In short: Sales Farmer auto-replies only handle short and clear agreements, using sentences prepared in advance. They are turned off by default. Keep manual approval until you know your typical replies, and remember that LinkedIn limits accounts that are very active: sending messages automatically can lead to restrictions.
This page describes a feature, not compliance advice. LinkedIn can change its rules at any time.
What exactly do auto-replies do?
They read each reply received after you turn them on, decide whether the case is simple, and then send a short message in your place. A simple case fits in one sentence: the person is interested, or they ask for information, a presentation or a meeting.
The message is not written freely by a model. It is a sentence prepared in advance, with the person's first name and your links. For a meeting request, the AI offers your booking link without inventing a time. For a request for a presentation, it sends the presentation link you saved.
This design has a useful consequence: the text is predictable. You can read it before you turn the feature on, and you find every message sent in the "What the AI did" log. The limit: a prepared sentence sounds less personal than a message written by hand. For a key account, reply yourself.
In which cases does the AI stop?
The AI stops as soon as a message calls for judgment. The list is closed, and it is long:
- a refusal, an unsubscribe, an angry person or an automatic absence;
- an objection, for example about price, or an unclear reply;
- a question about a price, a deadline, a client reference or a commitment;
- a request for a call, a competitor mentioned, a link, an email address or a phone number;
- a reply in a language the tool does not handle (French, English, German, Italian and Spanish are covered).
In all these cases, the conversation stays in your list, with the reason in one sentence. You reply by hand.
This caution costs replies. A person who writes "yes, but how much does it cost?" falls outside the automatic scope, because an answer about price commits your company. It is a design choice: it is better to answer fewer people than to say something false.
Which safeguards limit the sending?
Several rules stack up, and you cannot get around them from the screen.
- One reply per person. After a first automatic message, the AI no longer answers that person. The rest is up to you.
- A cap of 20 replies per day, within the message limits of your LinkedIn account.
- Working hours. Messages go out within your time slots, in waves, about every five minutes.
- A wait before sending, of a few minutes, so that nothing is answered to the second.
- A maximum age. A reply that is too old is not handled automatically.
- A check of the text before each send: length, language, allowed links, and no invented figure or amount.
The limit of these rules: they reduce errors, they do not remove them. A misread agreement is still possible, for example an ironic sentence read as interest.
The "I approve" mode: how does it work?
This is the default setting. The AI does not reply alone, and you approve each reply you send. The Replies page offers a draft, you read it, and you click. We describe this path in the article on the single inbox.
The second mode, "The AI replies to simple cases", needs an explicit activation. A window shows you what the AI will and will not do, and then asks you to acknowledge that these messages go out under your name without review and that the AI can be wrong. You can go back to manual approval at any time.
Only replies received after the activation are handled. The backlog from past weeks does not receive a burst of automatic messages.
When should you turn them on?
Turn them on when three conditions are met.
Your volume is more than you can handle in time. If you answer all your conversations within the day, automation brings you almost nothing. It helps most when the "yes" replies arrive in the evening or on weekends.
Your typical replies are stable. If the answer to an expression of interest is always "here is my meeting link", the prepared sentence fits. If every case needs some nuance, manual approval is better.
Your meeting link is ready. Without it, the AI does not reply alone: it warns you and you answer. Add the link in the settings before the activation.
Keep an eye on the log during the first days. If a message bothers you, pause the feature, and then fix your campaign profile.
When should you avoid them?
Avoid them in four situations.
- Your LinkedIn account is new or barely active. An account that suddenly sends many messages draws attention. Scale up step by step, following our guide to safe limits.
- Your target is narrow and sensitive. When every contact counts, a prepared sentence is a risk you do not want to take.
- Your offer needs nuance. A complex, regulated or expensive product cannot be summed up in a meeting link.
- You have not read your own replies yet. Without knowing the real cases, you do not know what you are automating.
How do you take back control?
Three actions exist, and they do not do the same thing.
- Pause: everything stops, and the messages that are ready or queued are removed.
- Take over: the AI no longer replies to one specific person, and its pending messages are removed. The button is in the header of the conversation.
- Remove pending messages: only the planned message is cancelled. The AI can still reply to the next message.
If the person writes again before the send, or if you handle the conversation yourself, the planned message is cancelled and the cancellation is noted in the log.
Limits
Here is what to keep in mind before you turn the feature on.
- The risk of restriction exists. LinkedIn limits accounts that are very active, and sending messages automatically can lead to it. No setting brings the risk down to zero.
- The AI can be wrong. It reads words, not intent. A message goes out under your name, without review.
- The scope is narrow. Only short and clear agreements are handled.
- The extension must be open. A message that is ready waits if Chrome is closed or if LinkedIn is not logged in.
- The feature can be switched off. A kill switch exists on the server side, and it stops all automatic sending.
For the general caps, see the LinkedIn limits page.
FAQ
Are auto-replies turned on by default?
No. The starting setting is manual approval. Turning them on requires reading an information window and confirming.
How many replies does the AI send per day?
Twenty at most, within the message caps of your LinkedIn account and within your working hours. The screen shows how many are waiting and how many went out today.
Does the AI reply to a refusal?
No. A refusal, an unsubscribe or an automatic absence receives no message. The person stays to be handled, and an unsubscribe places them in your "do not contact" list.
Can I read what the AI sent?
Yes. The "What the AI did" log shows each case, its reason and the exact text, whether it was sent, cancelled or left to you.
Decide with your own data
Start in "I approve" mode, read your first twenty replies, and then decide. Try Sales Farmer free for 14 days, no credit card needed.