Build a Five-Agent Content Agency

Build a five-agent workflow for researched, approved X and LinkedIn campaigns.

Introduction

30 Second Summary

A promising market insight often gets stranded in a browser tab. Moving it into prospect research and social content by hand makes evidence easy to lose.

In this project, you will build a five-agent content agency in Grok Bot for AI consulting firms that serve B2B organizations. One brief moves through current research, prospect qualification, channel-specific writing, human approval, and social scheduling in Buffer.

What You'll Build

You will watch one brief travel through five visible Bot handoffs until it becomes a sourced prospect shortlist and distinct X and LinkedIn posts ready for your approval.

By the end of this project, you'll have:

Trace a current market signal through visible handoffs until it becomes one focused campaign for AI consulting firms.
Review three qualified prospects with public fit evidence plus an unsent personalized opener for each firm.
Approve platform-specific posts before seeing each one scheduled on the correct Buffer channel.
Secret Mission: Switch the existing team to local fitness studios through a new configuration file without carrying over stale AI consulting context.

Are there any prerequisites?

You need Grok Bot desktop on Linux with a paid Cursor account. You also need X and LinkedIn profiles because this project connects both to a new Buffer Free account.

Before We Start

This checkpoint defines the campaign's commercial intent before any hands-on work begins. Your first run markets evidence-backed social content and market intelligence to founders or marketing leads at AI consulting firms serving B2B organizations. The intended buyer action is to engage your agency about the service.

Connect the Publishing Stack

Your campaign has a clear offer. It still cannot reach anyone until its publishing destinations are connected.

Buffer will hold your X channel. It will also hold your LinkedIn channel. Grok Bot will keep every public action behind your approval.

In this step, get ready to:

Prepare Buffer Free with both selected social channels.
Configure narrow Ask first rules for public actions.
Prove the channel connection through a read-only check.

Connect your Buffer channels

Buffer needs authorization to access each social profile. The authorization should belong to the exact profiles you want the agency to use.

Keep credentials under your control

Account sign-in is the only point where private credentials enter this workflow. You handle that moment yourself through Agent Computer takeover.

Ask Grok Bot to open Buffer in its cloud browser.
Create a Buffer Free account or sign in to your existing account.
Use Agent Computer takeover when Buffer requests private sign-in information.
Complete the private sign-in yourself.
Return control to Grok Bot after Buffer finishes signing you in.
Confirm Buffer shows the Free plan for your account.

Your Buffer Free account is now active in the cloud browser. Your private sign-in details stayed under your control.

Choose + New in Buffer.
Select Connect a New Channel.
Select X as the channel type.
Authorize the exact X profile you want the agency to use.
Confirm the selected X profile appears as a connected channel avatar.

The X avatar proves that Buffer can now identify the intended publishing destination.

Choose + New again.
Select Connect a New Channel.
Select LinkedIn as the channel type.
Authorize the exact LinkedIn profile you want the agency to use.
Confirm the selected LinkedIn profile appears as a connected channel avatar.

That connection completes the delivery path. Buffer can now show the exact X profile and LinkedIn profile that later campaigns may use.

Missing a channel avatar?

Check that the network session belongs to your intended profile before retrying the connection.
Complete any network authorization request through Agent Computer takeover.
Ask for targeted help with the connection by selecting Help me troubleshoot my Buffer channel authorization..
Share only non-sensitive error wording with the NextWork community if authorization still fails.

Set the approval boundary

An approval boundary separates preparation from public action. Ask first rules pause matching actions until you review them.

Return to Grok Bot desktop.
Open Settings.
Select General.
Open Auto-review.
Add a narrow Ask first rule for publishing content.
Add a narrow Ask first rule for sending any external message.
Confirm both rules appear in Auto-review.

Your safety boundary is now in place. Publishing content or sending an external message requires your review first.

Why use narrow rules?

Narrow rules target the consequential actions in this campaign. Research can continue while public posts remain under human control.

Rules not appearing?

Confirm you are viewing Auto-review under General.
Confirm each rule uses Ask first.
Ask for help refining the rules by selecting Help me check my Grok Bot approval rules..

