---
title: "How to Get Clients for a Marketing Agency from Inbound DMs"
slug: ai-setter-for-marketing-agencies
author: "Leonardo Maldonado"
category: "Coaching Niches"
articleType: how-to-guide
tags: ["marketing agency lead generation","get agency clients from DMs","white label DM setter","AI setter for client accounts"]
publishedAt: 2026-09-09T09:00:00.000Z
updatedAt: 2026-10-17T12:00:00.000Z
canonical: https://setluca.com/blog/ai-setter-for-marketing-agencies
---# How to Get Clients for a Marketing Agency from Inbound DMs

> How to get clients for a marketing agency from inbound DMs: reply fast, qualify on ad spend, book the call. To get clients for a marketing agency, publish proof-of-work content, reply fast to every inbound DM in your agency's writing style, qualify on ad spend, budget and category conflict, then book a discovery call and follow up on the quiet ones. Agencies can also run this on client accounts as a paid service. With Luca, auto-send is on by default on every channel, and you can switch any channel to review.

## Key takeaways

- Marketing agency lead generation is a pipeline: proof-of-work content, a DM conversation, qualification against your retainer, a booked call, and follow-up.
- In our experience, the DM stage is where agencies leak the most deals. Even dedicated reps spend just 28% of their week selling (Salesforce, 2023).
- Your buyers decide mostly without you. Gartner's 2025 survey of 646 B2B buyers found 67% prefer a rep-free buying experience.
- Agencies have a second use case coaches don't: running a setter on client accounts, priced as a retainer line item.
- Luca replies in your agency's writing style, qualifies against your offer, and books calls. Auto-send is on by default on every channel, and you can switch any channel to review.

---

Publish proof-of-work content, reply fast to every inbound DM, qualify on spend and budget, then book a discovery call and follow up. A lot of agency new business starts with a reply to a case study, or a quiet "do you take on new clients?" sent while you're in a client review.

This guide covers how to get clients for a marketing agency from the channel you already have open. Then it covers the part almost nobody writes about: running the same motion on your clients' accounts and billing for it. Leave a message sitting and people buy from whoever answered instead. Written by the team behind [Luca](https://setluca.com).

## How to get clients for a marketing agency, step by step

You get clients for a marketing agency by treating inbound DMs as a pipeline: content, DM, qualify, book, follow up. Most agencies work the first stage hard and then lose the rest to slow replies and follow-up that stops too early. The middle is where the deals go missing. Background in our [AI setter for coaches](/blog/what-is-an-ai-dm-setter) guide.

### Step 1: Publish proof-of-work content

Post the result you got a client, or a teardown of a campaign that isn't yours. A screenshot of a creative test with the losing version left in tends to pull more replies than a polished case study in our experience, because a stranger can check it.

### Step 2: Reply fast, in your agency's writing style

Answer new DMs quickly and like your brand, not like a template. Whoever messages you has usually checked you out first: **96% of prospects look into a company before they talk to a rep** ([HubSpot, 2026](https://blog.hubspot.com/sales/sales-statistics)), so the DM is late in their process, not early. Stall on a retainer worth thousands a month and it's gone.

### Step 3: Qualify against your retainer

Ask two or three questions that show you the fit before you pitch: what they want to grow, what they spend now, and whether the timing is real.

### Step 4: Invite the fits to a discovery call

Make booking easy. One invite, one link, two times to pick from. Don't bury the ask under three paragraphs of rapport.

### Step 5: Follow up the ones who ghost

In our experience, much of an agency's pipeline sits in leads who went quiet. A warm re-engagement a few days later books calls everyone else wrote off. Our [DM follow-up sequence](/blog/dm-follow-up-sequence) has the cadence.

