Your delivery cost on the top wholesale tier is $500 a month plus the $49 platform fee. Everything above $549 is yours, with no build invoice to earn back first. Here is the input-cost math, the three ways to set retail, and where the model breaks.
Lumina8 min read
Price the outcome, then check the arithmetic. That is the whole answer, and most people selling AI systems to local businesses do it backwards.
Here is the concrete version. On Lumina Sales your delivery cost for a full custom buildout is a wholesale monthly of $500, plus the $49 you pay us for the platform. Everything you charge above $549 is yours, every month, and there is no build invoice sitting in front of it that you have to earn back first. If you sell that buildout at $1,500 a month, you are keeping $951 a month per client from month one. That is the model. The rest of this is the math behind it, the three ways people set the retail number, and where the model breaks.
What does delivering an AI buildout actually cost?
There are two honest paths, and they cost very different things.
Add it up for one client doing 400 minutes a month and the raw usage is not the problem. It is maybe $40 of infrastructure. The problem is everything around it. Somebody has to write the prompt, wire the calendar, build the booking logic, handle the website chat, set up missed-call text-back, maintain it when a model deprecates, and answer the phone when the client says the agent quoted the wrong price. That labor is the actual cost, and it does not show up on anybody's pricing page.
Buy delivery wholesale. Our tiers are $500 a month for the top tier, which includes voice, $275 for tier two, $225 for tier one, and $40 to $350 a month for any single service bought on its own. Those are the numbers you pay us, not the numbers your client sees. The top tier replaces about $745 a month of the same work bought line by line.
The reason to look at both is that you should know your floor before you quote. If you cannot say what a client costs you to serve in month seven, you are not pricing, you are guessing.
Three ways to set the retail number
Cost-plus. Take your $549 and multiply. It is fast and it is defensible in your own head. It is also the worst of the three, because it anchors your price to your own efficiency rather than to what the work is worth. Get cheaper to deliver and you accidentally get cheaper to buy.
Hourly or project. This is where most agencies drift, and it is why so many of them stall at six clients. You quote a build fee, you quote maintenance, and the client now has two decisions to make instead of one. Worse, you have taught them the value is in your hours, so every renewal is a negotiation about hours.
Outcome pricing. You price against the money the system moves. For a home service business the arithmetic is usually missed calls. Ask the owner two questions on the call: what is your average job worth, and roughly how many calls go to voicemail in a week. If the answer is a $600 ticket and eight missed calls, and even a quarter of those convert when somebody actually picks up, that is a bit over $1,200 a week they are currently leaving on the table. Against that number, $1,500 a month is not expensive. It is obviously cheap, and they can do the math themselves while you are still talking.
That last part matters more than the pricing theory. The live demo has an ROI slider seeded with their own ticket and close rate for exactly this reason. You are not asking them to accept your figure. You are watching them arrive at it.
What the margin actually looks like over twelve months
Take a single client at $1,500 retail on the top wholesale tier.
Month one: $1,500 in, $500 wholesale out, $49 platform. Net $951.
Month twelve: same, assuming they stay. Cumulative net around $11,400 from one account.
Add a second client at the same price and your platform fee does not move. It is $49 whether you run one account or ten, so the second client nets $1,000.
The shape of that is the point. A build-fee model front-loads your revenue and then flattens. A wholesale-plus-markup model starts smaller and compounds, and it does it without a moment where you owe a developer money you have not collected yet. We wrote about how the pieces fit together in what Lumina Sales actually is if you want the product side rather than the money side.
One caution on the arithmetic above. Those are illustrative numbers using our published wholesale tiers and a retail price you might set. They are not a claim about what our subscribers earn. We do not publish customer results, because we would have to make them up.
Where this pricing model breaks
Three places, and they are worth knowing before you build a business on it.
The $49 buys demo capacity, not client capacity. You get 100 live voice minutes a month, 500 chat messages, 50 website audits, one published demo at a time, and a five minute ceiling on any single call. That is roughly forty demo conversations a month. It is sized for selling, not for running a client's phone line. Client delivery runs through the wholesale tiers, and that is a separate monthly commitment that starts when you order it. If you sell a client and they churn in month two, the wholesale monthly does not politely disappear on your schedule. Plan for that.
Outcome pricing requires you to hold a price. If you cannot sit through the silence after you say $1,500, you will discount to $700, and at $700 the model still works but the business is half as good. The pricing structure does not fix a sales problem. It just removes the excuse that the build fee killed the deal.
It is the wrong model if you want to own the stack. Some agencies genuinely should assemble their own infrastructure. If you have a developer in-house, a specific technical differentiator, or you are selling into a vertical with compliance requirements that need custom handling, you want the control more than you want the margin. Buying delivery wholesale means you are not the one who decides which model runs the agent next quarter. That is a real trade, and we would rather say so here than have you find out in month four.
How to set the number on the actual call
Do not lead with price. Lead with the demo running against their business, let them talk to it, then ask what happens today when that call comes in at 6pm. Once they have told you what a missed job costs, you name one number and stop.
One number, not three tiers. Tiered proposals are where a $1,500 deal becomes a $600 deal, because you have handed the buyer a menu and the cheapest item on a menu is a suggestion. If you want a second option, make it a bigger one, not a smaller one.
And quote monthly. The whole reason the no-build-fee structure exists is that a $4,000 setup charge is a decision and $1,500 a month is a subscription. Local businesses already buy software monthly. They do not already buy capital projects.
FAQ
What is the minimum I should charge?
Your floor is $549 a month per client on the top tier, which is what delivery plus the platform costs you. Anything at or below that is volunteering. Realistically, if you cannot get to at least double your cost, the account will not survive the support load.
Can I charge a setup fee anyway?
Yes. Nothing stops you. The structure removes the requirement, not the option. Some agencies charge a small onboarding fee to filter out tire kickers, and that is a legitimate use of a number that is otherwise a deal killer.
Do I have to buy the top tier?
No. Tier two is $275 a month, tier one is $225, and individual services run $40 to $350 on their own. Match the tier to what you actually sold. The top tier is the one to lead with because it includes voice, which is the part prospects react to.
What happens if I hit my usage caps mid-demo season?
The live parts pause until the next cycle and everything else keeps working. Your prospect sees a neutral message saying the demo is not accepting calls right now. There is no metered invoice and no overage bill on any plan, which is the tradeoff for the caps existing at all. The FAQ on the main site covers the rest of the limits.
Is now a bad time to lock in pricing?
Our prices rise on October 1. If you are going to test this, the arithmetic in this post gets slightly worse after that date.
What to do next
Do not start by picking your price. Start by finding out what a missed call is worth in one trade you already sell into, because that single number is the anchor for every quote you will write this quarter. Then open the live demo, run the ROI slider with real figures from that trade, and see whether the number it produces supports the price you were planning to charge. If it does, start your workspace and build the demo for your next call. If it does not, pick a trade with a bigger ticket.
