LinkedIn maxxing with Grok Bot: 12 prompts for a team of AI teammates

LinkedIn maxxing with Grok Bot means running your LinkedIn content system with a small team of AI teammates in xAI's Grok Bot app: an editor, a researcher and a writer, each with one job, all working on the same cloud computer. Twelve prompts cover the whole cycle, from a voice guide and research on X to drafts, reviews and monthly reports, and the teammates keep working on schedules after you close the laptop.

A model asked to "write a LinkedIn post about X" leaves out the part that makes a post worth reading: understanding your clients, choosing an angle, collecting proof and knowing which claims you can stand behind. The prompts below turn that part into a repeatable process, from studying your market to planning a calendar to tracing which posts led to real business conversations.

What you end up with:

• A voice guide built from the way you actually write and talk.
• Research that finds ideas before the calendar asks for them.
• Drafting and review with written quality standards.
• Monthly reporting that tells you what to write next.

Bookmark it for your next planning session.

What context does Grok Bot need before the first prompt?

More than a job title and an instruction to "sound human". Your teammates need to know who your clients are, what you think about your industry and which claims you can prove.

Start with this prompt:

This is the background for all my LinkedIn content. Save it as a reusable reference file and tell me where you put it.

• The business: [company name, website, what we sell, how long a typical sale takes].
• Who we serve: [client roles, company sizes, industries, locations]. Their problems: [problems]. What they object to: [objections]. What makes them start looking for help: [triggers].
• How we position ourselves: we help [audience] reach [outcome] using [approach]. Our clearest differences: [differences]. Who we are not right for: [disqualifiers].
• My opinions: three things I believe about my industry: [opinions]. Advice I disagree with and why: [common advice], [reason].
• Evidence you may use: [approved case studies, results, examples, personal experience]. Never make up customers, revenue figures, stories or quotes.
• How I sound: [10-20 posts] and [transcripts or other writing samples]. Things to avoid: [phrases, tones, formatting habits].
• What I want from LinkedIn: my main goal is [qualified conversations / support for open deals / recruiting / awareness]. My offer is [offer], and the next step I want from interested readers is [action].

Rules: work only with connected sources you are allowed to access. When a source won't open, ask me for an export or pasted text. Publishing, comments, connection requests and messages on LinkedIn are mine to do. Save drafts to [review destination], label what is evidence and what is inference, and flag anything that needs my input.

Update the file whenever your offer, audience or opinions change.

Where does Grok Bot keep this context?

Grok Bot is a separate xAI app from the Grok chat app. In it you create named AI teammates, each with one job, and every teammate on your account works on the same cloud computer, which has a browser, a filesystem and a terminal. So the context lives in two layers:

• A file that every teammate can read. Paste the prompt into the chat of a new teammate, say one called "LinkedIn Editor", and ask it to save the result as linkedin-context.md on its computer. Every teammate on the account shares the files on that computer, so the researcher and reviewer you add later read the same file and not a copy of it.
• Memory for each teammate. Each teammate also keeps a memory of its own: lasting preferences, its role and summaries of past work. Conversations don't cross between teammates, and only a few facts about you, such as your name and time zone, are shared. After the first session, ask the teammate what it remembers about you to check what it kept.

The first teammate's description can be this short:

Name: LinkedIn Editor
Job: Keep my LinkedIn content system running.
Description: Read linkedin-context.md before every job.
Draft and review posts in my voice. Keep a source for every
claim and mark what is inference. Never publish, comment,
connect or send messages. Leave drafts for me.

If you already built this context in another AI assistant, bring it over with Import Memory in Settings and import it into your main teammate instead of typing it again. You can later hand a colleague the same setup: open the teammate's details, choose Share, then Create a copy. The copy includes the teammate's configuration but not your computer, your logins or your conversations, so remove anything confidential from the description before sharing. For the "review destination" in the prompt, name a Page. Grok Bot can build Pages such as reports and dashboards, which you edit together and which the teammate keeps current.

Part 1: research and content production

Prompts 1-8.

