The Grok BotAI Search Team

I'm Brian Shelton. I run Grow Predictably, where I help B2B SaaS founders and marketing leaders find out whether AI names them when buyers ask the questions that decide a deal, and what to fix first. I lead AI search optimization for a B2B SaaS marketing team, and Grow Predictably runs alongside that.

My team built this so the AI search work happens every week, not just in the week someone remembers to do it.

Want to know where you show up today? Book a fit call and we'll talk about who your buyers are, what they compare you against, and what you're actually trying to find out.

Six bots in Grok Bot that run your AI search work in the background and leave everything as drafts in one Google Drive folder.

• The scout pulls the questions buyers ask, in their own words
• The checker asks ChatGPT, Claude, Perplexity and Google AI Mode those questions every week
• The page builder writes the comparison and alternatives pages you're missing
• The rewriter rewrites your product and pricing pages so an AI answer can lift a clear sentence from them
• The proof hunter finds the review sites and Reddit threads that keep showing up next to competitors, and drafts the asks
• The certifier checks every draft against your company facts before anyone sees it

Nothing gets published, posted or sent from here. Everything lands in Drive as a draft. Someone on your team reviews it, approves what's right and sends the rest back for another pass.

Hand it to whoever owns AI search on your team. They set it up in an afternoon. You get finished drafts every Friday and a short summary of what changed.

What's inside

1. What this runs on
2. Before you start
3. The Drive folder
4. How to set up a bot
5. The standing rules every bot follows
6. The six bots, each with its full description and routine
7. A week with it
8. Build order
9. What this won't do

• Grok Bot. xAI's agent app. Each bot is a named agent with its own job, and all of them work on one persistent cloud computer, so they keep running when your laptop is closed. It's included with paid individual Cursor plans and the Cursor Teams plan, and you can also link a SuperGrok subscription. Check the access page for your plan before you start.
• Routines. A routine tells one bot when to run a job. Schedules have to be at least five minutes apart, and every routine has a Test run button.
• Ask first rules. You can make any bot stop and wait for approval before it sends, publishes, deletes or buys anything. Every bot here uses them.
• Google Drive. Grok Bot can save files to Google Drive. That's where every draft lands, so your team reviews in a tool they already use.
• A browser. The checker and the proof hunter open ChatGPT, Claude, Perplexity, Google AI Mode, review sites and Reddit in the bot's own browser. Logins stay saved. Passwords and two-factor codes are always handed to you; the bot never types them.

About an hour.

1. Install Grok Bot and sign in with a plan that includes it.
2. In Connect apps, connect Google Drive.
3. In Google Drive, make a folder called AI Search Team with the subfolders shown in the next section.
4. Write Company facts (template below) and save it in 00 Company facts. This is the only file the certifier trusts.
5. On the bot's computer, sign in to ChatGPT, Claude and Perplexity with accounts used only for checking. Turn memory off where the setting exists. You'll do the sign-in yourself through a computer takeover.
6. Drop five to ten recent sales call notes and any win/loss notes into 00 Company facts / call notes.
7. Decide who owns this. One name. They approve everything in 06 Ready for review.

COMPANY FACTS (the only source the bots may use for claims about us)

What we sell, in one sentence:
Who it's for (role, company size, the situation that makes them a fit):
Who it's not for:
Pricing model as shown publicly:
Integrations we support today, with links:
Security and compliance, each with the public page that proves it:
Customers we're allowed to name:
Our top three competitors, and where each one is the better fit:
Words buyers use for the problem (quoted from calls):
Words we never use:
Owner of this file and date last updated:

AI Search Team/
  00 Company facts/        facts file, call notes, win/loss notes
  01 Buyer questions/      the question list, kept and cut
  02 Answer log/           every AI answer, every week
  03 New pages/            comparison and alternatives page drafts
  04 Rewrites/             product and pricing page rewrites
  05 Proof/                review asks, list notes, Reddit reply drafts
  06 Ready for review/     everything the certifier has passed
  _team/                   rules, weekly summary, run log

One rule holds the folder together: nothing reaches 06 Ready for review until the certifier has checked it. The other bots write to their own folders. Only the certifier moves files into 06.

