Picking a platform is the easy decision. The question that actually stalls agencies is what to charge for it, because white-label AI services don’t behave like a flat-fee tool. Voice minutes, chat volume, and message counts move every month, which means the cost underneath your price moves too. Get the pricing model wrong, and you either eat the variance yourself or hand your client a bill that swings unpredictably, and either one costs you the account eventually.
This picks up where choosing a white-labepl marketing automation platform leaves off. Once you’ve picked Chatley, GoHighLevel, or a combination, this is how to turn that cost into a price your agency actually profits from. If you’re building a recurring AI service offering for clients, a Chatley channel partner program can provide the infrastructure for delivering white-label AI services under your agency brand.
The three pricing models and where each one breaks down
Cost-plus markup. Take what the platform costs you and add a fixed percentage. Simple to calculate, and it guarantees your costs are covered. It breaks down with usage-based platforms; for example, your underlying voice-minute cost varies month to month, a fixed markup percentage means your invoice to the client varies too, and clients dislike a retainer that isn’t actually a flat number.
Flat monthly retainer. The client pays one number every month regardless of how much they use the service. This is what most clients actually want, since it’s predictable and easy to budget against. The risk sits entirely with you, if a client’s call volume spikes past what you priced in, your margin on that account shrinks or disappears. This only works if you’ve built in a usage ceiling or a buffer, not just an average.
Per-lead or per-outcome pricing. The client pays per qualified lead, per booked appointment, or per call handled. This aligns price with value better than two of the above, and it’s an easier sell to a client who’s skeptical of paying for software they might not fully use. It’s harder for you to forecast revenue, and it requires a clean way to track and prove the outcome count, which not every reporting setup handles well.
Most agencies that resell usage-based AI tools end up blending the second and third.A flat retainer that includes a set volume of calls or minutes, with overage billed separately. That structure gives the client the predictability they want and protects your margin from the usage swings that break a pure flat rate.
What markup actually looks like in practice
There’s no single correct number, but the ranges published across the agency-services industry cluster in a consistent place. A common baseline is a 50 to 100% markup on the underlying service cost, meaning if a service costs your agency $1,000 a month, you’d price it between $1,500 and $2,000 for the client. Agencies adding real strategic value on top, custom reporting, ongoing account management, integration into a client’s existing systems, often justify pushing that to 100 to 200% or more, because at that point the client isn’t paying for the software, they’re paying for the outcome and the oversight.
White-label SEO reselling, a mature and heavily benchmarked category, tends to run at the higher end of that range, commonly 100 to 200% markup, which works out to roughly 50 to 65% gross margin once the math is done. That gap between markup and margin trips people up constantly: a 100% markup means the client’s price is double your cost, and that’s a 50% margin, not 100%. A 200% markup means the client pays three times your cost, for a margin around 67%. If a competitor says they run “60% margins,” they’re describing something close to a 150% markup, not a 60% markup.
As a rough anchor across service categories: 20 to 40% margin is a volume play where you’re competing mostly on price, 40 to 60% is the healthy middle where most bundled agency services land, and 60 to 80%-plus shows up when an agency owns a genuine niche or a deep client relationship that makes the service harder to price-shop.
Why usage-based AI pricing complicates the flat retainer
This is the part that’s specific to reselling AI voice and chat rather than a flat-fee SaaS tool. A platform like GoHighLevel or DashClicks charges you the same base fee whether a client’s sub-account is quiet or busy. Chatley, like most AI voice and messaging platforms, charges based on minutes and message volume, which means your cost to serve a client isn’t fixed even if your price to them is.
Two things fix this without turning your invoice into a variable spreadsheet the client has to interpret:
Bundle the usage cost into the retainer instead of itemizing it. Presenting the platform’s usage cost as a visible line item invites the client to price-shop that specific number, and it reframes the entire relationship as “you’re marking up software” instead of “you’re solving a problem.” Fold it into a single retainer number the same way you’d fold in your own labor.
Price to a volume band, not an average. Estimate the call and message volume a client is likely to generate, price for the higher end of a realistic range, and set a stated ceiling above which overage applies. This protects your margin from the one client who calls in triple the expected volume without needing to renegotiate every account individually.
A worked example (numbers are illustrative; replace with your actual costs)
Say Chatley’s cost to your agency for a given client runs $180 a month once setup is amortized and usage is at expected volume. At a 100% markup, retail is $360 a month, a 50% margin. At 150%, retail is $450, closer to a 60% margin. If that client’s actual call volume runs higher than expected some months, that’s the gap the volume-band pricing above is meant to absorb, not something to renegotiate every time it happens.
Bundled into a broader package, the same math changes shape: if Chatley’s voice AI sits alongside GoHighLevel or another platform in a $1,500 monthly retainer, the Chatley component might represent $300 to $400 of that total rather than standing alone, and the client sees one number for one outcome, not a menu of software fees.
Conclusion
Bundle it when the client is already buying a broader package (SEO, ads, full marketing automation) and missed-call recovery or lead follow-up is a genuine gap in what you’re already delivering. This is the stronger sell for most agencies, since it turns “we also do AI voice” into part of a complete solution rather than an upsell the client has to evaluate on its own.
Sell it standalone when a client only has one problem, for example, a service business losing leads to missed calls or slow follow-up, and doesn’t need or want a broader marketing engagement. A branded dashboard, clear setup, and a flat monthly number work well here, closer to how you’d price access to a single tool rather than a full retainer. If you’re still evaluating the reseller model itself, see our guide to white-label AI reselling for more on how agencies can package and sell AI services to clients.
Avoid presenting Chatley by name as a cost input on the invoice. The client is buying missed-call recovery or faster lead response from your agency. What’s running underneath it is your business, not theirs to see itemized.