1. How do you build a voice guide in Grok Bot?

A voice guide is useful when it explains why your writing reads the way it does: how you set up an argument, where examples go and how you get from one idea to the next. "Conversational and authoritative" gives a teammate nothing to work with.

Study the posts and transcripts I've shared and turn them into a working voice guide for LinkedIn. Describe:

• how my posts open, how I explain an idea and how I wrap up an argument;
• my typical vocabulary, sentence length, paragraph length and use of lists;
• how I sound when I disagree, when I'm unsure, when I joke and when I'm confident;
• phrases I reach for naturally, quoted exactly with where each came from;
• habits from my weaker pieces that should not be turned into rules.

Then list what to avoid, including repetition that adds nothing, drama for its own sake, confidence the evidence doesn't support and stacks of one-line paragraphs. Close with three before-and-after rewrites that move bland copy into my voice with the meaning intact. The guide should be precise enough that a different writer could use it.

In Grok Bot: drop your posts and transcripts into the LinkedIn Editor's chat. In a chat only you can see, the teammate reads your own private files without asking first. Once the guide sounds right, ask the teammate to save it as a skill called "my-voice" and to show you the skill before saving. A skill is a reusable set of instructions available to all teammates, so every writer you add later follows the same voice rules. Add the openers you never use, such as "I'm thrilled to announce" or "Excited to share", to the avoid list by name. Keep the editor narrow: a short name, one main job and a clear description help a teammate gather more useful context over time than a general "content helper" would.

Why it matters: the guide makes your feedback reusable. You fix the instructions behind every draft once, instead of fixing the same phrasing every week.

2. How do you find the client questions worth answering?

Sales calls and customer conversations contain questions that never appear in a keyword tool. The ones worth posting about usually circle a decision someone is finding hard.

Read the sales calls, support conversations, customer interviews and lost-deal notes from [period]. Put the questions that come up again and again into these groups:

• problems clients are trying to understand;
• doubts that hold a decision up;
• options they are comparing us with;
• expectations that later turn into disappointment;
• questions they ask right before they sign.

For each theme, add an anonymized quote, where it came from, the number of separate conversations it appears in, the client's role and their stage in the decision. Offer two LinkedIn angles per theme: one helping the reader work out what the problem is, one helping them assess a solution. Rank the ten strongest by how often they recur, how much they matter to revenue and how solid our evidence is. Keep clients' own words separate from your interpretation.

In Grok Bot: open Connect apps and add the places where your conversations live, such as Gmail, Google Drive, Notion, Slack or your support tool's plugin from the marketplace. Grok Bot uses a connector when one exists and falls back to its browser for everything else; when a site asks for a login or a human check, the teammate passes that step to you. Before the first run, ask the teammate which apps it can reach right now, then tell it: "Add these to shared memory, so every future teammate knows where to find this." Keep the first run read-only ("read, do not reply or change anything") and ask for the ranked themes on a Page, so prompts 5 and 9 can use them.

Why it matters: you get a list of post ideas tied to conversations your company is already having, in words your clients use and with evidence you can check.

3. How do you find topics moving from X into your market?

A thread taking off on X can become a good LinkedIn post if you explain why it matters to your clients and bring something more than a summary.

With the X search tools you have and public sources you can open, review the last [seven days] of conversation about [topics] and from these accounts: [accounts]. Choose up to 10 developments or arguments that matter to [client role]. For each, give me:

• the original source, its author, the date and the link;
• a neutral account of what happened or what is being argued;
• the strongest competing interpretations;
• what it means for our audience in their daily work;
• an angle our own experience or evidence can support.

Present like and repost numbers as a snapshot, don't treat popularity as evidence and verify factual claims against the original source wherever possible. Recommend three ideas to develop and tell me what I'd need to add before each is a credible LinkedIn post.

In Grok Bot: this is the step where X is built in more deeply than anywhere else in the system. When you ask a teammate about your X account, it connects through your linked X account, and the X connector adds search on top. Two habits improve the prompt:

• Collect all week. When you see a thread worth answering on LinkedIn, tag Grok Bot under the post on X. That passes the post to your teammate, so by Friday the research starts from a shortlist you picked yourself.
• Give research its own teammate. Create a "Researcher" whose only job is this prompt, and add it to a group chat with the LinkedIn Editor. The researcher posts the ten developments, the editor checks them against your evidence, and you step in only to pick the three ideas.

