Skip to main content
Auto Follow-up is an automated follow-up messaging mechanism. When a client stops responding, the AI employee will reach out after a specified delay — naturally, within the context of the conversation, without copy-paste templates.
Auto Follow-up exists in two different places in the interface.
  • Auto Follow-up section in the left menu — a monitoring dashboard showing all chats with scheduled reminders across all employees.
  • Auto Follow-up tab in a specific AI employee’s settings — where you enable the feature, set delays, and write instructions for each step.

Monitoring Dashboard (Left Menu)

The Auto Follow-up section in the left menu is a unified read-only panel displaying all chats with scheduled follow-up messages across all your AI employees. You don’t configure chains here — you observe what’s already running. At the top of the page is a search bar: “Search by name, agent, or client”. You can search by:
  • Client name (e.g., “Michael”)
  • AI employee name (e.g., “Alexey”)
  • Any keyword from the chat name

Filter by Employee

Below the search bar is a dropdown: “All employees”. By default, chats from all AI employees are displayed. Select a specific employee to see only their chats.

Chat Table

The main part of the page is a two-column table:
ColumnWhat it shows
ChatClient name or chat title where auto follow-up is active
EmployeeName of the AI employee managing the conversation and sending the follow-up
Each row represents one active chat where the client has gone silent and an automatic reminder is scheduled. Click a row to open the full conversation history.
If one employee consistently has many chats in follow-up — consider reviewing their Persona or improving their initial response.
Only chats with active auto follow-up appear in this section. If an employee has auto follow-up disabled, their chats won’t show up here.

Configuring Auto Follow-up (Employee Tab)

Open the desired AI employee and navigate to the Auto Follow-up tab.

Enabling and Disabling

At the top of the screen is the “Enable Auto Follow-up” toggle — it turns the feature on or off entirely for this employee.

Reminder Chain Steps

Below the toggle are steps — sequential reminders that the employee will send to the client one after another if they continue to be silent.

Parameters for Each Step

ParameterTypeDescription
Delay (H / MIN)Numeric fieldsWait time before sending this reminder (hours and minutes)
Message instructionText field (required)Description of what the agent should write to the client. The agent composes the message on its own based on this instruction and the conversation context
Run once per chatCheckboxIf enabled, this step fires only once per chat — even if the client goes silent again

How the Chain Works

1

Client goes silent

After the client’s last message, the first step timer starts.
2

First reminder

When the first step’s delay expires, the employee sends a reminder.
3

Subsequent steps

If the client remains silent, each next step’s delay counts from the moment the previous reminder was sent.
4

Client replies — reset

If the client sends a message at any stage, the timer resets and the chain starts over from step one.

Step-by-Step Setup

1

Open the Auto Follow-up tab

Go to the desired AI employee’s settings and select the Auto Follow-up tab.
2

Enable the toggle

Switch Enable Auto Follow-up to the active position.
3

Add a step

Click Add step, set the delay (hours/minutes), and write the message instruction.
4

Build the chain

Add multiple steps with increasing delays for a complete reminder chain.
5

Save changes

Click the save button. Once a client goes silent, the chat will automatically appear in the Auto Follow-up section in the left menu.

Example Three-Step Chain

StepDelayInstruction
14 hoursRemind the user that we’re available and ready to help
224 hoursAsk if the question is still relevant
348 hoursSay a warm goodbye and let them know we’re always happy to help when needed

Recommendations

Don’t overdo frequent reminders. The recommended pause between steps is at least 2-4 hours. Messages that are too frequent annoy clients and may lead to blocking.
An instruction is not a ready-made message text — it’s a task for the AI. The employee reads the instruction, considers the conversation context, and composes the message on its own.Good:
“Gently remind that we’re available. Ask if there are any remaining questions. Don’t be pushy.”
Bad:
“Write: Hello! We’d like to remind you that you can…”
More examples of good instructions:
  • “Check if the client has made a decision. Offer help if they need more information.”
  • “Remind about the special offer that was discussed. Don’t push, just refresh their memory.”
  • “Ask if everything is okay. If the client doesn’t respond, say a warm goodbye.”
The more specific the task — the better the result. But don’t write the actual message text — let the employee formulate the message naturally.
Persona. The reminder style depends on the employee’s persona. If the persona uses a friendly tone — reminders will follow the same style. Auto follow-up doesn’t override the persona; it works on top of it.Working Hours. If the employee has a schedule configured — reminders will only be sent during working hours. Make sure delays and the schedule don’t conflict.Channels. Auto follow-up works across all connected channels — Telegram, WhatsApp, website widget, and others.

FAQ

Most likely, auto follow-up isn’t enabled for any employee. Check the Auto Follow-up tab in each employee’s settings — the toggle must be active and at least one step with an instruction must be added.
The timer resets. The countdown starts over from the first step.
No. You write an instruction — a task description — and the agent composes the message on its own, considering the conversation context and its Persona.
There is no limit on the number of steps. In practice, 2-3 steps is optimal. More than three reminders are usually perceived as spam.
Only for the first step. Each subsequent step’s delay counts from the moment the previous reminder was sent.
Yes, across all connected channels for the agent.
If the employee has Working Hours configured, the reminder will wait until the start of working hours.
Yes, each sent reminder consumes tokens just like a regular agent response.
No, auto follow-up is configured at the agent level and applies to all their chats. If a client asks not to be contacted, add the appropriate instruction to the Persona.