Verify the stack without publishing

A read-only check proves that Grok Bot can inspect the connected destinations without changing Buffer. This gives you evidence that access works before any campaign content exists.

Select New in Grok Bot's sidebar to start a temporary chat.

Before you send the check, consider whether the chat can name both channels without changing Buffer.

Verify the connected channels in read-only mode by sending this request:

Open Buffer in read-only mode.
List the connected X channel.
List the connected LinkedIn channel.
Do not create a post.
Do not edit a post.
Do not publish a post.
Do not schedule a post.

What does this check prove?

The first sentence limits the task to inspection.
The next two sentences request only the connected channel names.
The final four sentences block every post-changing action.

You should receive a response that names the connected X channel. It should also name the connected LinkedIn channel.

Confirm the response names your selected X profile.
Confirm the response names your selected LinkedIn profile.
Confirm the response reports no Buffer post changes.

Chat proposes a Buffer change?

Decline any approval request that would change Buffer.
Repeat the request with the read-only boundary preserved.
Ask for help correcting the task by selecting Help me keep this Buffer check read-only..

Before the final check, picture the safeguards you expect to see in Buffer. Picture the rules you expect to see in Auto-review.

Return to Buffer's publishing dashboard.
Confirm your selected X channel avatar is visible.
Confirm your selected LinkedIn channel avatar is visible.
Return to the Auto-review page from earlier.
Confirm publishing content uses Ask first.
Confirm sending any external message uses Ask first.

You should see both channel avatars in Buffer. You should also see both Ask first rules in Auto-review.

Your publishing stack is connected without creating or scheduling a post. Next, you'll test what happens when current market research has no niche to guide it.

Research Untargeted Trends

Your Buffer account now has your X and LinkedIn channels connected. Grok Bot still protects every publish or external message with an Ask first rule.

Fresh research can still leave an agency directionless. A trend scanner with no niche, buyer, or offer cannot choose a useful campaign angle.

In this step, get ready to:

Create a durable Trend Researcher with a narrow research-only role.
Run a deliberately untargeted scan that produces five dated market signals.
Inspect the report to identify the missing campaign constraints.

Create the Trend Researcher

A durable specialist Bot keeps one clear responsibility across future campaigns. The Trend Researcher needs strict evidence rules so its findings remain traceable.

Return to the Grok Bot desktop app.
Choose New in the sidebar.
Select Create new Bot.
Set the Bot name to Trend Researcher.
Open Edit Profile from the Bot menu.
Configure the Bot description with one responsibility by pasting this text:

You are the Trend Researcher for a multi-agent content agency.

You have one job: search current public web pages and X posts for market signals.

For every signal, preserve the observed date and public source link.

Keep observed evidence separate from interpretation.

Never contact anyone.

Never publish.

What Does This Role Enforce?

Current public web pages and X posts keep the evidence tied to recent signals.
Dates and links preserve traceability.
Separating evidence from interpretation makes unsupported conclusions easier to spot.
Read-only boundaries prevent contact or publication.

Complete the Bot profile setup.
Confirm that the new conversation is titled Trend Researcher.

Good work. Your first durable agency role now exists with a clear research boundary.

Does the Bot Have a Generic Role?

Return to the Trend Researcher conversation.
Open Edit Profile from the Bot menu.
Paste the research-only role description into the Bot description again.

Still stuck? Help me configure the Trend Researcher role.

Run the untargeted trend scan

This first task requests traceable market evidence without giving the Bot any campaign constraints. The resulting report tests how far source quality can take the agency on its own.

Before you send the task, do you think fresh source links alone can tell the agency what its content should achieve?

Switch back to the Trend Researcher conversation.
Test the untargeted research role by pasting this prompt:

Search the current web and X for five market trends that could inspire social content. Save the result to /workspace/flexible-agency/ai-consulting/raw-trends.md with the observed date, source link, evidence, and why each trend may matter. Do not infer a niche, target buyer, service offer, prospect criteria, or final recommendation.

What Does This Task Control?

Five results create a small sample that you can compare.
The observed date lets you judge when each signal appeared.
The source link lets you inspect the public evidence.
The evidence section records what the source supports.
The exclusions hold back campaign direction.