Skip requests for follower breakdowns: X doesn't provide that data, and Grok Bot will tell you so. Any like or repost counts it returns are a snapshot from the moment it checked. If you already filter X with Advanced Search operators such as min_faves:500, include them in the brief so the researcher starts from discussions that have already gathered a crowd, then ask it for the angle your clients haven't seen yet.

Why it matters: research hands you a starting idea; your judgment is what gives readers a reason to follow you. The LinkedIn version should tie the discussion to a decision your readers face.

4. How do you analyze posts that beat their authors' usual results?

Holding every post up against a celebrity account's best hit distorts what you learn. First compare a post with its own author's normal results, then look for what explains the difference.

Work through this LinkedIn swipe file and any performance data that comes with it: [file]. Where the data allows a fair comparison, measure each post's reactions against the median for the same author, format and period. Show the arithmetic and flag small samples, missing figures and posts that are too recent to judge. For the biggest outliers, set out:

• the topic and the reader it targets;
• the opening, the argument, the evidence and the ending;
• the format and approximate length;
• whether the comments suggest genuine interest from clients;
• other things that could explain it, such as a launch, a news story or unusually broad distribution.

Distill five structural patterns I can reuse and suggest an original idea for each from my own experience. Don't borrow anyone's distinctive wording, and don't describe an observed pattern as an algorithm rule.

In Grok Bot: the teammate has a terminal on its cloud computer, so ask it to calculate the medians with a small script and keep that script next to the swipe file. Next month you run the same calculation on fresh data instead of relying on a new estimate. Tables in the chat have a copy button for Slack or Markdown, which makes the outlier list easy to move into your notes. To check a single post by hand, use the LinkedIn engagement rate calculator. Ask the teammate to mark on the Page which outliers came from accounts far bigger than yours, since their structure transfers better than their numbers.

Why it matters: you get hypotheses worth testing and avoid assuming that a popular post's layout will bring you its results.

5. How do you turn recorded conversations into original posts?

A transcript often hides a strong post inside a story, a throwaway explanation or a disagreement that got less attention than it deserved at the time.

Read this transcript of an interview, podcast, webinar or meeting: [source]. Pick eight moments that could each carry an original LinkedIn post, giving priority to decisions, mistakes, surprising observations, practical explanations and opinions grounded in experience. For each moment, include:

• the exact quote with its timestamp or position in the source;
• the main idea and who would find it useful;
• context needed to represent the speaker fairly;
• facts we already have that support it;
• one follow-up question that would sharpen the idea.

Choose the three strongest and write a draft from each using my voice guide. Match the speaker's level of certainty, cut confidential details and leave placeholders where a result or example still has to be confirmed.

In Grok Bot: send the teammate to the recording's folder in Google Drive, or drag the transcript file into the chat. With no transcript yet, start a voice chat with the editor and talk the story through for ten minutes; it is often the quickest way to get raw material in your own words. Ask it to file the eight moments under the transcript's name on the shared computer, so the calendar in prompt 6 can refer to them.

Why it matters: you publish from knowledge that already exists, without asking anyone to invent a new opinion whenever the calendar has a gap.

6. How do you build a four-week content calendar?

The calendar has to weigh what clients need to learn against what you have the evidence to say. Filling every slot with a new version of your offer makes your feed repetitive.

Plan four weeks of LinkedIn posts at [three a week] from the research ideas I've approved. Combine:

• practical explanations that help clients solve a problem;
• opinions backed by experience or evidence;
• customer examples we have permission to use;
• answers to objections and decision questions;
• an occasional offer post with a relevant next step.

For each post, state the reader, the one objective, the central argument, the sources, a suggested format and the action you want from the reader. Check the plan for repeated topics, arguments that overlap, unsupported claims and too many sales posts. Give week one full briefs and weeks two to four shorter outlines. Hold one slot each week for a timely replacement so the plan can change.

