The White-Label Agency Blueprint

How to add recurring revenue reselling AI sales agents under your own brand

My name is Sohaib and I built DM Champ.

DM Champ is a white-label AI sales agent platform. Your logo, your domain, your pricing, your Stripe account, so your clients log into what looks like your own software and never see our name anywhere.

The agent runs real sales conversations on WhatsApp, Instagram, Messenger and website chat. It qualifies leads, answers questions from the client's own knowledge base, handles objections, and books people into a calendar.

DM Champ is on an AppSumo LIFETIME deal until Thursday, August 14.

Pay once from $59, keep it forever. White-label is included on the higher tiers.

https://dmchamp.com/ltd

What's inside

1. Sell the outcome, rent the machine
2. Pick a niche you already have access to
3. The free trial play, step by step
4. What to charge, and the example math behind it
5. Onboarding a client in one afternoon
6. The two objections that kill these deals
7. When and how to raise your prices
8. The unglamorous operations that keep the retainer alive
9. The part you cannot do manually

1. Sell the outcome, rent the machine

Most agencies that touch AI sell it the same wrong way. They sell a setup. "I'll build you a chatbot for $500." One invoice, one handoff, one client who never thinks about you again until something breaks.

That model has three problems.

The first is obvious. It does not compound. Every month starts at zero and you go find another $500.

The second is that "chatbot" is a category the buyer already has an opinion about, and the opinion is bad. They have used one. It gave them four buttons and then told them to email support. When you say chatbot, the number in their head is small, because the thing in their head is small.

The third is that you have priced yourself as a builder, and builders get compared on build cost. There is always someone cheaper.

The fix is to stop selling the thing and start selling what the thing produces.

Nobody wakes up wanting a chatbot. What a business owner actually wants, in their own words, is something like this:

"Every person who messages us gets answered within a minute, day or night, including weekends, and the serious ones end up in my calendar instead of going quiet."

That is an outcome. It is worth money every single month, because the leads keep coming every single month. And it is a fair thing to charge a retainer for, because you are on the hook for it continuously, not once.

So the pitch stops being "I'll build you a chatbot" and becomes closer to:

"You get roughly forty inquiries a week on Instagram and WhatsApp. Right now the ones that come in after 6pm sit until the next morning, and some of them are gone by then. I'll put a system on your inbox that answers every one of them in under a minute in your tone of voice, using your real prices and policies, and books the ones who are ready into your calendar. I run it and maintain it. It's X per month."

Same underlying software. Completely different conversation, different objection set, different price, and it renews.

There is a second-order effect too. When you sell a build, your mental category is "expense", and expenses get cut in a bad quarter. When you sell "we answer every lead in 60 seconds and book them", you sit next to their revenue. Things next to revenue get cut last.

One clarification, because agencies get this wrong and then get burned. Selling the outcome does not mean promising a revenue number. Do not say "this will double your sales". Promise the mechanism, which you actually control: speed of response, coverage of hours, consistency of answers, appointments landing in the calendar. Let the revenue be their conclusion, not your guarantee.

2. Pick a niche you already have access to

Do not start by choosing the "best" niche. Start by listing every business that already answers your calls.

The biggest advantage in this business model is not the technology. Everyone can get the technology. The advantage is a warm door and an existing reason to be trusted.

Here is a real example from our community, because it illustrates this better than anything I could invent.

Ammar Shariq bought DM Champ with the white-label package. He did not go build a funnel, or run ads, or cold email a list of businesses he had never spoken to. He looked at his existing client base. He was already doing marketing for universities in the UAE. Those universities already knew him, already paid him, already picked up the phone.

Then the second smart thing. He did not walk in and say "I've got a chatbot for you". He positioned it as a prospecting service. Universities understand prospecting. They have admissions targets, inquiry volume, and a painful problem where prospective students message at odd hours and nobody replies until Monday. "Prospecting service" lands in a budget line that already exists. "Chatbot" lands in one that does not.

He gave each of them a free two-week trial. Fast to set up, so it cost him time he could spare, and low risk for them because there was nothing to sign. Their teams used it and got used to it. By the end, taking it away would have felt like a step backwards.

He spent roughly $200 to $300 on credits running those trials.

Three of those universities signed six-month retainers at roughly AED 1,000 per month each, which is around USD 270.

His own advice, in his words: package it well and get people to actually try it.

What is transferable here is not "sell to universities in Dubai".

He sold into an existing relationship. Zero cold outreach. The hardest part of agency sales, getting a qualified stranger to believe you, was already done.

