What is an AI intake agent and what does one cost?
An AI intake agent is software that answers inbound inquiries the moment they arrive, asks the qualifying questions a human would, and books the appointment or hands the conversation to a person. In 2026 you can buy one three ways, at three very different price points. AI receptionist subscriptions run from about $25 to a few hundred dollars a month; Smith.ai, the best-known hybrid service, publishes AI plans just under $100 a month and human-backed plans around $300. Dedicated intake platforms for verticals like legal typically land between a few hundred and a couple thousand a month. A custom-built agent, trained on your own services, pricing and policies and wired into your channels and calendar, is usually a one-time project: ResonateOps builds one for USD 4,900 with an optional USD 1,200 a month operation plan. This guide explains what you get at each level and what actually moves the number.
What does an AI intake agent actually do?
An intake agent sits on the channel where prospects already reach you (web chat, WhatsApp, sometimes the phone) and does the first minutes of work a receptionist or intake coordinator would do. It answers questions about services, pricing and availability, collects the details you need to qualify the lead, and either books directly into a calendar or routes the conversation to a human.
The difference between a useful agent and a frustrating one is rarely the underlying model. It is the knowledge base behind it (does the agent know your services, prices, policies and tone, or is it improvising?) and the guardrails (does it know when to stop and hand off?).
- Instant first response on the channel the client already uses
- Qualifying questions asked consistently, every time
- Direct booking into your calendar
- Escalation to a human when the case falls outside its knowledge
How much does an AI intake agent cost in 2026?
There are three ways to buy one, and they are priced on different logic.
Subscription AI receptionists are the cheapest entry point. AI-only plans start around $25 a month at the low end and run to a few hundred at the high end. Smith.ai, the best-known hybrid provider, publishes pricing just under $100 a month for its entry AI plan and around $300 a month for plans backed by human receptionists, with per-call overage fees on top. These tools answer generically well, but they know your business only as deeply as the form you filled in during onboarding.
Vertical intake platforms (legal intake is the most developed market) bundle CRM-grade pipelines with automation and typically run from a few hundred to a couple thousand dollars a month depending on features and seats.
A custom-built intake agent is a project, not a subscription. Someone builds a knowledge base from your actual services, pricing and policies, wires the agent into your channel and calendar, tests it against real cases, and hands it over. You own the system. ResonateOps publishes its list price for exactly this build: USD 4,900 one time, with an optional USD 1,200 a month operation plan. One honest note on all the monthly figures above: they come from the vendors’ own published pricing, and vendor blogs routinely attach ROI claims to them that you should treat with suspicion.
What drives the price of a custom intake agent?
The expensive part of a custom agent is not the language model. Model usage for an intake workload is cheap. The cost is in the knowledge work and the reliability engineering around it.
- Depth of the knowledge base: how many services, price rules, policies and edge cases it has to answer correctly
- Channels: web chat only, or WhatsApp plus web plus handoff into your inbox
- Guardrails and testing: how much validation against real conversations happens before it goes live
- Integrations: calendar booking, CRM records, notifications to your team
- Languages and tone: bilingual operations cost more to test than single-language ones
SaaS receptionist or custom build: which one fits?
A subscription receptionist fits when your intake is mostly call answering with a standard script, your volume is low, and a generic answer plus message-taking is good enough. It is the right first step for many small operations, and at $25 to $300 a month the risk of trying one is low.
A custom agent fits when the answers themselves are the product: your pricing has rules, your policies have exceptions, your clients write on WhatsApp and expect real answers, and a wrong reply costs you the lead or creates liability. It also fits when you want to own the system instead of renting it, and when intake is the first step of a larger automation roadmap. The AI Intake Agent product page describes the full scope of that build, and the customer support automation page covers the adjacent use case of support tickets rather than new inquiries.
What ResonateOps charges for one
USD 4,900, fixed, agreed before the work starts. The build takes about 30 days: discovery of your FAQs and processes, a knowledge base trained on your services, pricing, policies and tone, an intake agent with guardrails and human handoff, connection to WhatsApp or web with direct booking into your calendar, and testing against real cases before production. You get a walkthrough video, training and documentation.
The optional USD 1,200 a month operation plan keeps the knowledge base current and the agent monitored, and it can be paused any month. The build carries the same 60-day operational guarantee as every ResonateOps system: if the agreed metric is not met, the work continues at no cost or you get a full refund. If intake is step one of a bigger plan, the same knowledge-base engine grows into the AI Operating System (three agents on one knowledge base) without throwing away the investment.
Related
Frequently asked questions
Is an AI intake agent a monthly subscription or a one-time cost?
Both models exist. SaaS receptionists rent you a generic agent for $25 to a few hundred a month. A custom build is a one-time project (USD 4,900 at ResonateOps) that you own, with monthly operation as an optional add-on rather than a requirement.
How long does a custom intake agent take to deploy?
About 30 days: discovery, building the knowledge base, wiring the channel and calendar, then testing against real cases before it goes to production.
What happens when the agent cannot answer a question?
A well-built agent hands off to a human. It runs with guardrails and a clear escalation path, so edge cases reach your team instead of getting an improvised wrong answer.
Is this only for law firms?
No. Legal intake is the most developed vertical, but any service business that loses clients to slow responses on inquiries and scheduling has the same problem: clinics, agencies, brokers, home services, consultancies.
Let's talk for 15 minutes about your operation.
A free call. I'll tell you straight which processes today's AI can solve and what your infrastructure needs for them to actually run. If it fits, we move forward. If not, I point you the right way, free.