## How to get marketing agency clients online (the 3 channels that work in 2026)

Three channels bring in most agency clients: proof-of-work content, referrals from clients and partners, and inbound DMs that follow that content. The table shows where each one lands. By the time somebody messages you, they've usually sized up three or four other shops on their own. Our guide on [how to get coaching clients](/blog/coaching-lead-generation) covers the same channels for service businesses.

| Source | What it does for an agency | Where it lands |
| --- | --- | --- |
| Teardowns and results posts | Proves you can do the work | Instagram DMs |
| LinkedIn posts and comments | Credibility with operators | LinkedIn DMs |
| Referrals from current clients | Best close rate, lowest volume | Text, WhatsApp |
| Partner directories | Category-qualified, slow trickle | A form, then email |
| Communities and Slack groups | Warm, peer-vouched | DMs on the platform |
| Your own paid campaigns | Predictable volume, colder leads | Lead form, then CRM |

Two of those carry both speed pressure and volume: Instagram DMs and WhatsApp. Automate those first. Luca drafts replies on Telegram, Instagram and Messenger today, and doesn't cover LinkedIn or email. Our [multi-channel DM playbook](/blog/multi-channel-dm-playbook) splits the coverage.

## How do you get clients for a marketing agency from DMs you have?

Fix the middle of the pipeline rather than chasing more reach. Luca reads every DM as it lands, drafts a reply the way you write, qualifies against your retainer, and moves the lead toward a booked call. You can review each draft, which matters in the first weeks, before it's read enough of your edits to be trusted on a lead worth $8k a month.

## How do you qualify a marketing agency lead for retainer fit?

Qualify on spend, budget, and category conflict before you pitch. Money early is not rude; it's the fastest way to find out whether this is a real conversation. Gong's call analysis put win rates about **10% higher when the seller raised pricing on the first call** ([Gong, 2023](https://www.gong.io/blog/sales-stats)). In the DMs, ask for a rough monthly number by the third message.

| Signal | How to ask it | Green light | Walk away |
| --- | --- | --- | --- |
| Monthly ad spend | "Roughly what are you spending a month?" | Your fee is a sane share of it | Your fee dwarfs the media |
| Retainer budget | "Any range in mind for management?" | A number, or willingness to hear one | "Depends on results" |
| Category conflict | "Who else in your space have you used?" | Adjacent to a current client | A direct competitor of one |
| Who decides | "Who else weighs in?" | They sign, or they name who does | Vague "the team" |
| Trigger and timing | "What changed that made this urgent?" | A launch, a departure, a flat quarter | "Just exploring" |

Category conflict is the row specific to agencies, and the one people skip. Sign two competing brands in one vertical and you'll spend a year defending which gets your best creative idea. Ask early, say no in the thread. More in our [lead qualification questions](/blog/lead-qualification-questions-for-coaches) guide and [the price conversation in DMs](/blog/the-price-conversation-in-dms).

## A DM script for marketing agencies: first reply to booked call (one worked thread)

Devon is an illustrative composite, not a customer record. He runs a paid-social shop in Austin. Eleven retainer clients, mostly DTC brands spending $20k to $60k a month on Meta. His inbound comes from creative teardowns.

> **Lead:** "Loved your teardown of that DTC brand's paid social. We're spending on Meta but the ROAS is flat. Do you take on new clients?"**Devon:** "Appreciate that. Flat ROAS is usually a creative or offer problem before it's a targeting one. Quick q so I point you right: roughly what's your monthly ad spend, and is this something you want fixed in the next month or two? Happy to book a 20-minute audit call if it's a fit."

The lead answers at 9:40pm: $38k a month, and their agency of record just lost the account lead. Spend clears Devon's floor, the trigger is real, and the brand competes with nobody on his roster. Devon's next reply names the trigger, gives a fee range, and offers two slots.

The second lane matters more. Devon's largest client, a skincare brand, gets sixty to eighty Instagram DMs a day about shade matching, shipping, and whether the serum works on rosacea. Some are buyers. Devon added DM management to that retainer as a separate line, ran the setter on the brand's own login, and sent a weekly export of qualified threads. That line now bills more than his creative fee, because the client sees it working daily.

## Can an agency run a setter on a client's account?

An agency can run a setter on a client's account, and it's the easiest line item to sell, because the client already knows that inbox is a mess. Work out the handoff to a person before you switch anything on. Customers expect one: **87% say a company using AI for service must still give them a way to reach a human** ([Gartner, 2026](https://www.gartner.com/en/newsroom/press-releases/2026-08-04-gartner-survey-finds-87-percent-of-customers-say-companies-using-genai-for-customer-service-must-provide-access-to-a-human-agent0)).