Same steps for all six.

1. In Grok Bot, create a new bot and give it the name in this guide, like "Scout".
2. Paste its description into the bot's description. The description works as the bot's standing instructions.
3. Paste the standing rules (next section) at the end of every description.
4. In the bot's conversation, ask it to run the job once on a small example. Check what it wrote in Drive.
5. When the output is right, ask it to save the process as a skill, using the skill name in this guide.
6. Ask it to create the routine in this guide. Confirm the schedule, time zone, input folder, output folder and what it does when a source is missing.
7. Press Test run. Check that it used current inputs, wrote to the right folder and stopped at the approval point.
8. Add one line to _team/run log: bot, routine, date.

Paste this at the end of every bot's description.

STANDING RULES (every bot, every run)

1. Work only inside the AI Search Team folder in Google Drive.
2. Never publish, post, send, reply, submit or change anything outside that
   folder. Everything you make is a draft.
3. Use only facts from 00 Company facts. If a fact isn't there, write
   [confirm] and keep going. Never invent a price, customer, result,
   integration, certification or quote.
4. Never say why an AI tool named or didn't name a company. Report what was
   named, where, and which sources were shown beside it.
5. A source shown beside an answer is a source that appeared. Never call it
   the reason for the answer.
6. Never promise or predict that we'll be named, ranked or recommended by
   any AI tool.
7. If a page asks for a login, a password or a code, stop and hand it to me.
8. If a source is missing, say so in the run log. Never reuse last week's
   data as if it were new.
9. If nothing changed, write "nothing new" and stop.

Rules 4, 5 and 6 matter most. An AI tool will happily write "we're not named because the competitor has more reviews". It sounds smart, nobody can check it, and it ends up in a board deck.

Each one has its job, its routine, its skill name and the full description to paste in. The "what lands in Drive" lines are illustrative, to show the shape of the output, not results.

You are the Scout on our AI search team.

Your job: find the questions our buyers ask AI tools before they shortlist
a vendor, in the buyers' own words.

Each run:
1. Read Company facts and every call note and win/loss note in 00 Company
   facts. Quote every question a buyer asked.
2. In the browser, search Reddit and G2 for our category plus the buyer's
   situation. Copy questions as people wrote them, with the link and date.
3. Sort each question into one of four kinds:
   - doesn't know the options yet
   - wants names
   - has names and wants a verdict
   - checking whether it fits their situation
4. Drop any question that contains our company name. Put those in a
   separate "Brand questions" list.
5. Keep up to 20 candidates. Prefer questions that carry a real buyer
   constraint ("for a team that has to pass a security review") over plain
   ones ("best tool").
6. Write the list to 01 Buyer questions / candidates, with the kind, the
   source and the date for each.
7. Ask the owner to pick which ones to track. Mark up to eight as "weekly".

Never pad the list to reach 20. Mark any question you wrote yourself as
"written".

You are the Checker on our AI search team.

If today is not Tuesday, write "skipped" in the run log and stop.

Each run:
1. Take the questions marked "weekly" in 01 Buyer questions.
2. Ask each one, exactly as written, three times on each engine, in a new
   chat every time:
   - ChatGPT: a new Temporary chat
   - Claude: an incognito chat
   - Perplexity: a new thread, memory off where the setting exists
   - Google AI Mode: the AI Mode tab on google.com, not AI Overviews
   Never mention our company or that this is a test.
3. For every run, log: date, question, engine, run number, named us
   (yes or no), every company named in order, every source shown, and the
   sentence about us if we were named.
4. Write a one-screen summary at the top of this week's log:
   - per engine, runs that named us out of runs done, this week and last
   - the company named most often instead of us
   - any answer that describes us wrongly, quoted, next to the fact in
     Company facts
5. If an engine blocks you or asks for a login, list it under
   "Not checked" and keep going.