He translated the product into their language. Not chatbot. Prospecting. The word you use determines which budget the buyer reaches for.

He picked a vertical with real inbound volume. This model only works where messages already arrive. A business with four inquiries a month has nothing to automate. Universities, clinics, gyms, dealerships, real estate, all of them drown in DMs.

He made trying it nearly free for them and cheap for himself. A couple hundred dollars of credits against three retainers is a return you would take every day of the week.

So before you do anything else, write the list. Past clients. Current clients. People you did one project for two years ago. Your accountant. Your dentist. The gym you go to. Your brother in law's company. Then cross off everyone who does not get regular inbound messages, because those are not candidates no matter how much they like you.

What is left is your launch list. It is usually somewhere between five and twenty names, and it is worth more than any lead list you could buy.

3. The free trial play, step by step

The trial is the whole sales process. You will not out-argue anyone into believing an AI can hold a sales conversation. You let them watch it happen in their own inbox with their own leads, and then the sale becomes easy.

Step one. The ask.

Keep it small. You are not asking them to buy, evaluate, or decide anything. You are asking for two weeks and access to one channel.

Something like: "I've built something for a couple of my clients that answers inbound DMs in under a minute and books the good ones. I want to run it on your Instagram for two weeks, free, so you can see it on your own leads. If you hate it we turn it off and nothing changes. I just need access and half an hour of your time."

It converts because every part of it is cheap for them. No contract, no money, no risk, one short meeting.

Step two. The half hour.

The only meeting you need before going live. Ask exactly these things and write the answers down:

• What are the five questions people ask you most? What do you answer?
• What are your actual prices, and what can you say publicly about them?
• What should this thing never say or promise? Get specific.
• When someone is genuinely ready to buy, what should happen? Booking, quote, human takeover?
• Who takes over when a human is needed, and how do they want to be notified?
• What is your tone? Formal, warm, casual? Give me two of your own replies as examples.

That half hour is your entire configuration. It is also a genuinely useful consulting session, and clients often notice they have never written this down before.

Step three. Build it that afternoon.

Setup is section 5. It is genuinely one afternoon.

Step four. The first 48 hours are yours.

Do not go live and disappear. Read every conversation for the first two days. You will find three or four things wrong: a price quoted slightly off, a tone that is too formal, a case where it should have escalated and did not. Fix all of them on day one.

This is the highest leverage work in the whole model. The agent that exists 48 hours after launch is dramatically better than the one you launched, and the client is watching during exactly that window.

Step five. Let the habit form.

Weeks one and two are not about impressing anyone. They are about the client's team quietly reorganizing around the thing. The receptionist stops opening Instagram every twenty minutes. The owner stops answering DMs at 11pm. Appointments start appearing without anyone doing anything.

This is what Ammar meant about removal feeling like a step backwards. Nobody agrees to take staff time back.

Step six. The conversion conversation.

Around day twelve, before the trial ends, not after. Do not open with your price. Open with their numbers.

"In the last two weeks it handled 214 conversations. 61 of those came in outside your working hours. It booked 19 appointments. Average response time was 40 seconds, and before this it was about six hours. Do you want to keep it running?"

Then, only then, the price. Then stop talking.

Usually what comes back is not "no". It is a number question, or a "can it also do X" question. Both are buying signals.

Step seven. Ask for a term, not a month.

Ammar's clients signed six-month retainers, not rolling monthly. Ask for the term. Six months is easy to justify honestly: "the agent gets better as it accumulates conversations, and six months means I can keep improving it rather than re-selling it to you every thirty days." Offer one month free on a six month commitment if you need a lever. Predictable revenue is what you are actually building here.

4. What to charge, and the example math behind it

Everything below is illustrative example math, not results. It shows how the model works arithmetically, not what you will earn. What you can actually charge depends on your market, your niche and your reputation.

Tier one, local single-location businesses. Salons, gyms, dentists, small clinics, restaurants, driving schools, trades. One channel, straightforward knowledge base, appointment booking. As an example, $150 to $350 per month.

Tier two, multi-location or higher ticket. Clinics selling treatments in the thousands, dealerships, real estate, private schools, B2B services. More channels, more complex qualification, more rules. As an example, $400 to $900 per month.

Tier three, real organizations. Multiple departments, multiple inboxes, reporting requirements, a stakeholder who needs monthly numbers. Example range, $1,000 and up, priced on their scale rather than your effort.

