_FBlogo_red.png)
An AI voice agent can sound finished long before the work behind it is paid for. The integrations that let it check tickets, create records, and book appointments take real build time, and if the only line on the quote is a monthly agent fee, that time is easy to give away. This guide walks through pricing the n8n integration work behind a RingLogix Agent Studio voice agent: separating one-time work from recurring work, estimating build effort and workflow executions, choosing where n8n runs, and turning a tested workflow into a repeatable service.
Key Takeaways
- Custom integration work is a professional service. Quote it separately from the recurring agent fee so your team’s build time gets paid for.
- Scope the job before you estimate it. A clear action and finish line keeps the build from growing after the quote.
- Build effort depends on the action. Templates and catalog tools are fastest; multi-step workflows and new systems that use OAuth take the most time.
- n8n Cloud bills by monthly workflow executions. Estimate them from tool calls per phone call plus scheduled workflows, and check concurrency limits too.
- A tested workflow can be duplicated for the next customer, but each deployment still needs configuration and testing. Price that work, too.
Before You Quote: Separate the Build From the Agent
When partners asked what to charge during the session, Jaime Norris, who hosted it, started with something MSPs already do in voice. Many charge a one-time setup fee for a phone system install or a per-user fee for installs. Integration work behind an AI voice agent belongs in that category, and it usually asks more of your team. As Jaime put it, this type of service that you’re offering now, really gets into a professional service
.
Her warning was about where the money goes missing: you don’t want to just build this agent and spend all this time and then charge your customer maybe a monthly fee, or usage, or whatever. You want to charge for all this building that you’re doing, too.
The monthly agent fee pays for the ongoing service. It doesn’t pay back the hours spent building the integrations that make the agent useful, so keep the two apart on every quote. For the recurring side, see our guide to AI voice pricing for MSPs.
1. Scope the Job Before You Estimate
An estimate is only as good as the definition behind it. Before you price anything, write down each action the agent needs to take, which system it touches, and what counts as done. Our guide to scoping an AI voice agent covers how to define one job and a clear finish line.
Then check how each action will be built. If the system already has a prebuilt tool in the Agent Studio catalog, the build line may be small; see Agent Studio’s one-click PSA integrations. PSA means professional services automation, the service-desk platform many MSPs use for tickets and customer work. If an action needs several steps or a system without a prebuilt tool, plan an n8n workflow. Our guide to when an AI voice agent should use n8n walks through that choice.
2. Estimate the Build by Tier
Effort depends on the action, not the agent. In the session, Jaime grouped typical builds into tiers, from fastest to slowest:
- Starting from a tested template. Most of the work is connecting accounts and adjusting fields for the new customer.
- One API action. A single request, such as looking up an order status. An API, or application programming interface, lets one system request information or actions from another.
- Multi-step logic. Several systems, conditional branches, data cleanup, and a success and error path for each action.
- A new system that uses OAuth. OAuth is a common way for one app to authorize access to another without sharing a password. Connecting a system with OAuth for the first time adds setup and testing time to any tier above.
Where a build lands within its tier also depends on your team’s experience with the systems involved. Your own time records are the best basis for estimates, so track how long each tier actually takes and price from that history rather than from a best-case build.
Count the Testing, Not Just the Build
A workflow that works when everything goes right still needs a clear response when a lookup fails or a system is offline, and the agent needs testing on both paths before callers reach it. That testing is part of the build. Our walkthrough on how to connect and test n8n tools in RingLogix Agent Studio covers the checks to run before you deploy.
3. Estimate Workflow Executions
If your workflows run on n8n Cloud, your plan is priced by monthly workflow executions. According to n8n’s pricing page, an execution is one complete run of a workflow, no matter how many steps it contains. For a webhook-triggered workflow, each trigger event is one execution, so every tool call your agent makes to n8n counts.
That makes the estimate straightforward:
Monthly executions ≈ (calls per month × n8n tool calls per call) + scheduled workflow runs
The session used two reference points. About 1,000 calls a month with roughly two n8n tool calls each comes to about 2,000 executions. A polling workflow that checks something every five minutes runs about 8,640 times in a 30-day month, whether or not anyone calls; n8n’s own estimate for a five-minute schedule is about 8,600 to 8,900 executions a month.
The following is a hypothetical example, not a customer result:
| Item | Executions per month |
|---|---|
| 1,500 calls × 2 n8n tool calls each | 3,000 |
| One workflow polling every five minutes | 8,640 |
| Estimated total | 11,640 |
In this example, the scheduled workflow uses almost three times as many executions as the calls themselves. That’s why scheduled workflows belong in every estimate.
Watch Concurrency, Not Just Volume
n8n Cloud plans also cap how many executions can run at the same time, from 5 on the Starter plan to as many as 50 on Pro. If several callers trigger tools at once, that cap matters as much as the monthly total. Check current limits before you commit to a price.
