Clone a LinkedIn Campaign for 5 Segments

You can clone a LinkedIn campaign for five segments in one coding agent session: the agent calls duplicate_campaign once per segment, then update_campaign_targeting, reads the criteria check, and starts a prospect preview for each copy. Your part is to read five criteria checks, mark the previews, and confirm five launch previews. This guide walks through the full session for a GTM engineer who works in Claude Code or Cursor.

The situation: one campaign works, sales wants it for five segments by Friday, and each copy built by hand is an afternoon of form fields. The session below uses the GTM Bud MCP server. For the wider picture of what the agent can do, see AI LinkedIn outreach for GTM engineers.

What does a clone keep, and what does it drop?

A duplicated campaign is a new draft that keeps the parts you want to reuse and drops the parts that belong to the original audience. A draft sends no messages, so the loop is safe to run before anyone has reviewed anything.

Part of the campaign	In the copy?	What you do per segment
Targeting	Kept	Replace it with update_campaign_targeting
Message sequence	Kept	Read it, and rewrite what does not fit
Reply playbook	Kept	Add a campaign rule if the segment needs one
Sending accounts, daily targets	Kept	Change them on gtmbud.com if the segment needs it
Leads, replies and results	Not kept	Nothing. The copy starts clean
Prospect previews	Not kept	Start a new preview

This is why cloning beats a new draft: the sequence and the playbook took the longest to get right, and they come along.

Step 1: Put the five segments in a file

Write each segment as one audience sentence, in a file the agent can read. A sentence is the input: the agent writes the criteria from it, and GTM Bud checks them. There is no list upload through the connector, so do not prepare rows.

The company in this guide sells energy monitoring hardware. The working campaign targets cold storage operators. The five new segments:

source_campaign: Cold storage operators
segments:
  - name: Breweries
    audience: Plant managers and heads of engineering at breweries with 50 to 500 employees in the US and Canada
  - name: Dairy processors
    audience: Plant managers and heads of engineering at dairy processing companies with 100 to 1,000 employees in the US and Canada
  - name: Injection moulders
    audience: Plant managers and heads of engineering at plastics injection moulding companies with 50 to 500 employees in the US and Canada
  - name: Commercial laundries
    audience: General managers and heads of operations at commercial laundry companies with 50 to 500 employees in the US and Canada
  - name: Colocation data centers
    audience: Facility managers and heads of data center operations at colocation providers with 50 to 1,000 employees in the US and Canada

Keep the role and the region stable where you can, and change one thing per segment. When five things change at once, you cannot tell later which change made the difference.

Step 2: Run the duplicate and retarget loop

Ask the agent to loop over the file. Each segment is two calls: duplicate_campaign, then update_campaign_targeting with the audience sentence. The agent writes the criteria from the sentence, and GTM Bud checks them and returns each flagged criterion with the second call.

The tool names below are the real ones. The data is invented.

you    Read segments.yaml. For each segment, duplicate the source
       campaign and set the targeting. Stop after the criteria checks
       and show me all five.
agent  list_campaigns
       Found "Cold storage operators" (active).
agent  duplicate_campaign            x5
       5 drafts created. Each one sends no messages.
agent  update_campaign_targeting     x5
       Criteria checks: every criterion passed for all five.
agent  Before the previews, three notes of my own:
       Breweries             "Breweries" also matches brewpubs and
                             taprooms. Include them?
       Injection moulders    I used both spellings, "moulding" and
                             "molding".
       Commercial laundries  "Commercial laundry" also matches
                             laundromat chains. Include them?

you    Breweries: production breweries only, not brewpubs or taprooms.
       Laundries: industrial laundries that serve hotels and hospitals,
       not laundromats.
agent  update_campaign_targeting     x2
       Both checks passed.

Read the criteria check and the agent’s notes, do not skip them. The check flags a criterion that cannot be read from a profile, a company page or public web content. It does not judge who else a criterion matches, so that part is the agent’s question and your answer. Two of five segments had a point that would have filled the campaign with the wrong companies. That ratio is invented for this example, but the pattern is real: a segment name that is clear to you often matches more companies than you meant.

Step 3: Preview prospects for each segment

A prospect preview shows real people who match the targeting, and you mark each one Good fit or Not a fit. Run one preview per copy. This is the step that tells you if the sentence found the people you meant.

you    Start a prospect preview for all five. When they are ready,
       show me the prospects per segment.
agent  start_prospect_preview        x5
agent  get_prospect_preview          x5
       Real matching prospects returned for each segment.

you    Open the second brewery prospect. What does the company make?
agent  get_prospect_profile
       Profile returned.

you    Mark them as I say.
agent  mark_prospect_fit
       Marks saved for each segment.

Do not hand the marking to the agent. It can show you each prospect and the reason it matches. The judgment is yours, and by our estimate it takes about ten minutes per segment.

A worked example: the brewery segment in detail

The names below are invented.

The audience sentence. “Plant managers and heads of engineering at breweries with 50 to 500 employees in the US and Canada.”

What the criteria check said. The check passed every criterion. “Breweries” was the problem that the agent saw: it would also match brewpubs and taprooms, which are restaurants with a small brewing system. The agent asked if those should be included. The answer was no, and the sentence became “production breweries, not brewpubs or taprooms”.

Two previewed prospects.

