.

What a live prompt should look like

A live prompt is coaching delivered during the call, on screen, triggered by what the prospect just said. Three rules make it usable.

1. One line. Under 15 words. The rep reads it while the prospect finishes the sentence. A paragraph is a battlecard with extra steps.
2. Triggered, not scheduled. It appears when the transcript matches a known objection, a competitor name or a pricing question. Not every ten seconds, not on a timer. Three or four triggers per call type is the ceiling before reps close the panel.
3. Drawn from a real call. The line is a paraphrase of what a top rep said in the same situation, on a call that booked. Not a copywriter's best guess.

Example, for "we already have a vendor": the prompt reads "Ask what they'd change about it before you pitch". That is the whole prompt. It encodes the one behaviour our benchmark of 1,000,000 B2B cold calls in France rewards most on objections: asking "What do you mean?" or a variant before answering lifts conversion by 197%.

How top performers train the prompts: the loop in five steps

The point of a feedback loop is that the content updates itself from the calls, so enablement is not rewriting PDFs every quarter.

1. Hear. Every call made through the dialer is transcribed. Objections are tagged by type: timing, incumbent, budget, authority, "send me an email".
2. Score. Each objection moment is scored on what the rep did and what happened next. Did the rep ask first? Did the call continue past 60 seconds? Did it book?
3. Promote. For each objection type, the answers with the best downstream outcome over the last two to four weeks are identified. If three reps asked "what would make you switch?" and two of those calls booked, that pattern wins.
4. Prompt. The winning pattern becomes the live one-liner for that objection, for everyone. The rep who found it gets credit in the Monday review, which is the cheapest motivation budget you will ever spend.
5. Drill. New objections that appeared this week, and the ones the team scored lowest on, become tomorrow's roleplay scenarios. Reps rehearse the real thing before the 10am block, not a library persona.

Run that loop weekly and the battlecard is never stale, because it is not a document any more. It is a ranking of answers, refreshed by outcomes.

The French objection set that belongs in your loop

From the calls we analysed, five objections cover most of what a French B2B SDR hears. Each needs its own trigger and its own prompt.

• "Envoyez-moi un mail." The default polite exit. Prompt: ask what they would want the email to answer.
• "On a déjà un prestataire." Prompt: ask what they would change about it.
• "Ce n'est pas le moment." Prompt: ask what changes at the moment it becomes the moment.
• "Ce n'est pas moi qui décide." Prompt: ask who, and what that person cares about.
• "On n'a pas de budget." Prompt: ask what the current solution costs them, in time.

Notice that every prompt is a question. 65% of reps open calls by pitching, and the same reflex shows up on objections. The loop's job is to keep surfacing the reps who ask first, because their calls book.

Guardrails: calibration and consent

Two things break live coaching in Europe if you skip them.

Calibration. Start with three triggers, measure how often reps dismiss the prompt, and only add a fourth when dismissal is under 20%. A prompt that fires on every "hmm" is surveillance with a UI.

Consent. The loop runs on transcripts, and transcripts are recordings. Under GDPR, recording and prompt announcement must be configurable per country, prospects must be informed, and deletion on request must be possible. Configure it before the first call, not after the first complaint.

How Sonik approaches live coaching

Sonik runs this loop inside the dialer. Calls are transcribed in French and English, objections are tagged, and each objection moment is scored on what the rep did and whether the call booked. The live prompt shown for an objection is derived from the answers that scored best on your team's own calls over the previous weeks, so it changes when your top reps find something better.

Prompts are one line and fire on a small set of triggers you control. New and low-scoring objections feed the roleplay scenarios automatically. Recording consent is set per country and per rep, and the score and summary are written to the CRM on call end.

We do not ship a battlecard library. In our experience the only answers reps trust are the ones a colleague used on a call that booked.

Conclusion

A static battlecard answers the average objection, late, from a document nobody opens. A live prompt answers the specific objection, in two seconds, with what worked for the rep next to you last week. This week, pull your ten most recent calls where a prospect said "on a déjà un prestataire", note what each rep said next and which calls kept going, and write the single best answer as a 15-word line. That is your first prompt, and your first battlecard that will not go stale.