Automating support is four steps: connect the channels (Instagram, Telegram, WhatsApp, email, web chat), build the knowledge base, set up tickets with priorities and SLAs, and define the rules for escalating to an operator. After that the AI answers the repeat enquiries itself and passes the hard ones to a human with context.
| Step | What to do |
|---|---|
| 1. Connect the channels | Instagram Direct, Telegram, WhatsApp, email, web chat — into one point, the single inbox |
| 2. Build the knowledge base | Frequent questions, delivery and returns terms, prices, opening hours, policies |
| 3. Set up tickets and SLAs | An enquiry that needs action becomes a ticket with a priority and a deadline |
| 4. Define escalation rules | What the AI resolves itself and what goes to an operator — with the whole exchange and the context |
Step 1. Connect the channels
Gather every channel your customers write to into one place: Instagram Direct, Telegram, WhatsApp, email, web chat. That way no enquiry is lost and the AI answers from a single point — the single inbox.
Step 2. Build the knowledge base
The knowledge base is what the AI answers from: frequent questions, delivery and returns terms, prices, opening hours, policies. The fuller the base, the more enquiries close without an operator.
- Collect the top 20 repeat questions.
- Add answers in Uzbek and Russian.
- Update the base whenever terms and prices change.
How the AI answers customers in their own language is covered in support in Uzbek and Russian.
Step 3. Set up tickets and SLAs
Every enquiry that needs action becomes a ticket with a priority and a deadline (the SLA). That makes it visible what is in progress, what is urgent and what has not been lost.
Step 4. Define the escalation rules
Decide what the AI resolves itself and what it hands to an operator — complaints, disputed amounts, non-standard cases. On escalation the operator receives the whole exchange and the customer context; the customer record lives in the AI-CRM.
What the business gets
- Routine closes without an operator — order status, prices, opening hours, returns.
- Nothing is lost — tickets and SLAs keep enquiries under control.
- Relief at night and at peak — the AI answers around the clock.
- Operators step in where it counts — with the full conversation context.
The full picture of AI support — from channels to escalation — is in the guide. The platform itself is on the AI support page. Onboarding takes from 2 hours for a small business and up to 30 days for an enterprise.
Frequently asked questions
Where do we start with automating support
By connecting the channels and building the knowledge base — that is exactly what the AI will answer customers from.
How long does launch take
For a small business, from 2 hours with the data ready. For an enterprise with complex integrations, up to 30 days.
What do we do with complex enquiries
Set escalation rules — the AI will pass those enquiries to an operator together with the context of the exchange.
Do we need a knowledge base in Uzbek
Yes — then the AI answers customers in both Uzbek and Russian, detecting the language of the enquiry automatically.
What is an SLA in support
It is the response and resolution time you have promised on a ticket. The AI tracks the deadline and flags anything approaching it.
Do operators stay in the process
Yes. The AI closes the routine, and operators work on complex and disputed enquiries with the full context in hand.