In Grok Bot: ask for the calendar as a Page the editor keeps current, so a swap in week three updates the Page instead of creating another file. If you also post on X, tell it that a LinkedIn slot and an X slot must not use the same line on the same day. With Google Calendar or Outlook Calendar connected, it can block each publishing slot in your calendar, where you will actually notice it. Grok Bot can also generate images and short videos from a description, so a brief marked "image" or "video" can arrive with a first visual for you to judge. Choose fixed days and hours with the best time to post on LinkedIn tool and keep them steady, and turn "carousel" briefs into slides with the LinkedIn carousel maker.

Why it matters: a working calendar cuts weekly planning time and still leaves room for better ideas, new evidence and changes in your market.

7. How do you draft posts that have something to say?

Settle the argument through research before Grok Bot touches the sentences. Reverse the order and you get smooth prose that needs several paragraphs to reach an obvious idea.

Write three LinkedIn drafts from these approved briefs: [briefs], using my business context and voice guide. Each draft should:

• argue one thing for one reader;
• start with a concrete observation, tension, result or question that the brief backs up;
• show the reasoning through a real example;
• use bullets only where parallel ideas scan better as a list;
• mix sentence and paragraph lengths, with the odd one-line paragraph for weight;
• finish on a useful implication or a natural next step.

Target [180-300 words] unless the idea calls for another length. Leave out repeated endings, invented personal stories, overblown promises and stock contrasts. Offer two other openings and say briefly how each changes the emphasis. Put missing evidence in a separate list, apart from the copy that can be published.

In Grok Bot: hand drafting to a third teammate, a "LinkedIn Writer", whose description holds one trust rule: never invent a number, quote, source or result, and keep any uncertainty the brief contains. Type / in the chat to call the "my-voice" skill. When you copy a finished draft out of the chat, headings, lists and links survive the paste into Google Docs or Notion, so what you review is what you publish. Get more opening options from the LinkedIn hook generator and paste the best two into the brief, so the Writer drafts around openings you already like.

Why it matters: tight constraints produce drafts that are easier to judge, so your review goes into the argument and not into cutting filler.

8. How do you set up a separate editorial review?

The reviewer needs one defined job: find what is weak in the draft and explain the fix. A comforting "This is strong" isn't a review.

Review these LinkedIn drafts against the brief, the sources, the business context and the voice guide. Give each a score out of 100, weighted as follows:

• relevance to clients: 25;
• concrete detail and supporting evidence: 25;
• strength and clarity of the argument: 20;
• match with my voice: 20;
• readability and formatting: 10.

Support each score with evidence from the draft. Verify every factual claim, mark generic passages and flag repetition, abrupt one-line fragments and paragraphs that should be lists. Return a revision ready to publish, a short explanation of the changes that matter and any open questions. Any invented evidence, exposed confidential information or misrepresented source blocks the draft, whatever its score. Keep all of it in the review document for a person to approve.

In Grok Bot: there is no model picker, so the separation comes from roles. The LinkedIn Editor reviews what the Writer produced; it has its own memory and didn't write the draft, so it has no reason to defend the sentences. Put the Writer and the Editor in one group chat, let them pass the draft back and forth, and ask for the final scores on the review Page. Before approving, check how the revision will look in the feed, including where "see more" cuts it, with the LinkedIn post preview.

Why it matters: the rubric puts your standards in writing. The score keeps feedback organized; the source checks and the editor's judgment decide whether the post is ready.

What should you do in the first hour after publishing?

Prompt 8 ends with a revision you have approved, and you publish it yourself. From there the method is the same for everyone: the first hour is when you want as many reactions on the post as you can get.

1. Ask your own people first. Colleagues, friends, family and clients, anyone you can actually reach, can react as soon as the post is up.
2. Use a community when your circle is small. Pick one where people act by hand from their own accounts and where an action that never happened can be reported.
3. Join checked groups. The groups in the table below were checked on Oct 21, 2026.
4. Try LinkedInLikes. It is a free community where LinkedIn members like, comment on, repost and save each other's posts by hand from their own accounts. You share a public link, never your password or cookie; comments are written in each member's own words; and an owner can report an action that wasn't done within 72 hours. See how comments work on LinkedInLikes.