**Whose account it is.** The client's. Connect through their own business login, with permission in writing. Never run a client inbox on a personal token that dies when someone changes a password. Telegram, Instagram and Messenger are live today, and WhatsApp is coming soon. There's no separate multi-account tier, so pick the plan whose reply allowance covers every client inbox you connect.

**Whose data it is.** The client's, and your contract should say so in a sentence. Threads, phone numbers, and emails picked up in their DMs belong to the brand. Agree the exit terms up front: a copy of the data, access switched off, a date.

**What the client sees.** A weekly count of conversations handled, qualified leads passed, and calls booked. Clients don't read dashboards. They read a short note with numbers.

**What you shouldn't promise.** Luca is a tool you run, not a white-label product you rebrand. Bill for the service, not for reselling software under your own logo. If a client asks whose software it is, say so.

## What should an agency charge for DM management?

Pick one of four models, based on how well you can count the result. The clearer the number at the end of the month, the more of your fee can safely ride on it.

| Model | How it bills | Fits when | Risk |
| --- | --- | --- | --- |
| Flat retainer | Fixed monthly fee per account | Volume is steady | You eat a viral week |
| Percentage of ad spend | 1-3% on top of media management | You already bill on spend | Fee falls when budget is cut |
| Per qualified lead | A price per booked call | They trust your definition | Arguments about what counts |
| Base plus performance | Small retainer, bonus on calls | You can forecast the account | Needs a quarter of data |

Your direct cost is the software, plus the time you spend approving drafts. Luca runs $99 to $499 a month. That review time is what makes month one on a new account your least profitable.

## What objections will you hear?

**"We already do outbound."** Different job. Cold sequences chase people who never asked. A setter answers people who already messaged you, and in our experience that side closes better because they started it.

**"We could just build this ourselves."** Sometimes true, usually expensive. A setter is writing-style training, a review queue, follow-up scheduling, channel connections, and AI disclosure that has to keep up with Meta's rule changes. You'd own every one of those, forever. Gartner expects **over 40% of agentic AI projects to be canceled by the end of 2027**, blaming rising costs, unclear value, and thin risk controls ([Gartner, 2025](https://www.gartner.com/en/newsroom/press-releases/2025-06-25-gartner-predicts-over-40-percent-of-agentic-ai-projects-will-be-canceled-by-end-of-2027)).

**"What about our brand writing style?"** The fair objection. A setter trained on one agency's chatty teardown writing style sounds wrong on a luxury skincare account. Each account needs its own writing sample, do-not-say list, and escalation rules. Edit drafts rather than rejecting them, because edits are the training signal. Our guide on [making AI DMs match the way you write](/blog/make-ai-dms-sound-like-you) has the method.

## Where do agencies leak inbound leads?

Almost all of it comes down to speed and to nobody owning the inbox. Buyers treat a fast reply as part of what they're buying, and the businesses answering them mostly don't see it that way: Forrester's work for Google put it at **54% of consumers against 29% of businesses** ([Forrester Consulting for Google, 2020](https://developers.google.com/business-communications/business-messages/files/google-what-businesses-need-to-know-about-communicating-with-consumers.pdf)). The five habits below are that gap in practice.

1. **Treating the inbox as everyone's job.** When three people can reply, nobody does, then two reply at once. Name an owner per inbox.
2. **Answering instead of qualifying.** A prospect asks about pricing, you send a rate card, the thread dies. Answer, then ask what they spend.
3. **Running one writing style across every client account.** Tone that works on your own feed reads as flippant on a client's.
4. **Booking before checking category conflict.** You learn on the call that they compete with your biggest retainer.
5. **Stopping follow-up after touch two.** RAIN Group's research puts an initial meeting at about **8 touchpoints**, with top performers needing 5 ([RAIN Group, 2020](https://www.rainsalestraining.com/blog/how-many-touchpoints-does-it-take-to-make-a-sale)). A quiet thread at day ten is normal.

## What edge cases come up on client accounts?

**A lead in your client's DMs fits your agency better.** Common on B2B accounts. Don't poach in a client's inbox. Note it, tell them, ask whether an introduction is welcome.

**A support complaint lands in a sales inbox.** Angry messages about a late order arrive in the same stream as buying questions. Route refunds, delivery problems, and health complaints to support.

**The client's own team replies at the same time.** Two answers to one person in four minutes reads as chaos. Agree who covers which hours, and pause drafting during their staffed hours.