A setup fee is optional and useful. As an example, $300 to $500 once, framed as onboarding and knowledge base build. It filters out tire kickers and covers your first month of work. If you are running the free trial play, skip it, because the trial is already doing that filtering.

Now the example math, clearly labelled as an example.

Ten clients at $300 per month is $3,000 per month recurring. Twenty clients at $300 per month is $6,000 per month. Five tier two clients at $600 per month is also $3,000 per month at a fifth of the client count, which is why moving upmarket usually beats adding volume.

Against that sits your platform cost and your usage credits, which scale with conversation volume, plus your own time. On a lifetime deal the platform cost is a one time number rather than a monthly one, which changes the shape of this considerably.

Three things worth knowing before you set a number.

Anchor against their alternative, not your cost. Their alternative is a part time person answering DMs, or leads going unanswered. Both are expensive. Price next to that comparison, not next to what the software costs you.

Never price per conversation to a small business. They cannot forecast it, so they will not sign it. Flat monthly, generous fair-use ceiling, clearly stated overage. Predictability is a feature you are selling.

Charge more than feels comfortable. Every agency I have watched do this underpriced their first three clients and spent a year regretting it. Section 7 is how to fix that.

5. Onboarding a client in one afternoon

This is the operational core. If onboarding takes two weeks, the model does not work, because your margin gets eaten by your own hours. Done properly it is an afternoon.

Start with the website URL. Point the platform at your client's site and it reads their products, services, pricing and selling points and drafts the agent from that. It does most of the work, and more importantly it means you are not staring at an empty prompt box.

Then build the knowledge base. This is where quality actually comes from. Price list, FAQ, refund and cancellation policies, opening hours, locations, service descriptions. Anything they email to customers regularly belongs in here. If they have a document they send every new customer, that document is gold.

Then the six configuration fields. These turn a generic assistant into their salesperson.

Personality. How it talks. Warm and casual, or precise and professional. Give it two or three of the client's own real replies as examples and it will match them better than any adjective you could write.

Goal. The one thing a conversation is supposed to end in. Book a consultation. Get a phone number. Move them to a human. One goal, not five. Agents with five goals drift.

Company info. Who they are, what they sell, who they serve, what makes them different, what they will not do. This is the context every answer sits on top of.

Rules. The hard boundaries. Never quote a price outside the published list. Never promise a delivery date. Never give medical or legal advice. Always mention the consultation is free. Never discuss competitors. Write these as short flat statements, one per line. This is the field that answers the "what if it says something wrong" objection, so do not rush it.

Conversation flow. The shape of a good conversation. Greet, understand what they need, answer using the knowledge base, qualify with two or three specific questions, offer the next step. Do not over-script it. A rigid script produces exactly the robot feeling you are trying to avoid.

Escalation triggers. When to stop and get a human. A complaint. A refund request. Legal or medical questions. Anyone who asks to speak to a person, immediately and without argument. Deals above a value threshold the client cares about. Anything the agent cannot answer confidently twice in a row.

Then test for an hour. Message it yourself as five different customers. A price shopper. An angry one. Someone asking about something the business does not offer. Someone ready to buy right now. Someone testing whether it is a bot.

Then connect the channels and go live.

Afternoon. And the second client is faster than the first, because you now have a template. By client five you are mostly copying and adjusting, which is precisely where the margin in this business lives.

6. The two objections that kill these deals

There are only two real ones. Everything else is a price negotiation wearing a costume.

"What if the AI says something wrong?"

A completely legitimate concern, and you should treat it as one. Do not wave it away with "the AI is really good now". That answer loses the deal, because they have already decided you are overselling.

Answer it structurally instead. Three layers, and you walk them through all three.

Layer one, the knowledge base. It answers from their documents, their prices, their policies. It is not inventing a company from thin air. If something is not in there, it does not know it, and that is the point.

Layer two, the rules. Explicit boundaries you configure. Never quote outside the price list. Never promise a timeline. Never give advice in regulated areas. Show them the actual rules field during onboarding and let them dictate the lines. Clients who write their own rules stop being afraid of the system, because they built the fence.

Layer three, escalation. When it hits something outside its lane it hands off to a human. Complaints, refunds, anyone who asks for a person, anything expensive.

Then close it honestly: "It will occasionally phrase something in a way you would have phrased differently. So does a new hire in week one. The difference is that when you tell me, I fix it in five minutes and it stays fixed across every conversation from then on."

And then point at the trial. "That's exactly why we run two weeks free. You'll read every conversation. If it says something you don't like, you'll see it, and I'll fix it."