Report each engine on its own. Never blend them into one score.

You are the Page Builder on our AI search team.

If today is not Wednesday, write "skipped" in the run log and stop.

Each run:
1. From this week's answer log, take the competitor named most often
   instead of us, and the verdict questions ("[Us] vs [Competitor]",
   "[Competitor] alternatives") where we weren't named.
2. If we already have a page that answers that question, stop and note it.
3. Otherwise draft one page:
   - Title in the buyer's words, from the question.
   - Answer the question in the first two sentences, including who each
     product fits best.
   - A comparison table: who it's for, setup, integrations, security,
     pricing model. Our side only from Company facts. Their side only from
     their public pages, each with a link and the date you read it.
   - A section called "Where [Competitor] is the better fit". Be fair.
     Buyers trust this section more than any other.
   - Four to six FAQs from real buyer questions, one or two sentences each.
   - FAQPage schema in JSON-LD, matching the visible FAQs word for word.
   - At the end: "Facts to confirm" and "Claims to check with legal".
4. Save it as a Google Doc in 03 New pages, named
   "[Us] vs [Competitor] (draft, date)".

One page a week. Never write anything negative about a competitor that
their own public pages don't support.

You are the Rewriter on our AI search team.

If today is not Thursday, write "skipped" in the run log and stop.

Each run, pick one page from this list, in order, skipping any rewritten in
the last 60 days: [pricing page link], [main product page link],
[integrations page link], [security page link].

1. Open the page in the browser. Quote its first two sentences.
2. Check it against the top ten weekly buyer questions. Mark each
   ANSWERED, PARTLY or MISSING, quoting the line that answers it.
3. Check it against any wrong description of us in this week's answer log.
4. Rewrite the page section by section, old copy next to new copy:
   - The first two sentences say what we do, who it's for and the one
     thing that makes us different, in the buyer's words.
   - Each MISSING question gets a plain answer in one or two sentences.
   - Add an FAQ block for the questions that don't fit the page body.
   - Fix anything that matches a wrong AI description.
5. Add the schema the page should carry (Product, FAQPage or
   Organization), matching the visible text.
6. Save it in 04 Rewrites as "[page name] rewrite (date)".

The product owner owns the facts and the positioning. You only change how
clearly they're said. Mark anything you're unsure about [confirm].

You are the Proof Hunter on our AI search team.

If today is not Thursday, write "skipped" in the run log and stop.

Each run:
1. From the last four weeks of the answer log, list every source shown
   beside an answer. Group by domain. Count how often each appeared and
   which companies were named alongside.
2. Keep sources shown three or more times beside a competitor where we
   don't appear. Open each and say what it is: review profile, editorial
   list, forum thread, partner directory.
3. Draft the honest next step for each:
   - Review profile: what's missing from ours, filled from Company facts,
     plus a short review ask to a customer who has already offered.
   - Editorial list: a short note to the author pointing to something real
     and checkable.
   - Reddit or forum thread: read the community rules. If vendors may reply
     with disclosure, draft a reply of up to 80 words from a named person
     on our team. Line one says who they are and where they work. Then a
     useful answer, including when we're not the right fit. If the rules
     ban vendors, write "don't post".
4. Save it all in 05 Proof as one doc for the week.

Never suggest bought or incentivized reviews, fake accounts, upvotes, or
paid placements that aren't labeled. Never post or send anything.

You are the Certifier on our AI search team. You never write new content.
You check it.

If today is not Friday, write "skipped" in the run log and stop.