• Miguel A., Director of Engineering at a regional production brewery with about 220 employees in Colorado. Marked Good fit: right role, a plant with refrigeration and compressed air, inside the size range.
• Stacy W., General Manager at a brewpub group with 60 employees in Oregon. Marked Not a fit: inside the size range, but the staff are restaurant staff and there is no plant to monitor.

One value offer. The offer of the source campaign was about freezer rooms, which means little to a brewer. The agent called get_value_offer_brief, read the GTM Bud instructions, researched the company website and wrote: “A two-week reading of the power that your glycol chillers and your packaging line use, with a written list of the three largest sources of waste.” It called save_value_offers, and GTM Bud checked and saved the offer.

The results question, for two weeks later. “Give me the results for the source campaign and the five copies in one table, with the as-of time for each row.”

Step 4: Readiness, then launch with confirmation

Before launch, ask for the launch readiness check on every copy. It lists what is missing. Then ask for launch one campaign at a time. Each launch shows a preview and waits for your confirmation.

you    Run the readiness check on all five drafts.
agent  get_launch_readiness          x5
       Breweries             Ready.
       Dairy processors      Ready.
       Injection moulders    Ready.
       Commercial laundries  Missing: value offers.
       Colocation            Missing: the messages check found a
                             problem in the first message.

you    Fixed both. Launch Breweries.
agent  launch_campaign
       Launch preview shown. Waiting for your confirmation.
you    Confirmed.
agent  launch_campaign
       Campaign launched.

Repeat the last four lines for each segment. There is no way to confirm five launches with one word, and that is intended. Each preview is the last time a person reads the campaign before prospects do. The preview shows the first message of each channel, so read the full sequence of each copy with get_campaign_sequence first.

Two things are done on gtmbud.com and not in the terminal: changing the sending accounts of a campaign and setting the sending schedule. Each copy keeps the sending accounts of the original. Change them on gtmbud.com if a segment should send from a different account. Launch also needs an active plan or trial.

Five prompts for the cloning session

• “Read segments.yaml, duplicate the source campaign once per segment, set the targeting of each copy, and stop after the criteria checks.”
• “For the laundries copy, change the audience to general managers and heads of operations at industrial laundries that serve hotels and hospitals, with 50 to 500 employees in the US and Canada. Show me the new criteria check.”
• “Start a prospect preview for every copy and print the prospects per segment with one line on why each one matches.”
• “Show me the message sequence of the dairy copy next to the source sequence, and mark every sentence that mentions cold storage.”
• “Run the readiness check on all five drafts and write the missing items to outbound/launch-todo.md, grouped by campaign.”

Three objections a GTM engineer raises

“I have the account lists for all five segments already. Let me load them.” The connector does not take a list upload. Targeting is an audience sentence that the agent turns into criteria. GTM Bud checks the criteria and uses them to find leads. You lose control of the exact rows, and you get a criteria check and a preview of real matches in return. If the segments exist only as fixed lists in another system, this loop is not the right tool for them. To turn a list into a sentence, start from how to build an ICP for outbound.

“Five confirmations is not automation.” Correct. The agent removes the form work: ten tool calls where you would have filled five long forms. It does not remove the decision. Launch needs your confirmation of a preview, once per campaign, in a session with you in it. A scheduled job that launches campaigns is not possible.

“Five campaigns will split my volume five ways.” They share the sender accounts that you have, and the sending schedule is set on gtmbud.com. More campaigns do not create more sending capacity. If you have two senders, think about launching two segments this week and three later, and leave the other three as drafts until a sender is free. A draft sends no messages.

Frequently asked questions about cloning a LinkedIn campaign

What does a duplicated GTM Bud campaign keep from the original?

The copy is a new draft that sends no messages. It keeps the targeting, the message sequence, the message settings, the reply playbook, the sending accounts and the daily targets of the original. It does not keep the leads, the replies, the results or the prospect previews, so each copy starts clean.

Can I change the audience of a campaign that is already live?

Not with update_campaign_targeting. That tool changes a draft. For a new segment, duplicate the live campaign and change the targeting of the copy. The original keeps running with its own audience.

Can the agent launch all five copies in one step?

No. Each launch shows its own preview and waits for your confirmation. Five segments mean five previews to read and five confirmations. The agent can prepare all five and bring them to you one after the other.

Do five campaigns need five LinkedIn accounts?

No. One GTM Bud account can hold several LinkedIn sender accounts, and each copy keeps the sending accounts of the original. Change them on gtmbud.com if a segment should send from a different account. The setup steps for the terminal are on GTM Bud for coding agents.

Should each segment get its own messages?

Often the first message and the value offer should change, because the pain differs by segment. The copy keeps the sequence of the original, so the default is the same messages. Read the sequence of each copy and decide for each segment.

Clone the campaign once per segment, and review five times

To clone a LinkedIn campaign for five segments by Friday, let the agent run the loop and keep the three reviews for yourself: the criteria check, the prospect preview and the launch preview. Setup is one command, claude mcp add --transport http gtm-bud https://gtmbud.com/mcp, and the steps for Cursor are on GTM Bud for coding agents. When the copies run, read LinkedIn outbound results in your terminal to compare them. If the role is new to your team, our explainer on what a GTM engineer is gives the background, and the page on AI LinkedIn outreach for GTM engineers shows the rest of what the agent can do with GTM Bud.