02 / GIFSChatbots and AI agents
The customer asks.
The bot helps.
Answers from your materials, clarifies details and passes enquiries to an employee. Inside the company, it helps find instructions and prepare documents.
I need a website for 80 products.
The photographs are already available.
The photos are ready.
Clarify: features, timeline and budget.
bear fruit.
From a question to the next step.
Choose who the bot helps and go through the example at your own pace.
The bot will clarify.
A person will take over.
Go through a short conversation to see the result. The data in this example is for demonstration.
- Question
- Details
- Handover to an employee
This is what a scenario could look like. We will configure your bot with your own materials, rules, and available actions.
Each has its own job.
Bot
Finds answers in your materials. Clarifies the request. Prepares an enquiry or a draft.
Employee
Confirms prices, deadlines, and important actions. Steps in when a question goes beyond the agreed scope.
Let’s start by talking about your business.
You can come with a question or an example. We will help define the scope, stages, and what is needed to start.
- 01
We will choose one scenario
We will use common customer questions or a team workflow task.
- 02
We will set it up and test it
We prepare the materials, responses, integrations and handover to a person.
- 03
We will hand it over for use
We will test routine and complex questions and agree on support.
Price and timeline follow clarification of the task.
We agree on the scope of work before starting.
Before the project begins
Who can an AI agent be built for?
We develop chatbots and AI agents for businesses across Kazakhstan. For employees: searching materials, documents, and meeting summaries. For customers: answers, selection, and inquiries. For the whole company: one agent or multiple roles. We agree on sources, access permissions, and handoffs to people.
Can enquiries be received from WhatsApp and Telegram?
Yes. If both messengers are needed, we first check access and agree on where to send requests. The manager will see which channel an inquiry came from and what the customer has already told the bot.
Does a bot always need artificial intelligence?
Buttons and rules may be enough for a short, familiar scenario. AI is needed when it is important to interpret free-form questions and find answers in approved materials. Before development, we choose the approach, communication channel and test questions. If the bot does not know the answer, it should seek clarification or hand over to an employee.
What is needed for setup?
You need up-to-date instructions or answers, anonymised examples of requests and a person responsible for review. You can start with one scenario and expand the solution. If the agent cannot find a verified answer, it asks for clarification or passes the question to a person. For how to separate materials for customers and employees, see the guide on preparing a knowledge base for an AI agent.
Can existing business services be connected?
We check the required channel, CRM or documents, available actions and permissions. We agree the integration scope after review; an employee approves prices, deadlines, important messages and disputed decisions.
What determines the implementation cost?
It depends on the tasks and number of roles, the condition of materials, communication channels, connections to working systems, and review requirements. We discuss development, service fees, and support separately. After reviewing the process, we propose the scope of work and an estimate; multiple agents are needed when the task justifies splitting the roles.
How do you accept a finished solution?
We test routine and complex requests, access to internal data, enquiry saving, connection failures and handover to a person. With multiple agents, we test the entire chain. Use test questions and acceptance criteria for an AI agent. A broader process can be discussed in the section AI implementation and business process optimization.