Wait for Trend Researcher to finish the research task.
Confirm that Trend Researcher reports saving /workspace/flexible-agency/ai-consulting/raw-trends.md.
Display the saved report for review by pasting this prompt:

Open /workspace/flexible-agency/ai-consulting/raw-trends.md and display the entire file without editing it.

Why Display the Saved File?

Displaying the saved file lets you inspect the durable artifact directly. Later Bots will depend on this exact workspace path.

You should see five market-trend entries. Every entry should include an observed date, a public source link, evidence, and an explanation of why the signal may matter.

Is the Report Incomplete?

Ask Trend Researcher to complete the original task if the file does not exist.
Ask Trend Researcher to revise the same file if an entry lacks a date or source link.
Ask Trend Researcher to remove any inferred targeting or final recommendation.

Need help checking the output? Help me audit my raw trend report.

Identify the operational gap

Sourcing proves what is happening in the market. Campaign targeting determines which evidence supports a useful commercial decision.

Before you audit the report, which campaign decision do you expect it to leave unresolved?

Count the entries in the displayed raw-trends.md report.
Verify that every entry includes an observed date.
Open every public source link from the report.
Compare each evidence statement with its linked source.
Check that every entry explains why the signal may matter.
Inspect the report for any inferred niche or target buyer.
Inspect the report for any service offer or prospect criteria.
Inspect the report for any final recommendation.

You should find five dated signals with links to their public evidence. You should find no campaign targeting or final recommendation.

The result is current and traceable. However, it cannot tell the agency which trend fits its offer, which consulting firms to approach, or what the content should achieve.

What Did the Experiment Expose?

Freshness and source quality make research auditable. Shared targeting constraints make that research operational.

The Trend Researcher has completed its job correctly. The missing direction belongs to the agency planning layer.

Review the Trend Researcher transcript for any contact or publishing action.
Confirm that the conversation contains only research and workspace file creation.

No post or external message should appear in the conversation. Your Buffer channels remain untouched.

That experiment did its job. You now have a sourced report that still cannot choose a commercially useful direction.

Next, you will give the agency shared targeting constraints so its research can guide a campaign.

Define the Offer and Find Prospects

Your Trend Researcher has proved that fresh evidence can still leave a campaign directionless. Its five market signals have dates and source links.

The scan has no shared standard for choosing a useful angle. It also cannot identify which firms deserve attention.

A durable agency brief gives every Grok Bot one commercial frame. Two focused roles can now turn the evidence into a campaign direction and a prospect shortlist.

In this step, get ready to:

Create an Agency Director with a durable B2B AI consulting brief.
Create a Prospect Scout with a public-evidence standard.
Run a visible handoff that produces a campaign plan and three sourced prospects.

Create the Agency Director and brief

An agency brief acts as a decision standard for the whole team. It gives planning a clear niche while preserving the source and approval rules you already established.

Choose New in the Grok Bot sidebar.
Select Create new Bot.

You should see the Bot creation screen where you can define a durable specialist.

Set the Bot's name to Agency Director.
Configure its role with this instruction: "Own campaign planning. Sequence visible handoffs between specialists. Reject unsupported claims. Keep every external message behind explicit human approval. Keep every public post behind explicit human approval."

You should now see Agency Director as a reusable Bot with a focused planning role.

Ask Agency Director to create /workspace/flexible-agency/ai-consulting/agency-brief.md with this request: "Create the pilot agency brief. Set the niche to AI consulting firms serving B2B organizations. Set the buyer to a founder or marketing lead. Set the offer to social content and market intelligence. Use English-language public sources. Prioritize implementation, governance, workflow redesign, scaling, and measurable ROI. Retain a source link and date for every factual claim. Require human approval for every send and post. Save the completed brief at the requested path. Display the saved brief when finished."

That's the targeting gap closed. The saved brief now gives every later decision one commercial standard.

You should see the pilot niche, buyer, offer, evidence rules, priority themes, and approval boundary in agency-brief.md.

Is the brief missing a constraint?

