AI sales agent
It runs the conversation to your script, asks the questions that actually decide something and qualifies the enquiry. The meeting lands in a manager’s calendar.
- Buildfrom $1,700
- Timeline2–4 weeks
- Supportfrom $570/mo
Who this suits: A flow of enquiries where a good part never reaches a conversation: they arrive in the evening and at weekends, and qualifying happens at the meeting itself.
What is in the first version
- conversation script
- allowed promises
- qualifying criteria
- booking the meeting
The figures above are the package this sits on: AI agent or process automation. Anything past its edges the calculator adds before the work starts, not after.
How the path goes
- 01
The answer first, the script second
A person gets an answer to their own question first, and the script follows it instead of standing in front.
- 02
Qualifying takes three or four facts
The set is yours and holds only what changes the outcome, and nothing already said is asked again.
- 03
Booking happens inside the conversation
It offers time genuinely free in the right calendar, holds the slot and passes the manager the whole conversation.
What decides whether this works or annoys people
Promises come from a list
Prices, terms and timings are quoted from a table, and what is not in it goes to a person.
Qualifying is not a questionnaire
The criteria are what to find out, not the order to ask, and a meeting may be booked with fields empty.
An offered slot is held
A slot is held from the moment it is offered rather than accepted, and the hold expires by itself.
What to measure once it is live
- share of enquiries that reach qualification
- meetings booked per week
- no-show rate on booked meetings
- conversations where a human was asked for
Why this comes out faster
The shape of this one is known: the states, the edge cases and the things that usually go wrong have been decided before. Nothing here is a template, and the saving is not in your half of the work. It goes into your process, your content and the systems this has to talk to, which is the part nobody can have solved in advance.
A likely stack for this
Picked against the task when we scope it, not decided in advance. This is the shelf it usually comes off.
- Agent and modelsPythonFastAPIOpenAIAnthropicpgvector
- Conversation, channels and dataTelegram Bot APINode.jsPostgreSQLRedis
- Scheduling and operationsTemporaln8nDockerGrafana
Price it yourself, right here
Five steps, and you can see the number without leaving a contact. The estimate accounts for the kind of work, what you already have and what it will need inside.
These bills do not come from us
- AI model calls at the provider’s rate
- Hosting and the server with the conversation log
- The CRM and calendar plans
- SMS and telephony for meeting reminders
Asked before the first call
- Will it give discounts and haggle?
- Only inside the frame you have written down: the step, the floor and the conditions. Outside that table it does not negotiate, it hands the conversation over.
- What if the customer asks for a human?
- The agent goes quiet and hands the thread over at once, with no attempt at the objection. The same happens on a complaint or an off-script subject.
- Where does it get our qualifying criteria?
- From your closed deals and what your managers already ask. Afterwards you edit the list without us: a criterion is a row in a table, not a prompt.