**Someone asks whether they're talking to a bot.** Answer honestly, every time. Meta's platform terms and California's SB 1001 both require disclosure of automated messaging. Our [Instagram DM automation rules](/blog/instagram-dm-automation-rules-2026) guide covers it.

## How do you troubleshoot the agency inbox?

Match the symptom to its cause before you change the script.

| Symptom | Likely cause | Fix |
| --- | --- | --- |
| Plenty of DMs, few booked calls | You're answering questions, not qualifying | Ask for a spend number by message three |
| Calls booked, poor fit | No spend floor in your criteria | Add a floor, disqualify in-thread |
| Drafts sound wrong on one client | Writing style trained on your account, not theirs | Re-sample from that client's past replies |
| Client complains about a reply | No escalation rule for support topics | Route complaints out of the sales queue |
| Threads die after the price | The number landed before the value did | Pair the range with the outcome |
| Great threads that stop | Follow-up ends at touch two | Extend the cadence past day ten |

## What does an AI setter cost versus hiring a setter?

An AI setter costs a fraction of a human one. A US appointment setter averages **$41,420 a year, about $3,452 a month**, before commission ([Salary.com, 2026](https://www.salary.com/research/salary/hiring/appointment-setter-salary)). Luca costs $99, $249 or $499 a month for 1,250, 4,000 or 8,500 replies, with a 7-day free trial (card required).

For the full comparison, including commission and ramp time, see [AI setter vs human setter](/blog/ai-setter-vs-human-setter) and [Luca's pricing](/pricing). A human setter's tone drifts, and the good ones leave.

## What an AI setter does and does not do for marketing agencies (inbound only)

An AI setter handles the first reply, your qualifying questions, the follow-up and the booking link. It works only with people who messaged you first. It doesn't send cold DMs, and it shouldn't quote custom fees, promise results or answer for a client's brand on refunds or safety. Those stay with you.

With Luca, auto-send is on by default on every channel, and you can switch any channel to review. Telegram, Instagram and Messenger are live today, and WhatsApp is coming soon.

## Should your agency build, buy, hire, or do it yourself?

Decide on three things: how much inbound you get, how many accounts you'd run it on, and whether DM work is something you sell. Everyone has AI in their stack now, which is not a reason to build your own.

- **Do it yourself** under about fifteen inbound DMs a week with no client inboxes. You need the raw material first.
- **Buy a setter** if inbound arrives steadily on Instagram or WhatsApp, you reply hours late, and follow-up dies early.
- **Hire a human setter** if leads arrive mainly by referral, phone, or LinkedIn, or if deals need relationship work first.
- **Build your own** only if DM operations are the product you sell. Otherwise, in our experience, upkeep soon outruns the licence fee.
- **Do both** at scale: the setter covers after-hours and first replies, and a human owns your ten named targets.

## The honest limit on how to get clients for a marketing agency

An AI setter won't out-argue a great human on your hardest pitch. A serious prospect sometimes pushes back on scope, or has a budget set up in a way that needs a custom proposal, and a closer reading the room wins that one.

How to get clients for a marketing agency from DMs: reply fast, qualify on ad spend, and book the call before you quote.

Running a setter on someone else's account raises the stakes on a bad send. It looks like this: a drafted reply quotes a delivery time the brand can't hit, or answers a product-safety question it had no business answering, and the customer screenshots it.

The ladder we'd run: anything touching refunds, health claims, legal threats, press, or a named executive gets flagged rather than queued, and waits for a person. Anyone asking to speak to a person at the company gets a name and a handoff in the next message. Any account in its first week stays manual.

If a bad reply goes out, correct it in the thread within the hour rather than deleting it, tell the client before they find it, and add the phrase to that account's do-not-say list. Deleting a message someone already screenshotted is how a small mistake becomes the story.

Start on one channel, and approve what goes out from the first week. Luca costs $99, $249 or $499 a month for 1,250, 4,000 or 8,500 replies, with a 7-day free trial (card required).


## FAQ

### How do marketing agencies get clients online?

Most agencies get clients from three places: proof-of-work content like teardowns and case studies, referrals from current clients and partners, and inbound DMs that follow that content. HubSpot found 96% of prospects research a company before talking to a rep, so the DM comes late. Reply fast and qualify on spend.