Ask the Agency Director to compare the saved file with your original request. Name the missing constraint so it updates the existing file instead of creating a second brief.

Still missing a requirement? Help me check my agency brief against the required pilot configuration.

Create the Prospect Scout and planning group

Prospect qualification needs its own durable role because company research follows a different standard from campaign planning. The Scout must connect each candidate to public evidence without collecting sensitive personal data.

Choose New in the Grok Bot sidebar.
Select Create new Bot.

You should see a fresh Bot creation screen for the prospecting specialist.

Set the Bot's name to Prospect Scout.
Configure its role with this instruction: "Find AI consulting firms with clear B2B offers. Record the public evidence that makes each firm relevant to the selected campaign angle. Draft one personalized opening line for each firm. Avoid sensitive personal data. Never send a message."

You should now see Prospect Scout as a reusable Bot with a clear research boundary.

Choose New in the sidebar.
Select Agency Director, the existing Trend Researcher, and Prospect Scout for the new group.

You should see all three Bots selected before the group opens.

Open the group.

The group should show Agency Director, Trend Researcher, and Prospect Scout as participants. This shared conversation makes each handoff visible.

Don't see all three Bots?

Return to the group creation screen if one role is missing. Confirm that both new Bots were saved before selecting the existing Trend Researcher.

Need help checking the roster? Help me identify which Bot is missing from my three-role planning group.

Run the evidence-to-prospect handoff

A visible handoff shows who owns each decision. The Director selects the commercially useful trend before the Scout begins prospect research.

Why keep the handoffs visible?

Visible handoffs preserve the reasoning between specialist roles. You can trace the selected trend back to its dated source before reviewing any prospect or outreach opener.

Send this kickoff request in the planning group: "Use /workspace/flexible-agency/ai-consulting/agency-brief.md with /workspace/flexible-agency/ai-consulting/raw-trends.md. Agency Director must select one AI implementation trend against the agency brief. Preserve the linked and dated evidence. Record the selection rationale. Define the approved campaign direction. Include the target buyer, content objective, and approval boundary. Save the result as /workspace/flexible-agency/ai-consulting/campaign-plan.md. Hand the approved angle visibly to Prospect Scout. Prospect Scout must find three AI consulting firms with clear B2B offers. Include public fit evidence relevant to the selected angle. Draft one personalized unsent outreach opener for each prospect. Avoid sensitive personal data. Save the result as /workspace/flexible-agency/ai-consulting/prospects.md. Do not send outreach. Do not publish content. Do not make any changes in Buffer."

You should see the Director explain its selection before handing the approved angle to Prospect Scout. The Scout should then return the saved prospect file without sending any message.

Before you inspect the files, do you expect every prospect to have a visible connection to the selected trend?

Open /workspace/flexible-agency/ai-consulting/campaign-plan.md in the Grok Bot cloud computer's file browser.
Confirm the plan selects one trend from raw-trends.md with linked and dated evidence.
Confirm the plan includes its selection rationale, campaign direction, target buyer, content objective, and approval boundary.
Open /workspace/flexible-agency/ai-consulting/prospects.md in the same file browser.
Confirm the file contains exactly three AI consulting firms with clear B2B offers.
Confirm each prospect has public fit evidence tied to the selected campaign angle.
Confirm each prospect has one personalized opener marked as unsent.
Confirm the group transcript shows that no prospect outreach was sent.

You now have a sourced campaign direction and a three-firm prospect shortlist. Every opener remains behind the approval boundary.

Is a prospect missing evidence?

Ask Prospect Scout to revise only the unsupported entry. Require a public source that connects the firm's B2B offer to the selected campaign angle.

Need help reviewing the shortlist? Help me audit each prospect for public fit evidence and an unsent personalized opener.

Your planning layer can now turn raw market evidence into a focused opportunity. Next, you will add the specialists that convert this direction into channel-specific content and a reviewable publishing proposal.

Assemble the Content Pipeline

Your Grok Bot planning group has selected a sourced campaign angle. It has tied that angle to three relevant consulting firms.

A plan plus a prospect list still cannot create market visibility. This step adds specialists that turn the evidence into channel-specific drafts.

