Open the call with the prospect’s own AI answer, signed out. A forty-prospect quarter by hand costs EUR 1,800 to EUR 3,600.
You have a call booked on Thursday with a named account and forty minutes to prepare for it. You need one opening sentence the prospect has not heard forty times this month, and you need something to send afterwards that is about them. The first ten names on the list get that treatment. By name eleven the prep time is gone, the opener collapses back into “I noticed you’re in aviation”, and the follow-up goes out carrying the same PDF everyone else got.
Meanwhile your prospect’s own buyers have started asking AI before they ask anyone. So there is one specific thing worth knowing before you dial: whether AI names your prospect when their buyer asks the category question, and which of their rivals it names in their place. Nobody in their marketing team can tell you, which is exactly why it opens a conversation.
What it costs today
Sales time loaded at EUR 45 an hour.
| Line | Rate | Time | Cost |
|---|---|---|---|
| Pre-call research that yields one specific opener | EUR 45/h | 30 to 60 min | EUR 23 to EUR 45 |
| The tailored leave-behind after the call | EUR 45/h | 30 to 60 min | EUR 23 to EUR 45 |
| A 40-prospect quarter, both ends | EUR 45/h | 40 to 80 h | EUR 1,800 to EUR 3,600 |
Most reps skip the second line entirely, which is why the follow-up is a generic PDF.
When the claim has to survive their internal meeting after your call, the bill moves from sales time to analyst time. Tracing one buyer question properly means running it on three models, following every brand each answer names, and logging which of those brands hold up on the next hop. At EUR 90 an hour that is two to four hours a question, so EUR 180 to EUR 360. A twenty-question set runs 40 to 80 hours and EUR 3,600 to EUR 7,200, one to two weeks of one analyst. Half of that work cannot be bought at any price, because no consumer chat window shows you the searches a model fires before it writes its answer.
The workflow
Three playbooks, in the order a rep actually uses them.
1. Open a cold call with the prospect’s own AI answer · 5:34 · no account
Start on a public page, signed out. Three fields: the prospect’s company, what they sell, their city. Rankfor.AI suggests the questions their buyers ask, in their market’s language, and puts the one you pick to Gemini, ChatGPT and Grok. A couple of minutes later you have a deck showing which brands each model named, who owns the citations behind the whole answer, and where your prospect sits on one track running from Invisible through to Recommended. That position is your first sentence. One free check per address per week, so spend it on the name that matters most that week.
2. Walk into the call with their answer already traced · 6:29 · any plan, including Free
Once you have a seat, you stop rationing the check. Type the prospect’s brand straight into the Answer Trail brand field from whatever project you already have open. The field is free text, nothing about them has to be set up or scanned first. About a minute later three models have answered, and every brand in the answer carries a verdict from a fresh search: Survived, Dropped, Emerged or Absent, with the sources behind each one. The cross-model screen is the one you share on the call. Then mint a share link, tag it with the deal name so each account reports separately, and send that in place of the one-pager you were never going to write.
3. Follow one buyer question across three AI models · 6:29 · any plan, including Free
Same run, read all the way down, for the moment your champion says they checked ChatGPT themselves and looked fine. Same question, one model over: a different set of searches, different sources, a different set of brands left standing. This playbook also opens the citation split behind every verdict on the Signals bar: other independent, review, tier-one news and brand-owned, so you can show a prospect when a model is mostly quoting them about themselves. Tick the gaps the models’ own searches exposed and each becomes a content plan their marketing lead can pick up. That is the second meeting.
Playbooks 2 and 3 run on the same screens, so the second one is a quicker watch.
What this does not do
It does not work your list for you. One run per project at a time, and a three-model comparison counts as one run, so a rep on Free traces one prospect at a time. The public page allows one free check per address per week. Nothing here runs on a schedule, and nothing messages you when an answer changes. You run it when you prepare for a call.
Only half the quarterly number comes off the bill. The research half shrinks to a one-minute run plus the few minutes you spend reading it. The leave-behind half is a swap: the link replaces the one-pager wherever the account will open a link, and where the buyer wants a document in the follow-up, that half of the number stays exactly where it is.
Three models, and only three. Gemini, ChatGPT and Grok. No Claude, no Perplexity, no fourth model on any screen in this set.
We measure what the models say. We do not promise a position. Nothing here moves an answer on demand, and nobody can promise you one. A content plan produced from a run is an outline plus SEO, so do not tell a prospect they are getting finished pages.
Start here
Watch Open a cold call with the prospect’s own AI answer. It is 5:34, and you can run the thing itself in about the time it takes to watch: no account, no card, nothing to install.
Put the name at the top of your list into it and read where they land on the track. If they land on the left, you have your opening sentence and you spent nothing finding it. Then come back for the other two, which need a seat and give you the version you can run every week.