### Is Instagram or Telegram better for marketing agencies?

Instagram is better for winning new clients, because teardowns and case studies there start most agency DMs. Telegram suits client communication and private groups. Luca is live on Telegram, Instagram and Messenger today, so agencies can run client chats on whichever app each client uses.

### What should a marketing agency say in the first DM reply?

Thank them, give one useful observation, and ask about spend and timing. For example: "Flat ROAS is usually a creative or offer problem before it's a targeting one. Roughly what's your monthly ad spend, and is this something you want fixed this quarter?" Save the fee range for message three.

### Can an AI setter book calls for marketing agencies?

Yes. An AI setter can answer people who message you first, ask your qualifying questions and send your Calendly or Cal.com link when the lead fits. With Luca, auto-send is on by default on every channel, and you can switch any channel to review. Telegram, Instagram and Messenger are live.

### How much does an AI setter cost for marketing agencies?

Most AI setters cost between $25 and $500 a month. Luca is $99, $249 or $499 a month for 1,250, 4,000 or 8,500 replies, with a 7-day free trial (card required). A US appointment setter averages $41,420 a year before commission (Salary.com, 2026), so software costs far less than a hire.

### Can a marketing agency run an AI setter on a client's account?

Yes, and it's a common paid line item. You connect the setter through the client's own business login with written permission, keep the writing-style profile separate from your agency's, and route support topics to their team. Gartner's 2026 survey of 3,566 buyers found 87% expect access to a human when a company uses generative AI, so build an escalation path first.


## Sources

1. [Salesforce, State of Sales (5th ed.): reps spend 28% of their week selling, 2023, retrieved 2026-08-08](https://www.salesforce.com/news/stories/sales-research-2023/)
2. [HubSpot, Sales Statistics (96% of prospects research a company before engaging a rep), 2026: retrieved 2026-08-08](https://blog.hubspot.com/sales/sales-statistics)
3. [Gartner, Sales Survey Finds 67% of B2B Buyers Prefer a Rep-Free Experience (646 buyers, surveyed Aug-Sep 2025): retrieved 2026-08-13](https://www.gartner.com/en/newsroom/press-releases/2026-03-09-gartner-sales-survey-finds-67-percent-of-b2b-buyers-prefer-a-rep-free-experience)
4. [Gartner, 87% of Customers Say Companies Using GenAI for Customer Service Must Provide Access to a Human Agent (3,566 customers, Feb-Mar 2026): retrieved 2026-08-13](https://www.gartner.com/en/newsroom/press-releases/2026-08-04-gartner-survey-finds-87-percent-of-customers-say-companies-using-genai-for-customer-service-must-provide-access-to-a-human-agent0)
5. [Gartner, Over 40% of Agentic AI Projects Will Be Canceled by End of 2027, 2025: retrieved 2026-08-13](https://www.gartner.com/en/newsroom/press-releases/2025-06-25-gartner-predicts-over-40-percent-of-agentic-ai-projects-will-be-canceled-by-end-of-2027)
6. [Forrester Consulting for Google, What Businesses Need to Know About Communicating With Consumers (54% of consumers vs 29% of businesses on response speed), 2020: retrieved 2026-08-13](https://developers.google.com/business-communications/business-messages/files/google-what-businesses-need-to-know-about-communicating-with-consumers.pdf)
7. [Salary.com: Appointment Setter Salary, United States (average $41,420 a year, about $20 an hour; 10th to 90th percentile $34,611 to $48,323; as of October 1, 2026). Retrieved October 17, 2026.](https://www.salary.com/research/salary/hiring/appointment-setter-salary)
8. [Gong, Sales Statistics (win rates 10% higher when pricing is raised on the first call), 2023: retrieved 2026-08-08](https://www.gong.io/blog/sales-stats)
9. [RAIN Group, How Many Touchpoints Does It Take to Make a Sale? (8 touchpoints average; top performers 5), 2020: retrieved 2026-08-08](https://www.rainsalestraining.com/blog/how-many-touchpoints-does-it-take-to-make-a-sale)
10. [California Legislature, SB 1001, Bots: disclosure, retrieved 2026-08-13](https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=201720180SB1001)

---

Published by SetLuca.