The Content Creator writes distinct X and LinkedIn drafts from the approved evidence. The Publisher prepares an exact Buffer proposal while your approval rules block public action.

In this step, get ready to:

Create durable Content Creator and Publisher Bots with separate responsibilities.
Bring all five Bots into the B2B AI Consulting Agency group.
Run one coordinated pipeline that stops at an approval packet.

Create the content specialists

Content Creator owns channel-specific drafting. Its role keeps factual evidence intact without gaining publishing authority.

Choose New in Grok Bot's sidebar.
Select Create new Bot.

You'll see a new Bot profile ready for its durable role.

Choose Edit Profile on the new Bot.
Enter Content Creator as the Bot name.
Set the Content Creator's role by pasting this description:

You are Content Creator.
Your one job is to turn selected evidence into distinct X and LinkedIn drafts.
Preserve every factual source link and date.
Do not invent metrics or customer claims.
Never publish.

What does this role protect?

Distinct drafts let each social channel use its own structure.
Source preservation keeps factual statements traceable to dated evidence.
The publishing boundary keeps Content Creator focused on preparation.

Confirm the profile changes.
Check the Bot profile for the name Content Creator.

That's one specialist ready: Content Creator now owns channel-specific drafting without publishing authority.

Content Creator role missing?

Return to the Bot's Edit Profile view.
Check that the Bot name is Content Creator.
Paste the role description again if the instructions are missing.

Still stuck? Help me configure my Content Creator Bot.

Publisher owns the final quality check. Its approval boundary prevents a prepared proposal from becoming a public post without your decision.

Choose New in Grok Bot's sidebar.
Select Create new Bot.

You'll see another new Bot profile ready for a focused role.

Choose Edit Profile on the new Bot.
Enter Publisher as the Bot name.
Set the Publisher's role by pasting this description:

You are Publisher.
Your one job is to quality-check approved copy and prepare Buffer actions.
Never schedule or publish without my explicit approval of the exact text, destination, and time.

What does this approval boundary do?

The quality check catches copy that conflicts with approved evidence.
The approval packet makes the proposed public action reviewable.
Explicit approval limits permission to the exact campaign you inspect.

Confirm the profile changes.
Check the Bot profile for the name Publisher.

Your second specialist is ready: Publisher can prepare a precise action without gaining permission to execute it.

Publisher boundary missing?

Return to the Bot's Edit Profile view.
Check that the Bot name is Publisher.
Paste the role description again if the approval rule is missing.

Need help? Help me fix my Publisher Bot's approval boundary.

Build the five-Bot agency group

A shared group gives every specialist the same campaign context. A visible handoff records which Bot completed a stage plus which Bot owns the next one.

Choose New in Grok Bot's sidebar.
Select these five Bots: Trend Researcher, Agency Director, Prospect Scout, Content Creator, and Publisher.
Open the new group.

You'll see one group conversation containing all five durable roles.

Use the group header's name control to rename the group B2B AI Consulting Agency.
Confirm that the group header shows B2B AI Consulting Agency.

The renamed group is now the agency's shared campaign room.

Why use one group?

The group keeps every handoff in one transcript. You can see when evidence becomes a campaign angle.

You can also trace that angle through prospect fit, channel-specific drafting, and final approval preparation.

Missing a Bot from the group?

Check the group participant list for all five Bot names.
Create a fresh group from New if any role is absent.
Rename the complete group B2B AI Consulting Agency.

Still missing a role? Help me assemble my five-Bot group.

Run the coordinated pipeline

One kickoff message gives the agency a fixed sequence. It also defines the files plus the exact point where automation must stop.

Launch one coordinated run in the B2B AI Consulting Agency group by sending this kickoff message:

Run one coordinated campaign pipeline using /workspace/flexible-agency/ai-consulting/agency-brief.md and /workspace/flexible-agency/ai-consulting/campaign-plan.md.

