An agency proposes an assistant to a client. The client is polite and unconvinced.
She has used bad ones. She has been through a loop with a company that would not let her reach a person. She thinks her customers want to speak to somebody, she is worried about what it might say, and she does not have time to look after another thing.
Every one of those is a reasonable position based on her actual experience, and an argument that begins by disagreeing with all four is an argument that loses.
Script 1: I have used ones that were useless
Agree. Most are, and the difference is entirely in whether somebody wrote proper answers and set boundaries, which is the work being proposed.
Script 2: my customers want to speak to a person
For some conversations they do, and those should reach you faster than they currently do. The rest are people asking what time you open.
Script 3: it will say something wrong
A real risk, addressed by deciding what it must never answer and what must escalate, which is a decision the client makes rather than a hope.
Script 4: it will sound impersonal
It sounds like whatever is written, in the client's own words. Showing two answers, one generic and one in their voice, settles this immediately.
Script 5: I do not have time to maintain it
Which is why the service includes maintenance, and is the reason to pay somebody rather than to install something, as in reselling assistants to your own clients.
Script 6: we already have a contact form
Which collects a message and answers nothing. The comparison is not with having no form, it is with the evening enquiry that got no reply.
Script 7: we are too small for this
Volume matters, and honestly assessing it is better than selling into a business that will see nothing, which protects the agency's reputation.
Script 8: what if it replaces my staff
It handles the questions they resent answering, which is usually welcomed once framed that way rather than as a headcount conversation.
Script 9: how do I know it is working
Reporting, monthly, with a number, which is the commitment described in pricing assistant builds for clients.
Rule 1: agree with the objection first
Every one of them is based on a real experience, and dismissing it tells the client you have not had the same experience they have.
Rule 2: demonstrate rather than explain
Ten minutes with their own questions is worth an hour of description, and it is the only thing that reliably changes a sceptical client's mind.
Rule 3: offer a limited trial
One section of the site, or one category of question, which removes the risk and produces evidence for the wider conversation.
Rule 4: never argue that people prefer machines
They do not, and claiming otherwise loses the room. The argument is that people prefer an immediate answer to a delayed one, which is different and true.
Answer the objections with a demonstration
An agency that agrees with every objection, shows an answer written in the client's own voice and offers to prove it on one part of the site converts a sceptical client who has used bad ones before. The distinctions are set out in ai assistant vs chatbot vs ai employee, and the local market approach in agencies selling chatbots as a service to local clients. Start at try, see features, and the sector view sits under agencies.