Which LinkedIn groups can you join for the first hour?

These free communities were checked on Oct 21, 2026; the member counts are what each platform showed that day. Each group has its own admins and rules, so read the pinned message before sharing anything. If you would rather skip group chats, LinkedInLikes for likes works from one public link to your post.

Part 2: pipeline and performance

Prompts 9-12.

9. How do you write content around deals that went quiet?

Some of your best posts answer a concern that several prospects raised privately. The goal is to explain it well enough that future clients come to you already informed.

Go through the CRM records and call notes for qualified deals that went quiet during [period]. List the objections that keep returning and sort them into:

• objections people stated outright in conversations;
• explanations that are possible but unconfirmed;
• delays on the business side that content is unlikely to fix.

For each objection that repeats, note the client's role, the deal stage, the evidence we hold and the question a useful LinkedIn post would answer. Recommend five content briefs. Each should treat the concern fairly, show where our approach fits, be honest about where it doesn't and name the proof required. Write drafts of the two strongest without naming prospects or revealing anything about the deals. Separately, list approved resources a salesperson could share in a matching conversation.

In Grok Bot: connect your CRM through the plugin marketplace; Salesforce, as one case, lists each of its servers as its own row so you can sign in to each one. Before the first run, give the teammate a plain standing rule, "never edit CRM records", and ask it to remember that. For anything that writes, add an Always Ask rule in Settings -> Permissions. Keep the anonymized briefs on a Page your sales team can open. Every teammate shares one cloud computer, so a CRM export saved there is readable by all of them, and xAI's own advice is not to treat separate teammates as a security boundary. Delete raw exports once the briefs are finished.

Why it matters: it ties marketing to the questions salespeople answer over and over, and keeps you from writing persuasive posts about problems your clients don't consider important.

10. How do you prepare briefs for LinkedIn conversations?

When someone answers a post, the right next move depends on what they said and on whether a business conversation would be welcome. A reaction by itself doesn't show commercial intent.

Here are LinkedIn conversations I've gathered: [comments, inbound messages or interactions I noted by hand]. Split them into:

• questions worth a useful public answer;
• conversations where a resource would help;
• clear interest in our offer;
• general discussion with no commercial next step.

For every relevant conversation, summarize the need the person stated, the business context we know and what we still don't. Draft a reply that answers their actual message before raising anything else. Where it fits, add one follow-up question or a resource we already have. Don't assume commercial intent from a like, don't pretend to know someone you don't, and don't push a meeting request into a conversation that won't hold one. Keep every reply as a draft for me to review and send.

In Grok Bot: there is no official LinkedIn connector, so the pattern is simple: the teammate prepares and you post. Paste the comments and messages in, or forward them from email, and ask for each reply as a draft card. When an inquiry comes in by email, the teammate can prepare a Gmail reply as a draft; the draft card has a Discard button, and nothing is sent until you send it. Answer every useful comment on your post while the thread is still fresh. If salespeople should use these briefs as well, publish the Editor as a shared teammate in Slack so they can request a brief on a commenter without opening the app. Between your own posts, leave considered comments under other people's posts in your niche; the LinkedIn comment generator gives you a first version to rewrite in your own words.

Why it matters: preparation makes replies more thoughtful without turning each exchange into a sales sequence. Sometimes a good answer is the whole outcome.

11. How do you report on business results with clear attribution?

Impressions describe distribution. They can't say whether the right people became interested or whether a post helped a deal along.

Bring together my LinkedIn analytics, tracked website visits, inbound inquiry records and the relevant CRM data for [period]. Report on:

• reach and interactions by post, topic and format;
• qualified inbound conversations;
• resource sign-ups and demo requests where tracking exists;
• opportunities with a documented contact through LinkedIn;
• time spent producing and reviewing content.

Separate conversions tracked directly, discovery people reported themselves and possible influence. Never attribute a whole deal to LinkedIn because one contact reacted to a post. Call out posts that attracted relevant prospects, posts that reached many people of little relevance, and posts with too little data to judge. Finish with three findings, three limitations and one recommended experiment. Where you can, compare posts at the same age so recent posts aren't measured against older ones.