4. Decide Where n8n Runs
Where n8n runs changes what you pay for and what your team maintains:
| n8n Cloud | Self-hosted | |
|---|---|---|
| Setup | Quick to start, with no infrastructure to manage | Runs on infrastructure you control, such as a VPS (virtual private server), Docker, or Kubernetes |
| What you pay for | A subscription priced by monthly workflow executions | Hosting, plus your team’s time |
| Capacity | Plan limits on monthly and concurrent executions | Set by the infrastructure you provide |
| Data location | Stored in the EU, on servers in Frankfurt, Germany | Wherever you host it |
| Maintenance | Handled by n8n | Your team handles updates, backups, and security |
Self-hosting the Community Edition has no license fee. When Jaime described it as free in the session, she added a caveat: Besides the storage you need to host it.
The server, the setup, and the time to keep it patched and backed up are real costs. Weigh them against your expected call volume, how many calls will hit tools at once, where customer data needs to live, and who will own the server.
If you plan to serve several customers from one self-hosted instance, review n8n’s Sustainable Use License first. It allows consulting work such as building workflows for clients, but not making your instance available for customers to connect their own accounts and build their own workflows.
5. Structure the Quote
Separate the work by how it gets paid for:
| Line item | What it covers | How to bill it |
|---|---|---|
| Integration build | Scoping, workflow build, testing on success and failure paths, and launch | One-time professional services fee |
| AI voice agent | The agent service your customer uses every month | Recurring, per your agent pricing |
| Workflow hosting and upkeep | Hosting, monitoring, fixes for API changes, and credential updates | Recurring fee |
| Executions | Workflow runs on an execution-based n8n Cloud plan | Usage-based, or built into the recurring fee |
The upkeep line matters because integrations change after launch. A connected system updates its API, a token or password expires, or the customer adds a step to its process. Pricing that work up front keeps you from absorbing it later.
6. Price Reuse Without Giving It Away
The professional services fee covers the first build. Reuse is how integration work becomes a repeatable service. Jaime called this out as one of n8n’s practical advantages: you can duplicate tested flows and point it at new credentials
.
This matters most if you serve a vertical, where several customers need the same kind of workflow. A tested ticket lookup or appointment flow built for one customer becomes the starting point for the next.
Reuse still takes work. Each new deployment needs the customer’s own credentials, fields, and routing, and it needs its own testing before it goes live. Price later deployments to cover that configuration and testing time, even when they take far less effort than the first build.
Where to Look When the Numbers Don’t Work
Start with the symptom and work back to the line item that missed it:
| What you see | What to check first |
|---|---|
| Margins shrink months after launch | Whether the recurring fee covers upkeep: API changes, expired credentials, and process changes. |
| The n8n bill is higher than expected | Scheduled workflows that weren’t in the execution estimate. |
| Callers wait during busy periods | The plan’s concurrent execution limit and the slowest step in the workflow. |
| The second customer takes as long as the first | Whether the workflow was built to duplicate, with customer-specific values kept separate. |
| The build runs past the estimate | Whether each action and its finish line were defined before you quoted. |
Charge for the Work That Makes the Agent Useful
An AI voice agent earns its place by completing work in your customers’ systems, and the integration work behind it deserves its own line on the quote. Scope the job, estimate the build and the executions, decide where n8n runs, and price reuse so later deployments are paid for, too.
Ready to price your next integration? Watch the hosting and pricing discussion on YouTube, then run your next quote through the steps above. If you’re evaluating the platform, bring an integration to an Agent Studio demo so the conversation starts with the work you want to price.
Frequently Asked Questions
Should MSPs charge a setup fee for AI voice agent integrations?
Yes, when the agent needs custom integration work. Building, testing, and launching integrations is professional services work. Quote it as a one-time fee, separate from the recurring agent fee, so your team’s build time is paid for.
How do I estimate n8n executions for an AI voice agent?
Multiply your expected monthly calls by the average number of n8n tool calls per call, then add the runs from any scheduled workflows. For example, 1,000 calls with about two tool calls each comes to roughly 2,000 executions a month before scheduled workflows.
How many executions does a workflow that runs every five minutes use?
About 8,640 in a 30-day month. n8n estimates 8,600 to 8,900 executions a month for a five-minute schedule, and those runs count whether or not anyone calls.
Is self-hosting n8n cheaper than n8n Cloud?
It can be, but it shifts the cost rather than removing it. The self-hosted Community Edition has no license fee, but you pay for hosting and for your team’s time on setup, updates, backups, and security. n8n Cloud is a subscription priced by monthly workflow executions, with limits on monthly and concurrent executions.
Can I reuse the same n8n workflow for multiple customers?
Yes. You can duplicate a tested workflow and point it at a new customer’s credentials. Each deployment still needs its own configuration and testing, and you should review n8n’s license terms before giving customers access to your n8n instance.
What should a recurring integration fee cover?
It should cover hosting for the workflows, monitoring, fixes when a connected system changes its API, and credential updates when tokens or passwords expire. Pricing these up front keeps ongoing support from turning into unpaid work.
