The reply depends on who happened to be free
The message arrives now, while the person who knows the answer is away or busy.
Conversations and AI
An AI employee that replies, qualifies, and books from your own business information
It takes customer inquiries on the channels you connect, answers from an approved knowledge base, collects what's needed, opens a booking or follow-up, then hands anything requiring human judgment to your team with the full context.
✓ Not a standalone chatbot✓ Answers from sources you approve✓ Human handover when needed
Illustrative example with sample data — results vary by business, setup, and conversation volume.
ServiceNot a ready-made chatbot: how we set up the AI Employee for your businessSee how we prepare its knowledge, permissions, and testing before go-live →Replies outside working hoursonce the channel and scenario are enabled
Conversations that carry contextthe customer doesn't start over each time
Booking inside the conversationonce the calendar and services are connected
An orderly handover to the teamwith a conversation summary and the reason for handover
The problem it solves
Your customer doesn't see your tools — only how fast you reply and how clear the experience is
When knowledge and follow-up are split across staff and channels, messages get lost, questions get repeated, or the booking is delayed.
Information doesn't travel with the customer
They're asked to repeat their name, their request, and what they were told every time they move between a channel and a person.
Booking and follow-up sit apart from the conversation
The answer happens in one place, the appointment in another, and follow-up depends on memory.
What does it actually do?
From the question to the next step — inside one conversation
What it does is set by the information, permissions, and channels you approve before it goes live.
Answers the recurring questions
It draws on the information, policies, and services you supplied and approved.
It doesn't invent information that isn't in its sources.Collects qualifying details
It asks about service type, location, timing, budget, or whatever the scenario requires.
The questions are defined in advance for your business.Opens the right booking
It offers available times or moves the customer into booking when the request fits.
Requires a connected calendar plus configured services and permissions.Records the outcome in the CRM
It saves the customer's details, their source, what they asked for, and the stage they reached.
According to the fields and workflow that were configured.Starts follow-up and reminders
It fires a message, a task, or a reminder based on the conversation's outcome instead of relying on memory.
Timing, content, and stop rules are approved before go-live.Hands over to a person when needed
When a case exceeds its permissions it passes to the team with the context, instead of the conversation starting over.
Handover rules are tested before activation.How the conversation moves
The conversation doesn't end at the reply — it turns into an action you can track
The path below is a general example. The details vary by your business, the channels you connect, and the permissions you grant.
The message arrives
from SMS, phone calls, Instagram, Messenger, or website chat once the channel is connected.
It understands the request
It reads the question and refers to the approved knowledge base and instructions.
It collects what's missing
It asks the qualifying questions and saves the answers to the customer record.
It executes the step
A booking, a follow-up, a task for the team, or a clear answer within its permissions.
It hands over when needed
It passes the conversation on with a summary and clear context when human judgment is required.
For the customerA faster answer and an experience that doesn't make them repeat themselves.
For your teamA clear record, a defined next step, and a conversation you can review.
Knowledge base and instructions
We don't train a new model — we configure an employee that works inside your rules
The knowledge, behavior, and permissions are built and tested before it ever deals with a customer.
What we prepare before go-live
We turn your business information into clear sources and instructions, then test the questions and the sensitive cases.
- Services, prices, policies, locations, and the recurring questions.
- Style, language, and the reply tone that fits the brand.
- The qualifying questions and what must be saved to the customer record.
- What it may do, and when it stops and hands the conversation over.
Services and pricesApproved source
Policies and recurring questionsApproved source
Behavior instructions and permissionsOperating rules
Handover and escalation casesClear limits
Illustrative example
An answer built on your business knowledge
It answers from the sources you approved, asks for clarification when information is missing, and takes no action beyond its permissions.
✓ It doesn't guess at information it doesn't have.
✓ It doesn't promise the customer anything your business hasn't approved.
✓ It passes sensitive or unclear cases to the team.
Not a conventional chatbot
The difference isn't how the chat looks — it's what it knows and what it can carry out
Combinu's AI employee
- Answers from your approved business knowledge.
- Saves the customer's details and the conversation context.
- Qualifies and opens the booking or follow-up.
- Works as part of the CRM, the calendar, and workflows.
- Hands the conversation to the team with a clear summary.
A conventional reply bot
- Usually relies on fixed paths and canned replies.
- May repeat questions when the customer moves channel.
- Answers without any real link to the next step.
- Runs as a tool separate from the customer record.
- Human handover may start with insufficient context.
Preparation before activation
Four steps before it replies to a single customer
We gather your business information
Services, prices, policies, locations, and recurring questions — from the sources you approve.
We define its role and permissions
Reply style, qualifying questions, what it may carry out, and when it stops and hands over.
We test it on real cases
Questions, objections, and sensitive cases — so we correct the knowledge and instructions before customers do.
We connect the channels and switch it on
After your approval it starts replying on the channels that have been connected and tested.
To be clear: We don't train a new AI model from scratch. We prepare its knowledge base, instructions, behavior, and permissions — then test them before activation.
Human control and measurement
It runs automatically within its limits — and your team keeps the decision
When the customer needs a person
Handover comes with a summary of what happened, what the customer asked for, and the suggested next step.
Can the appointment be changed after confirmation?
This case needs the team to review it. I'm handing the conversation over now.
Handover summaryConfirmed appointment · change requested · customer prefers evenings
How do we know operations are improving?
We agree on metrics that serve the actual goal — not generic numbers that fit any business.
Time to first replyQualification rateBookings producedHandovers to staffConversations with no follow-up
There are no guaranteed results — your offer, price, fit, and the customer's readiness are independent factors.
Questions worth asking before go-live
Limits, integrations, and human intervention
These answers are general — the scope of each channel and action is settled during setup and testing.
Does it start replying to my customers immediately?
It starts after the knowledge, instructions, and permissions are prepared, the channels connected, the scenarios tested — and you approve go-live.
What does it do if it doesn't know the answer?
It isn't supposed to guess. It asks for clarification or says the information isn't available, then hands the conversation to the team per the handover rules.
Can I step into the conversation?
Yes. The team can take the conversation over and continue it, keeping everything that happened before the handover.
Does it only work on one channel?
No. It can run across supported channels that are connected and tested, such as SMS, phone calls, Instagram, Messenger, and website chat.
Can it book and follow up automatically?
It can, once the calendar, services, and workflows are configured and connected, and the right permissions and stop rules are set.
Would the AI employee suit how your business works?
We review your channels, your recurring questions, your booking and follow-up — then decide what can be automated and what should stay with your team.
Book a diagnostic session