1. Trend Researcher refreshes the chosen signal using current public evidence. Retain every source link and date.
2. Agency Director approves one campaign angle against the agency brief.
3. Prospect Scout confirms that the angle fits the three consulting-firm prospects in /workspace/flexible-agency/ai-consulting/prospects.md.
4. Content Creator creates distinct channel-specific drafts. Save them as /workspace/flexible-agency/ai-consulting/x-post.md and /workspace/flexible-agency/ai-consulting/linkedin-post.md.
5. Content Creator records the selected angle, draft assets, proposed destinations, and proposed publishing times in /workspace/flexible-agency/ai-consulting/content-calendar.md.
6. Publisher quality-checks the approved copy. Return one approval packet with the exact proposed X text, exact proposed LinkedIn text, connected X and LinkedIn destinations from the prior read-only verification, and proposed schedule times.

Keep every handoff visible in this group.
Preserve source links and dates for every factual statement.
Keep all prospect outreach unsent.
Do not open Buffer.
Do not create, edit, schedule, or publish any Buffer post.
Stop after Publisher returns the approval packet with the status awaiting explicit human approval.

How does the pipeline work?

The fixed sequence gives every Bot one defined input plus one defined output.
The shared workspace path keeps the campaign assets available to the whole agency.
The Publisher's stopping rule converts a possible public action into a reviewable proposal.

Before you follow the run, which Bot do you expect to own the final handoff?

Follow the group transcript from Trend Researcher's refresh through Publisher's final response.
Open /workspace/flexible-agency/ai-consulting/x-post.md from the completed Content Creator handoff.
Open /workspace/flexible-agency/ai-consulting/linkedin-post.md from the completed Content Creator handoff.
Open /workspace/flexible-agency/ai-consulting/content-calendar.md from the completed Content Creator handoff.
Read the Publisher's final approval packet.
Review the transcript for any Buffer action or sent outreach.

What should you see?

The transcript shows a research refresh followed by angle approval.
The next handoff confirms prospect fit before drafting starts.
The x-post.md file contains an X-specific draft with dated source links.
The linkedin-post.md file contains a distinct LinkedIn draft with dated source links.
The content-calendar.md file records the proposed destinations.
The content calendar marks the proposed publishing times as pending approval.
The final packet includes the exact proposed text for each network.
The final packet names the connected destinations plus the proposed schedule times.
The final status says that the packet is awaiting explicit human approval.
The transcript shows no Buffer opening or scheduling.
The transcript shows no sent prospect outreach.

Pipeline stopped before the approval packet?

Check the final visible handoff to identify which Bot still owns the work.
Check that all five Bot names appear in the group participant list.
Resend the kickoff message if the original message was incomplete.

Need help tracing the gap? Help me diagnose my missing agency handoff.

You now have a five-agent pipeline that turns sourced evidence into distinct social drafts plus a precise approval packet. Every consequential action still belongs to you.

Your agency has prepared the complete campaign without touching Buffer. Next, you'll review the exact proposal before authorizing the two scheduled posts.

Approve and Schedule the Campaign

Your five-Bot Grok Bot group has completed its visible handoffs. You now have channel-specific drafts plus a Publisher approval packet.

Public distribution through Buffer affects your live X and LinkedIn accounts. This step keeps that action behind an exact human approval boundary.

In this step, get ready to:

Review the exact X text plus the exact LinkedIn text against the approved campaign direction.
Approve the destinations plus publishing times before Publisher schedules either post.
Verify both scheduled posts in Buffer Calendar plus record the completed action.

Review the Publisher approval packet

The approval packet binds the campaign assets to specific destinations plus publishing times. Every detail needs to match the files you already reviewed.

Publishing affects live accounts, so your Publisher still has no permission to act until your reply matches the packet exactly.

Return to the B2B AI Consulting Agency group from earlier.
Locate the Publisher approval packet in the group transcript.
Compare the proposed X text with x-post.md.
Compare the proposed LinkedIn text with linkedin-post.md.
Review the selected evidence plus campaign direction in campaign-plan.md.
Check the proposed X destination against your connected X profile.
Check the proposed LinkedIn destination against your connected LinkedIn profile.
Compare the proposed publishing times with content-calendar.md.

What counts as explicit approval?

Explicit approval gives Publisher a narrow instruction that can be checked before any public action.

