Call operations
Set a rule for callers who ask if they are speaking to a bot
A clear disclosure rule for callers who ask whether they reached a bot, with a plain answer, a human transfer path, and a monthly review.

Table of contents
A caller who asks whether they reached a bot is usually deciding whether to keep talking. The question is not an insult and it is not a technical quiz. It is a request for honesty about who is on the line and what will happen to the information being shared. A small team needs one plain answer, one transfer path, and one record of how often the question appears.
Key takeaways
- Answer the question directly in the first sentence, without hedging.
- Say what the caller is speaking with and who else may hear the call.
- Offer a human transfer before the caller has to demand one.
- Never claim to be a person when the caller asks a direct question.
- Count the question each month and review the script with the team.
Why the question deserves a direct answer
Callers ask about bots for practical reasons. They may be about to share a payment card, a medical detail, or a complaint they want a person to hear. A vague answer makes them repeat the question or hang up. The first sentence should settle the issue so the rest of the call can continue.
A direct answer also protects the team. An agent who avoids the question may sound evasive even when the service is fully human. A written rule removes that pressure and gives every agent the same words. The rule should apply to live agents, recorded greetings, and any automated menu on the line.
Write the disclosure in one plain sentence
Use a sentence that names what the caller reached and what happens next. "You are speaking with a live assistant who can take your details and pass them to the office" is clear and short. If the line uses an automated greeting before a person answers, say that as well. The caller should not have to infer the answer from background noise.
Put the sentence in the call script beside the greeting. Keep it free of jargon such as virtual agent, conversational AI, or language model. A caller who hears a familiar word is more likely to trust the answer. The agent should say the line in a normal tone rather than reciting it like a warning.
Give the agent a script for the follow-up question
Most callers ask a second question after the first answer. They want to know whether the call is recorded, who reviews the notes, or whether a person will call back. Prepare one short response for each of those questions. The agent should answer what is known and offer to find out what is not. Guessing about recording or review is worse than a brief pause.
A useful follow-up line is: "I can note that you want a person to review this, and I will tell you when to expect the call." That line names an action and a time without overpromising. It also gives the caller a reason to stay on the line. The agent should write the promise in the same record the next shift will read. The caller should hear the same line from any agent who takes the case.
Decide what the agent may and may not claim
The rule should list claims that are never allowed. An agent must not say "I am a real person" when the caller asks directly and the agent is not. An agent must not deny that a call is recorded when a recording notice has played. An agent must not promise that no software touches the notes if the team uses transcription or summarization tools.
The rule should also list claims that are safe. An agent may say that a named person will review the request. An agent may say that the caller can ask for a supervisor. An agent may say that the team follows a written privacy policy and point to where it is published. These statements are checkable and do not depend on guesswork.
Handle the caller who wants a person immediately
Some callers will not continue with anyone who is not the person they expected. The agent should offer a transfer or a callback without arguing. Ask for the best number and a time window, then repeat both back. If no person is available, say when one will be and what will happen in the meantime.
Do not route every bot question to a supervisor. That turns a simple disclosure into a queue problem and slows the calls that truly need escalation. Route to a person when the caller asks, when the request is sensitive, or when the caller cannot get a clear answer. The transfer rule should name who receives the call and what the receiving agent must say first.
Record the disclosure and review it monthly
Add a short field to the call record for the bot question. The field can be as simple as yes, no, or not asked. Add a second field for whether the caller requested a human. Those two counts show whether the script is working and where callers lose confidence.
Review the counts with the team at the end of each month. Read a few calls where the caller asked and the call ended early. Ask the agents which wording felt natural and which felt defensive. Change one line at a time so the effect is visible. The callback promise guide shows how to keep the follow-up promise consistent with the disclosure.
The goal is not to win an argument about technology. The goal is to let the caller decide what to do with a truthful answer. A short rule, a plain sentence, and a real transfer path are enough for most small teams. Write the rule down, test it on live calls, and keep the record.