We do not publish a price list, because the cost of an AI system depends on what it has to do. Here is an honest breakdown of what actually drives the number, and how we structure an engagement so you know what you are paying for before we build.
The biggest driver is how much you are automating. A single voice agent for after-hours booking is a smaller build than a suite of agents, chatbots, and back-office automations across departments. We scope by the number of distinct workflows and the decisions each one has to make, so the estimate maps to the actual surface area of work.
Every system the automation connects to adds work. Reading your calendar, writing to your CRM, and posting to Slack each require wiring, testing, and handling for when an API is slow or returns an error. Tools with clean, documented APIs go quickly, while closed or legacy systems that need middleware or custom connectors take more engineering.
Voice agents and chatbots carry running costs that scale with use. Telephony minutes through Twilio, voice platform time on Retell or VAPI, and language model tokens are all metered, so a line handling a few calls a day costs far less to run than a high-volume outbound campaign. These platform fees are usage-based and separate from the build, and we size an estimate to your expected volume during scoping.
Straightforward flows built on n8n or Make cost less than systems that need custom reasoning, multi-step agent logic, memory, or retrieval over a large knowledge base. The more edge cases, branching, and judgment a workflow has to handle correctly, the more design and testing it takes. We flag where a problem is genuinely custom versus where an off-the-shelf pattern will do.
A build is a one-time cost to design, integrate, test, and ship the system. Running it is different work: monitoring execution, reviewing transcripts and logs, refining prompts and scripts from real behavior, and upgrading models as they improve. You can take ownership after the build or keep it under an ongoing plan, and the two are priced separately so you only pay for what you use.
Handling sensitive data or working in a regulated industry adds requirements that shape both build and run. Recording disclosures and consent prompts on calls, tighter control over where data is stored and which model providers see it, and audit logging all add scope. We scope the specific rules that apply to you before building rather than assuming a one-size-fits-all setup.
For chatbots and marketing systems, the state of your source material affects cost. A clean, current set of docs indexes quickly, while scattered or outdated content needs cleanup before the system can answer from it reliably. Getting your knowledge base in order up front keeps the build tighter and the results more accurate.
Most engagements start with a fixed-scope build. We map the workflows, agree on what the system will and will not do, and quote the design, integration, testing, and launch as a defined project. You know the deliverable and the cost before we write anything, and larger systems are phased so early workflows go live while later ones are still being built.
After launch you can keep the system under an ongoing plan where we monitor execution, review real transcripts and logs, tune scripts and prompts, and upgrade to stronger models as they ship. This is priced separately from the build and is optional. If you would rather run it yourself, we hand off with documentation so your team can maintain it.
The metered costs from third-party platforms, such as Twilio telephony, the Retell or VAPI voice layer, and language model tokens, are usage-based and passed through to you. We are transparent about which platforms your system uses and roughly what drives their cost, so your running spend tracks your actual volume rather than a marked-up flat fee.
There is no single price, because cost depends on what you are automating. The main drivers are the number of workflows or agents, the systems they integrate with, the complexity and custom engineering involved, your call and message volume, and whether you want an ongoing run-and-tune plan after the build. We scope your specific use case and give you a defined quote before any work starts, so you are never guessing. Tell us what you want to automate and we will put a number to it.
A voice agent has two parts. The build is a one-time cost to design the scripts, configure the agent on Retell or VAPI, connect Twilio, and wire it to your calendar and CRM, and that scales with how many call flows and integrations it needs. Running it is usage-based, combining telephony minutes, voice platform time, and language model tokens, so a low-volume after-hours line costs far less to run than a high-volume campaign. We size both to your expected call volume during scoping and quote them separately.
Only if you want ongoing support. The build itself is a one-time, scoped project. After launch you can choose an ongoing plan where we monitor the system, refine it from real usage, and upgrade models over time, which is a recurring cost, or you can take ownership and run it yourself with no monthly fee to us. Separately, the usage-based platform costs like telephony and model tokens are billed on what the system actually uses. Ask for a quote and we will lay out exactly what is one-time and what is recurring.
Send us where the manual work hurts most and we'll come back with a system that handles it, and a date it goes live.
We'll get back to you by email. Prefer to email directly? Send here →