Hook: Type a prospect's brand into Answer Trail and read what three AI models say about them in about a minute, then send the same run as the follow-up. A forty-prospect quarter costs EUR 1,800 to EUR 3,600 of your own time today.
What You'll Be Able To Do
- Trace any named account from a project you already have, with nothing about them scanned first
- Open an opening sentence backed by what three AI models said when they checked their own work
- See the searches each model fired before it wrote a word, and the pages it read
- Answer "I checked ChatGPT myself and we were fine" with the same question one model over
- Mint a link the prospect opens with no account, tagged with the deal name so each account reports separately
Runtime: 6:29 | Difficulty: Advanced
You have a call booked with a named account and forty minutes to prepare. You need one opener the prospect has not heard forty times this month, and something to send afterwards that is about them. It goes wrong at prospect eleven. The research time is gone, the opener collapses back into "I noticed you're in aviation", and the follow-up carries the same PDF everyone else got.
What it costs today
An account executive 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 |
Only one of those halves is a saving. The research shrinks to a run of about a minute plus the few minutes you spend reading it. The leave-behind is a swap: you send a link and write nothing, so those 30 to 60 minutes come off the bill wherever the account will open a link. Where the buyer wants a document, that half of the number does not move.
What the video shows
- Answer Trail in the sidebar, with the Models block switched from "One model" to "Compare models". Three chips appear, Gemini, ChatGPT and Grok, all ticked.
- The prospect's buyer question goes into the prompt field, and the prospect's brand into the brand field. It is free text, so the name is enough.
- Region goes in below it, and the persona picker stays on "General audience", because the personas come from your own project's scan.
- Run comparison. About a minute of waiting, three cards filling in, each naming how that model got to its answer.
- The Gemini trail: every search the model fired before it wrote a word, badged as an intent query or a site probe, over the pages it read.
- The brand verification block. Each brand gets a verdict from a fresh search, Survived, Dropped, Emerged or Absent, with the citations behind it.
- ChatGPT, then Grok. A different fan-out, different sources, a different set of brands still standing.
- Cross-model: one presence score for the prospect, then one row per brand cut into one segment per model, red for absent and green for survives.
- The Prompts tab, where the models' own searches become the pages the trail says the prospect should own. Tick two and press Produce.
- The share panel: public or PIN-locked, the deal named as a campaign, then copy the link.
- The closing frame is that shared page as the prospect sees it, signed out, with no sidebar and no account.
What you need
- Any plan, including Free. Answer Trail carries no tier gate and needs no completed scan of the prospect.
- A project of your own with a finished scan, because the persona list is filled from it. Free and Marketer allow one project per workspace; Strategist and Enterprise three.
- Member access to that project. A share link minted from outside it is refused.
- The prospect's buyer question, commercial in shape, in the language of their market. The prompts that come back follow the language of the question.
- A prospect you are comfortable naming, because the brand and the question print on the shared page.
Questions people ask
Do I need the prospect set up as a project first? No. The brand field takes any string, and Answer Trail forces a verification pass on whatever sits in it. Type their name in from whatever project you already have open. Nothing about them has to be scanned first.
Which AI models does it check? Gemini, ChatGPT and Grok. Three, and only three. The whole group counts as one job in the queue, so three models cost the slot of one.
Does picking a persona change what the models answer? No. The run sends your prompt as written. The persona rides into any content plan you produce from that run, so pick it before you press run.
Can I trace ten prospects at once? No. One run per project at a time, so a rep on Free works a list one prospect at a time. And a produced plan is an outline plus SEO, so do not promise finished pages.
Watch it once, then run it on the account you have booked for Thursday.