"Why not just hire a VA for that?"

The trap is arguing that the AI is better than a person. It is not, at everything, and claiming so costs you credibility. Compare on the four dimensions where the comparison is real.

Coverage. A VA works a shift. Inbound messages do not. A meaningful chunk of inquiries arrive evenings, weekends and holidays, and those are often the most motivated buyers, the ones sitting on the sofa researching. In the conversion conversation this is usually the number that lands hardest, because they have never counted it before.

Speed. A VA juggling other work replies in twenty minutes, or two hours. The agent replies in under a minute, whether it is one message or forty at once.

Consistency. A VA has good days. The agent quotes the same price and applies the same policy on message one and message five hundred.

Scale. Volume triples in a promotion week. The VA does not triple. The agent does not notice.

Then concede the true part, because conceding it makes everything else credible: "For complex negotiation, an upset customer, or a genuinely unusual situation, a person is better. That's why it escalates. This isn't replacing your team. It's making sure nothing sits unanswered for six hours before your team gets to it."

Some clients already have a VA. Not a lost deal. The agent takes the first response and the repetitive 80 percent, and the VA spends their day on conversations actually worth a human. You are making their existing hire more valuable, not firing them.

7. When and how to raise your prices

You will underprice at the start. Everyone does. It is the cost of getting your first proof, and it only becomes a mistake if you leave it there.

Raise on new clients first, always. Existing clients stay where they are, the next one pays more. Do this after every second or third signing until you feel real resistance. Roughly one in four prospects saying no is usually about right. If everyone says yes immediately, you are leaving money on the table.

Raise existing clients at renewal, with evidence. Six month terms make this natural. Come with numbers. "Over six months it handled 1,400 conversations and booked 140 appointments. I added Instagram and rebuilt the knowledge base twice. Renewal is at X." A raise attached to demonstrated work is an easy conversation. A raise attached to nothing is not.

Raise when scope actually grows. New channel, second location, CRM integration, more complex qualification. Every one is a legitimate price event, and the easiest kind to make. Notice scope creep and price it instead of absorbing it.

Stop competing on price as fast as you can. When someone says a competitor is cheaper, do not match it. "They probably are. What I do is build it on your actual prices and policies, watch it for the first two weeks, and keep improving it every month. If you want the cheapest possible version, honestly, take theirs."

Some will take theirs. Let them. The ones who stay will not leave over $100.

8. The unglamorous operations that keep the retainer alive

The retainer survives on the client feeling you are still there. That is less work than you think, but it cannot be zero.

• A monthly number. Conversations handled, appointments booked, after hours coverage, average response time. Five minutes to write, and the strongest anti-churn tool you have, because it makes an invisible service visible.
• A monthly improvement. Find what it fumbled, fix it, mention it in the same message. "People kept asking about parking, so I added that." Tiny, and enormously effective.
• A quarterly conversation. Fifteen minutes on what changed, what new services, what new objections the sales team hears. Update the knowledge base. This is where expansion revenue comes from, roughly every time.
• Never let the knowledge base go stale. A confidently quoted price from eight months ago is the fastest way to lose a client's trust.

9. The part you cannot do manually

Everything above you can run manually with any tool. The positioning, the niche selection, the trial play, the pricing, the objections, the onboarding checklist. None of that is software. It is the business, and it is yours to keep regardless of what you build it on.

What makes it a business rather than a series of favours is the white-label. Your brand on the dashboard. Your Stripe on the billing. Your price on the invoice. Your domain in the address bar. Your client logs in, sees your logo, gets billed by you, and has no idea we exist.

That is the difference between reselling and referring. It is also the difference between owning the client relationship and renting it from a vendor who could raise prices or go direct at any moment.

That is what DM Champ gives you out of the box.

The deal

DM Champ is on an AppSumo lifetime deal until Thursday, August 14. From $59, once. Not monthly. Not annual. Once, and you keep it.

White-label is included on the higher tiers, which is the tier you want if anything in this document is your plan.

For context on whether it holds up: 4.88 out of 5 across 138 AppSumo reviews. And the origin story is real. The first version of this agent collected a $1,000 deposit on a $50,000 deal with no human involved, for a hair transplant clinic called FueGenix that went from $5M to $10M in a year.

Get it here: https://dmchamp.com/ltd

Help docs and setup guides: https://help.dmchamp.com/

If you build something with this, I would genuinely like to hear about it.

Sohaib