Agency new business in one hour a week is possible when a campaign already runs and the hour has a fixed shape: 15 minutes to review results, 30 minutes to approve replies that your AI wrote from a playbook, and 15 minutes to test one targeting change on a duplicated campaign. This routine is for the owner of a 5 to 30 person marketing, design or development agency whose week belongs to delivery.
It covers the agency’s own pipeline. If you want to run campaigns for your clients, read white label LinkedIn outreach for agencies.
The routine uses the GTM Bud connector, so the whole hour happens in a chat with Claude or ChatGPT. The minutes are our estimate for a normal week. A week with many replies takes longer, and that is stated plainly in each step.
Why does delivery always win against new business?
Delivery wins because it has a client and a deadline, and new business has neither. Nobody emails you on Thursday to ask where this week’s prospecting is. The task with no owner and no date loses every time, to work that is honestly more urgent.
Owners usually answer this with a resolution: “I will do two hours of outreach every morning.” It lasts until the first client escalation. The resolution fails for a structural reason, not a personal one. It asks for a large, open block of time from a person who has none.
A routine that survives has three properties:
- It is small. One hour is hard to argue away.
- It is closed. Each part has a time limit and a finished state.
- It does not depend on you between sessions. Sending and lead finding continue while you are in client work.
That third property is what a launched campaign gives you. If you have no campaign yet, build one first. The steps are on our page about AI LinkedIn outreach for agency owners.
The one hour at a glance
| Block | Minutes | Job | Finished when |
|---|---|---|---|
| Results | 15 | Read each campaign, pick one action | You have written one sentence about what you change |
| Replies | 30 | Approve, edit or reject each draft | No Interested reply waits for an answer |
| Test | 15 | One change on a duplicated campaign | The copy has passed a prospect preview |
Put it in the calendar as a meeting with yourself, with the same title every week. Open the chat, not the dashboard.
Minutes 0 to 15: What should you look at in the results?
Look at three things for each campaign: how many replies came in, how they split across Interested, Warm and Not interested, and what the Not interested leads said. Then decide one action. More than one action a week means that you cannot tell which change worked.
Start with one prompt: “List my campaigns and give me the results for each with the as-of time. For the campaign with the most replies, summarize what the Not interested leads said in three lines.”
The as-of time matters. The numbers are a snapshot, and you want to compare this Friday with last Friday, not with a number you half remember.
Use this table to choose the action:
| What you see | What it usually means | The one action |
|---|---|---|
| Few replies of any kind | The offer or the first message is weak | Rewrite the value offer in block 3 |
| Many replies, mostly Not interested | The audience is wrong | Narrow the targeting on a copy in block 3 |
| “We have an agency” in most Not interested | Right people, wrong timing | Add a playbook rule for that answer |
| Interested replies that wait from earlier weeks | The routine slipped | Give block 2 the extra time, skip block 3 |
Write your action as one sentence in the chat, so that next week’s session starts with it. GTM Bud does not send a scheduled report. This block is the report, and you run it.
Minutes 15 to 45: How do you clear replies in 30 minutes?
You clear replies in 30 minutes by letting a playbook do the writing and keeping your attention for the review. Your AI reads each reply, writes an answer that follows your rules, and shows it to you for review. GTM Bud instructs the AI to send only the exact text that you approve.
A reply playbook is a short set of rules for how your agency answers leads. GTM Bud keeps one for the account and one for each campaign. For an owner who sells, these are the rules that save the most time:
- Interested leads get an offer of a 25 minute call with the owner, with two time options that the playbook holds.
- Budget questions get no number. Scope decides price, and the call decides scope.
- Requests for case studies get one relevant link from the website, and no attachment.
- Projects you do not take get a polite no in one line. Name the kinds: for example, one-off logo work or rescue jobs with a two week deadline.
- Warm leads get one question about timing and nothing else.
Then work in label order. Ask: “Read the new replies and label each one. Write answers for the Interested ones with the account playbook and show me each.”
An example. A 20 person brand and packaging studio runs a campaign to marketing directors at food and drink brands. A lead wrote: “We are redoing our range next spring. Too early?” The AI labeled it Warm, and wrote:
“Not too early, Jan. Spring ranges usually need the design locked by late autumn to make the print slots. When do you plan to brief it?”
The owner changed “usually” to “in our projects”, because the first version made a general claim she could not back. Then she approved it.
Three rules for the 30 minutes:
- Interested first, always. If time runs out, the Warm list waits a week. The Interested list does not.
- Edit the playbook, not only the reply. If you make the same edit twice, add a rule. That is how next week gets shorter.
- Respect the limit. Each LinkedIn account sends up to 40 reviewed replies a day. A normal week for one owner is well inside that. If it is not, your problem is capacity for calls, and you should slow the campaign.
For the wording of hard replies, see our guide on how to handle cold outreach replies.
Minutes 45 to 60: Why change targeting on a duplicated campaign?
