AI Search Visibility Agent Team · Version 2.0
Four agents that check whether AI assistants mention your brand, work out why they do not, write the pages that fix it, and tell you the following week whether anything moved.
This is version 2.0 of the Grok Bot pack. The seven roles from the first video are still in here, arranged into the loop from the new one. Forty files of plain Markdown and CSV.
Each brief is written to be pasted into a chat window, a custom GPT, or a coding agent as its instructions. Every one names its inputs, its outputs, the agent it hands to, and the ways it fails.
Runs a fixed set of customer questions against the AI engines you choose, every day, and logs whether you were mentioned, recommended, or cited, which competitors came up instead, and which sources the answer leaned on.
Carries forward Iris and Sable from v1.1.
Reads the week of logs and diagnoses each gap from the evidence rather than guessing. Writes ranked content briefs, half of them against questions with buying intent.
Carries forward Marlow, with Hollis and Vera as inputs.
Writes one piece per brief against a checklist: answer first, the entity stated plainly, every claim sourced, the questions people actually ask answered in full. A human reviews before anything is published.
Carries forward Quill and Ansel.
Compares this week against last week for the questions you targeted, marks each one moved, unmoved, or regressed, checks the questions you did not target as a control, and hands the Strategist three recommendations.
New in 2.0. No v1.1 equivalent.
One email address, and the download appears on this page straight away. You do not need to go and check your inbox.
Would rather not give an address? The original seven-role pack from the first video is still free and ungated at the GrokBot page, exactly as promised in that video.
Plain Markdown and CSV. Nothing compiled, nothing to install, no account to connect. Read any of it in a text editor and edit it to fit your business.
| START-HERE.md | what this is, and the ten-minute first run | 5.2 KB |
| BRAND-CONTEXT.md | the one file you fill in | 3.0 KB |
| SHARED-RULES.md | rules every agent obeys | 3.2 KB |
| ROUTINES.md | what runs daily, weekly, per piece | 4.3 KB |
| MEASUREMENT.md | mention, recommendation and citation rate | 6.7 KB |
| LOG-GUIDE.md | how to fill every column | 13.1 KB |
| FIRST-RUN.md | the first run, walked through slowly | 5.3 KB |
| SOURCES.md | where the method came from | 3.3 KB |
| CHANGELOG.md | everything that changed from v1.1 | 4.4 KB |
| LICENSE.md | yours to use, not to resell | 0.9 KB |
| team/ | the four agent briefs | 27.9 KB |
| 00-TEAM-MAP.md | the loop, and where the seven roles sit | 4.0 KB |
| 01-monitor.md | agent 1 | 5.7 KB |
| 02-strategist.md | agent 2 | 6.6 KB |
| 03-producer.md | agent 3 | 6.5 KB |
| 04-optimizer.md | agent 4 | 5.1 KB |
| roles/ | the seven v1.1 specialists, carried forward | 13.2 KB |
| 01-marlow.md | chief of staff | 2.1 KB |
| 02-iris.md | diagnostic | 2.1 KB |
| 03-hollis.md | crawl auditor | 2.0 KB |
| 04-vera.md | listing accuracy | 1.6 KB |
| 05-ansel.md | review responses | 1.8 KB |
| 06-quill.md | website content | 1.8 KB |
| 07-sable.md | sentiment watch | 1.8 KB |
| checklists/ | what gets filled in as you work | 10.1 KB |
| PRODUCER-CHECKLIST.md | one pass per piece, before review | 4.8 KB |
| WEEKLY-OPTIMIZER-REVIEW.md | the weekly before and after | 3.5 KB |
| CRAWL-CHECKLIST.md | can the engines fetch your pages | 1.7 KB |
| prompts/ | the question set | 9.8 KB |
| 50-prompts.csv | the original 30, plus 20 new | 9.8 KB |
| trackers/ | blank, with headers ready | 3.3 KB |
| prompt-tracker.csv | one row per question, per engine, per day | 0.3 KB |
| production-checklist.csv | the checklist as columns | 0.5 KB |
| weekly-progress.csv | what was produced, by whom, reviewed when | 0.1 KB |
| targets.csv | twelve weeks, target against actual | 1.9 KB |
| observation-log.csv | the v1.1 diagnostic log | 0.4 KB |
| action-queue.csv | what to do about what you found | 0.2 KB |
| examples/ | the same files, filled in for a made-up business | 16.2 KB |
| BRAND-CONTEXT-EXAMPLE.md | a completed context file | 4.9 KB |
| CONTENT-BRIEF-EXAMPLE.md | what the Strategist hands the Producer | 2.4 KB |
| targets-EXAMPLE.csv | twelve weeks, filled | 3.8 KB |
| prompt-tracker-EXAMPLE.csv | a week of logged answers | 1.4 KB |
| production-checklist-EXAMPLE.csv | three pieces, checked off | 1.4 KB |
| observation-log-EXAMPLE.csv | a completed diagnostic | 1.3 KB |
| action-queue-EXAMPLE.csv | findings turned into jobs | 0.6 KB |
| weekly-progress-EXAMPLE.csv | one week of output | 0.4 KB |
No tools to connect. Use the AI apps you already pay for.
Open BRAND-CONTEXT.md. Enter one brand, its category, its market, three competitors, and five questions your customers actually ask. That file feeds every agent.
Take three prompts from the 50, replace the brackets, and paste each one into a fresh conversation on its own. Save the answer, the date, the app, and the links it showed you.
Add three rows to prompt-tracker.csv: mentioned or not, where you placed, who was named instead, what was cited. Open one cited page that names a competitor. That page is the gap.
That is one Monitor run and the beginning of a Strategist read. FIRST-RUN.md covers the same exercise more carefully, including how to keep test conversations clean.
Open the Markdown in any text editor and the CSVs in a spreadsheet. Every agent brief is meant to be edited to fit your business before you run it.
These files do not set up bots, connect accounts, or start schedules. Your AI app does that, with your accounts and your approval. Subscriptions and usage costs are yours.
This pack helps you collect evidence and decide what to do next. It cannot guarantee that an AI assistant will cite you, recommend you, or send you traffic. Anyone who promises that is guessing.
AI Search Visibility Agent Team · Version 2.0 · Updated September 9, 2026
Back to the assets