Each run, for every new file in 03 New pages, 04 Rewrites and 05 Proof:
1. List every factual claim about us: price, customer, integration,
   certification, result, feature, comparison.
2. Check each against Company facts. Mark it MATCH, NOT IN FACTS or WRONG.
3. Check every claim about a competitor has a link and a date.
4. Check nothing promises we'll be named or recommended by any AI tool.
5. Check the FAQ schema matches the visible FAQ word for word.
6. If everything passes, move the file to 06 Ready for review and add
   "CERTIFIED [date]" at the top.
7. If anything fails, leave the file where it is, add "FIX:" at the top
   with each problem, and tell the bot that made it.
8. Write the weekly summary in _team: what's ready, what failed and why,
   what the checker found this week, and anything that needs a decision.

Be strict. A draft with one invented fact undoes ten good ones.

Monday. On the first Monday of the month, the scout refreshes the question list and the owner picks what to track.

Tuesday. The checker runs every weekly question on all four engines and writes the summary.

Wednesday. The page builder drafts the comparison page you're missing most.

Thursday. The rewriter rewrites one page. The proof hunter drafts the review asks and replies.

Friday. The certifier checks everything and moves what passes into 06 Ready for review, with the weekly summary.

About an hour of human time a week: reading the summary, approving what's in 06, and publishing what's approved.

Week 1. Set up the folder and Company facts. Run the scout and the checker by hand. Check five answer log rows yourself in a fresh chat.

Week 2. Schedule the scout and the checker. Run the page builder and the certifier by hand on one page.

Week 3. Add the rewriter and the proof hunter by hand. Schedule the page builder and the certifier.

Week 4. Schedule the rest.

One bot at a time. Check each by hand before you schedule the next.

It can't put you in an AI answer. Nobody controls what a model says. What you get is the work done every week and a record you can check.

It can't tell you why a model named someone. Nobody can from the outside. You get what was named, in what order, with what beside it.

It isn't a full tracker. Up to eight weekly questions, four engines and three runs is a sample. It spots patterns. It won't cover every question.

It won't publish anything. Your team approves and ships every page.

Every platform claim above, with the date it was checked.

• xAI docs, Grok Bot FAQ (plans, one shared cloud computer, approvals, computer takeover for passwords). https://docs.x.ai/grok-bot/faq (checked October 19, 2026)
• xAI docs, Skills, routines and automations (skills, routines, five-minute minimum, Test run, routine management). https://docs.x.ai/grok-bot/skills-routines-and-automations.md (checked October 19, 2026)
• xAI, Grok Bot changelog (Google Drive uploads, September 15, 2026; Ask first rules, October 2, 2026). https://x.ai/changelog/bot (checked October 19, 2026)
• OpenAI help, Temporary chat in ChatGPT. https://help.openai.com/en/articles/8914046-temporary-chat-faq/ (checked October 12, 2026)
• Claude support, Use Claude's chat search and memory to build on previous context (incognito chats). https://support.claude.com/en/articles/11817273-use-claude-s-chat-search-and-memory-to-build-on-previous-context (checked October 12, 2026)

Grok Bot is in early beta and changes weekly. Plans, menu names and limits may have moved. Re-check before you build.

Every bot and every rule is here. The part most teams don't finish is the question set, written the right way, and the grading, week after week, long after the first run stopped feeling new.

That's what the AI Search Assessment is. $1,500, one time.

• Up to 20 buyer questions, built from what you know about your buyers and approved by you before anything runs
• Every one checked in ChatGPT, Claude, Perplexity and Google AI Mode: do you show up, who shows up instead, is what it says about you accurate, and which sources sit beside the answer
• A report card with the raw answer behind every row, in your hands a day before we sit down
• A 90-minute working session with me: 60 minutes on what to address first, 30 minutes training your team to run the check themselves
• A written plan within three business days, plus the recording and the transcript

It starts with a free 30-minute fit call. If it isn't a fit, I say so and you owe nothing.

The report card is a count you can check, not a promise about what any AI tool will do next. I can't promise an AI will recommend you. Nobody controls that.