Change targeting on a duplicate because the original campaign is your control. If you edit the campaign that works, you lose the thing you compare against, and a bad change costs you the whole pipeline instead of one test.
The method is the same every week:
- Ask: “Duplicate the food and drink campaign.”
- Give the one change as an audience sentence: “Change the audience of the copy to marketing directors at pet food and pet care brands with 20 to 250 employees in Germany and the Netherlands.”
- Read the criteria check. In this example it passed, and the AI asked if retailers that sell pet products should be included. They should not: the studio designs for brands that own the packaging.
- Ask for a prospect preview and mark each person. Mieke V., Marketing Director at a pet food brand with about 90 employees in Utrecht, was marked Good fit. Stefan B., Category Manager at a pet store chain with 1,400 employees, was marked Not a fit: a retailer, and a buyer of products, not of packaging design.
- Stop there if the 15 minutes are over. The copy waits as a draft.
Launch is a job for the next week’s block 3, and it takes the readiness check and your confirmation of a preview. The copy keeps the sending account of the original. Change it on gtmbud.com if the test should send from another account. If the test needs a new value offer, such as “three shelf mock-ups of your current pack next to the two category leaders”, your AI writes it with GTM Bud instructions and GTM Bud checks and saves it.
Trade-off: A copy splits your sending across two campaigns, so each one learns more slowly. One change a week on a copy is slow and readable. Five changes at once is fast, and you learn nothing from it.
When a test has run for a few weeks, ask the results question: “Compare the food and drink campaign with the pet care copy as of now. Which has more Interested replies for the messages sent?” Keep the better one. End the other.
What do you do in a week when delivery takes everything?
Use the campaign controls before the week starts, not after it goes wrong. Each launched campaign has a sending mode with three settings, and each one fits a different kind of bad week.
| Your week | Sending mode | Effect |
|---|---|---|
| Busy, but the hour will happen | All | Nothing changes |
| A launch week, the hour is at risk | Follow-ups only | No new first messages, earlier conversations go on |
| Out of office with no review possible | Stopped | Nothing is sent until you start it again |
Ask: “Set the food and drink campaign to follow-ups only until I say otherwise.” When you return, ask for it to be set back to all. A change that starts sending needs your confirmation, so nothing restarts by accident.
Three objections owners raise about a one hour routine
“My replies will not fit in 30 minutes.” In some weeks they will not. The 30 minutes is an estimate for a normal week on one LinkedIn account, and the system allows up to 40 reviewed replies a day for each account. When the Interested list alone takes more than the block, drop the test block for that week. If it happens three weeks in a row, the campaign is producing more than you can handle, and follow-ups only is the right setting until you catch up.
“I do not want to talk to a chat window to see my numbers. Send me a report.” GTM Bud does not send a scheduled report, so that option is not there. You ask, and the answer carries its as-of time. The cost is that nothing arrives unless you open the session. The gain is that the question can change every week: last week by label, this week by what the Not interested leads said.
“One change a week is too slow. I need clients now.” Then the routine is the wrong tool for this month. It keeps a pipeline moving. It does not fill an empty one fast. If the need is urgent, do the two week build first and come back to the hour when a campaign runs.
This is the honest limit of the one hour claim. The hour covers the routine. It does not cover the calls that the routine produces, and it does not cover the first build. The broader case for why agencies with no free hours need a system at all is in our article on lead generation for marketing agencies.
Frequently asked questions about agency new business in one hour a week
How many hours a week should an agency owner spend on new business?
Enough to keep it regular. One fixed hour every week does more for a pipeline than ten hours in one quiet month, because outreach rewards consistency. Sales calls with interested leads come on top of that hour, and they grow as the routine works.
Which day is best for the weekly new business hour?
The day that delivery interrupts least. For many agencies that is Friday morning, after the weekly client work is shipped and before the planning for next week. The exact day matters less than keeping the same day every week.
What if I miss a week?
The campaign keeps sending, and the replies wait for you. The cost is slower answers to interested leads. If you know a hard week is coming, set the campaign to follow-ups only before it starts, so that fewer new conversations begin while you are away.
Can someone on my team do the weekly hour for me?
The review is done by the person who is signed in to the GTM Bud account, and this routine assumes that person is the owner. The owner knows which projects to decline and what the agency can promise. Hand the hour over only when the playbook is good enough that your edits are rare.
Does one hour a week work without a campaign that already runs?
No. The hour is a maintenance routine for a launched campaign. Building the first campaign is a separate job that takes a few sessions. If the reason you need one is urgent, follow the two week plan in agency lost its biggest client, and see AI LinkedIn outreach for agency owners for the setup.
Put the hour in the calendar before the next deadline arrives
Agency new business in one hour a week works because the hour is small, closed and the same every week: 15 minutes on results, 30 on replies, 15 on one test. The campaign does the sending between sessions, and you keep every decision. To start, add the connector with the steps on GTM Bud for ChatGPT or create an account at GTM Bud, then build your first campaign from AI LinkedIn outreach for agency owners.