The X approval identifies the exact reviewed X text.
The LinkedIn approval identifies the exact reviewed LinkedIn text.
The destination approval identifies both connected social profiles.
The schedule approval identifies the publishing time for each post.

Request a revised approval packet if any statement lacks support or any destination or time is wrong.
Review the revised packet against the same three campaign files.
Reply to Publisher with explicit approval for the exact X text plus the exact LinkedIn text at the named destinations plus times.

Good work. Your Publisher now has a narrow mandate that covers two exact posts without granting broader publishing permission.

Schedule the approved posts

Channel-specific customization preserves the approved voice for each network. Publisher must keep the X draft separate from the LinkedIn draft throughout the Buffer action.

Ask Publisher to open Buffer in the Grok Bot cloud browser.
Tell Publisher to choose + New Post.
Tell Publisher to select the connected X profile through + Channels.
Tell Publisher to select the connected LinkedIn profile through + Channels.
Tell Publisher to choose Customize for each network.
Tell Publisher to place the approved x-post.md text in the X box.
Tell Publisher to place the approved linkedin-post.md text in the LinkedIn box.

Why customize each network?

The two posts use one verified campaign angle while serving different reading habits. Separate network boxes keep each approved draft intact.

Tell Publisher to choose Set Date and Time.
Tell Publisher to apply the approved publishing times from the packet.

Check the three-part match

Publisher must compare three details before scheduling. A complete match covers content, destination, plus time.

The X box contains the exact approved X text.
The LinkedIn box contains the exact approved LinkedIn text.
The selected channels match the approved X plus LinkedIn destinations.
The displayed publishing times match the approved schedule.

Ask Publisher to compare the displayed action with the complete approval packet.
Tell Publisher to stop the action if any detail differs from the approval.

This is the highest-stakes click in the project, with your exact-text approval plus Ask first checkpoint keeping the action contained.

Instruct Publisher to choose Schedule Post when every displayed detail matches.
Review the Ask first checkpoint when Grok Bot pauses the publishing action.
Approve the checkpoint when its action still matches the exact text plus destinations plus times.

That is the consequential part done. Buffer now holds one approved X post plus one approved LinkedIn post for their approved times.

Details do not match?

Stop the scheduling action before approving the checkpoint.
Return to the Publisher approval packet in the agency group.
Ask Publisher to replace the mismatched content or destination or time.
Restart the Buffer action after Publisher returns a corrected packet.

Need help isolating the mismatch? Help me compare my Publisher approval packet with the Buffer scheduling action.

Verify the calendar and action log

Before you inspect the calendar, how many scheduled entries do you expect to find across the two connected channels?

Return to the Buffer publishing dashboard from earlier.
Select Calendar.

You should see one scheduled X entry plus one scheduled LinkedIn entry at the approved times.

Open the scheduled X entry from the Calendar view.

You should see the connected X destination plus the exact approved x-post.md text at the learner-approved time.

Close the scheduled X entry.
Open the scheduled LinkedIn entry from the Calendar view.

You should see the connected LinkedIn destination plus the exact approved linkedin-post.md text at the learner-approved time.

What belongs in the action log?

The action log proves that Publisher stayed within the approved scope.

It records the connected X plus LinkedIn destinations.
It records the approved scheduled time for each post.
It records the visible Buffer status for each entry.
It confirms that no prospect outreach was sent.

Return to the B2B AI Consulting Agency group.
Ask Publisher to return the short action log in the group transcript.
Ask Publisher to update content-calendar.md plus x-post.md plus linkedin-post.md with the approved scheduling state.
Review the returned action log against the four required fields.

The log should name both destinations plus times plus visible statuses. It should also confirm that all three prospect openers remain unsent.

Missing a scheduled entry?

Check that Calendar is showing the approved publishing date.
Check that both connected channels are included in the current Calendar view.
Ask Publisher to compare the missing entry with the approval packet before taking another action.

Still missing one post? Help me trace a missing scheduled post without creating a duplicate.

You have completed the full evidence-to-publication workflow. Both channel-specific posts are scheduled behind human approval while every prospect message remains unsent.