In Grok Bot: make the report a dashboard Page that the teammate refreshes every month, so this month's figures sit beside last month's instead of in a new file. Drop the LinkedIn analytics export into the shared folder; the CRM side arrives through the plugin from prompt 9. Shared Pages also show up in the Library, so your team reads the same numbers you do. To see whether a post's reach was normal for an account of its size, compare it with the LinkedIn impressions calculator.

Why it matters: better reporting gives you a reason to keep an approach or change it, and it keeps uncertainty visible instead of turning incomplete tracking into a confident revenue claim.

12. How do you turn your edits into a better system?

Each edited draft records something about your standards. Write the repeat lessons down so you spend less time fixing familiar problems.

Compare the original drafts with my edits, the published versions and whatever performance data exists for [period]. Find:

• voice corrections I make again and again;
• claims that needed more evidence;
• ideas that became clearer after I reframed them;
• formatting changes that made posts easier to read;
• topics that started useful conversations with clients.

Propose changes to the voice guide, the research briefs and the editing checklist. For each new rule, include an example, its purpose and the cases where an exception makes sense. Keep matters of style apart from guesses about performance, and never build a permanent rule on a single post that did well or badly. Give me a short change log to approve and one controlled experiment for next month.

In Grok Bot: approved changes go into the "my-voice" skill and linkedin-context.md, which every teammate reads, and not into one teammate's memory. Ask the Editor to show you the edited skill before saving and to list which teammates use it, so you know who the change affects. Keep the change log on the shared computer with a date in the file name. Memory is for preferences; for facts that matter, ask the teammate to check the current source instead of recalling it.

Why it matters: this is how the system gets easier to run. Your corrections become written guidance, and performance ideas stay open to testing instead of hardening into superstition.

How should you roll out all 12 prompts?

Add the work in stages so you can check each output before taking on another recurring job.

• Week 1, the foundation. Load your business context, run prompt 1 and check the voice guide against writing you would actually publish.
• Week 2, the research backlog. Run prompts 2-4 on a manageable set of customer conversations, market sources and LinkedIn examples.
• Week 3, the first batch. Use prompts 5-7 to turn approved material into a calendar and a few drafts.
• Week 4, the review. Run prompt 8, publish the posts you approve, work the first hour on each one and note the changes you make.
• Weeks 5-6, conversations. Add prompts 9 and 10 once you have enough sales context and inbound activity for them to be useful.
• Every month after that. Use prompts 11 and 12 to review results, improve the instructions and pick the next experiment.

For every recurring job, decide its inputs, where the output goes, how often it runs and who owns it. Test one run before trusting the schedule, and have Grok Bot report missing access or unfinished work.

Skills and routines

Grok Bot splits recurring work in two. A skill describes how to do something; a routine gives that skill to one teammate and says when to run it, on a schedule or, where supported, after an event. Run a skill once by hand before turning it into a routine. Routines run on the cloud computer in your time zone, even when the app is closed:

Any message a routine wants to send waits for you as a draft, and approvals it asks for expire after about 30 minutes, so schedule routines for times when you are around. The iOS and Android companion apps show the same teammates and conversations, so you can approve a routine's output from your phone. Publishing on LinkedIn stays with you.

You don't need 12 teammates on day one. Start with the job that takes the most time and produces something you can reliably judge.

What can't Grok Bot do for you?

Grok Bot can organize research, find useful evidence, prepare drafts and learn from your edits. The quality of the posts still comes from the experience and judgment you bring.

That means holding a real opinion, being willing to explain it, and noticing when a polished draft says something you would never defend in conversation.

Every post this system produces still needs a first hour with some activity. Share it with people you know, and if your circle is small, create a free LinkedInLikes account: confirm your email to get 100 starter points for your own posts. Otherwise, load your context and run prompt 1, because better inputs make every later step more useful.

More articles

• How to build an AI content system for LinkedIn from one empty folder (2026 setup)
• The 2027 LinkedIn marketing playbook
• LinkedIn maxxing with Grok
• LinkedIn maxxing with Claude Code