Secret mission

Switch the Agency to a New Niche

Reuse your five-agent agency for local fitness studios through one new configuration. Prove the new campaign stays isolated from the scheduled AI consulting campaign and stops before any publishing action.

Clean Up Your Resources

Your campaign stays within Buffer Free's limits because it uses two connected channels with one scheduled post on each. Decide whether to keep the campaign active, pause its scheduled posts, or remove its resources.

Resources you used:

Two approved Buffer posts scheduled to the selected X and LinkedIn channels.
Buffer authorizations for the selected X and LinkedIn profiles.
Five durable Grok Bot roles: Trend Researcher, Agency Director, Prospect Scout, Content Creator, and Publisher.
The B2B AI Consulting Agency group with visible handoffs between all five Bots.
The shared /workspace/flexible-agency/ folder containing the AI consulting campaign and isolated local-fitness campaign.

Keep everything running

No action is needed. Choose this option if you want the approved posts to publish at their scheduled times.

Buffer keeps the approved X post and LinkedIn post scheduled.
Your selected X and LinkedIn profiles remain authorized in Buffer.
The five durable Bots remain available for future campaigns.
The B2B AI Consulting Agency group retains its visible handoff history.
The Ask first rules continue protecting publishing and external messages.
Both campaign directories remain stored under /workspace/flexible-agency/.
Buffer Free supports up to three connected channels and ten scheduled posts per channel. This project uses two channels with one scheduled post on each.

Pause - I'll come back to this later

Remove both scheduled posts before publication. Your Bots, group, channel connections, and workspace files stay ready for another campaign.

Cancel Scheduled Posts

Return to Buffer's Calendar view in the Grok Bot cloud browser.
Open the scheduled X post.
Open the post's three-dot menu.
Choose Delete.
Open the scheduled LinkedIn post.
Open the post's three-dot menu.
Choose Delete.

Your public campaign is now safely paused. You should see neither approved campaign post in Buffer's Calendar view.

Delete - I don't want to use this again

Remove the scheduled posts, channel authorizations, Bots, group, and shared files if you no longer want this agency. Deleting the shared workspace is final, so this path cancels anything that could publish before removing the stored work.

Cancel Scheduled Posts

Return to Buffer's Calendar view in the Grok Bot cloud browser.
Open the scheduled X post.
Open the post's three-dot menu.
Choose Delete.
Open the scheduled LinkedIn post.
Open the post's three-dot menu.
Choose Delete.

You should see neither approved campaign post in Buffer's Calendar view. This confirms that the campaign cannot publish while you complete the cleanup.

Revoke Channel Access

Open the authorized-app controls for the connected X profile.
Revoke Buffer's access to the X profile.
Open the authorized-app controls for the connected LinkedIn profile.
Revoke Buffer's access to the LinkedIn profile.
Confirm that Buffer no longer appears in either profile's authorized-app list.
Return to the Buffer session in the Grok Bot cloud browser.
Sign out of Buffer on the shared computer.

Remove Agency Resources

Delete the /workspace/flexible-agency/ folder from Grok Bot's shared workspace.
Confirm that /workspace/flexible-agency/ no longer appears under /workspace.
Hide or delete Trend Researcher, Agency Director, Prospect Scout, Content Creator, and Publisher through each Bot's menu.
Hide or delete the B2B AI Consulting Agency group through its group controls.
Confirm that the five Bots and group no longer appear in your active Grok Bot workspace.

You have closed the agency's public and stored project footprint. Your Buffer Free account remains available for other work.

Nice Work!

You did it! You built a five-role Grok Bot agency that turns current evidence into a focused B2B AI consulting campaign while keeping every public action under your control.

You've learned how to:

Convert dated market evidence into a source-linked campaign plan for AI consulting firms serving B2B organizations.
Coordinate five durable Bot roles through visible handoffs that produce a sourced prospect shortlist plus unsent outreach drafts.
Move one verified market angle through channel-specific X and LinkedIn drafts to a human-approved publishing schedule in Buffer.
Secret Mission: Reconfigured the same five-Bot agency for local fitness studios with isolated files and zero stale AI consulting references.

Ready to quiz yourself?