Automating support: a bot that answers customer questions instead of your team
Half of support inquiries are the same questions asked in different words. An AI bot takes them on and answers from your knowledge base, while the team focuses only on what genuinely needs a person.
Why support can't keep up, even when the team is doing fine
It’s usually not that employees work slowly. It’s just that nine out of ten new questions are variations on the same thing: how to pay, what’s included in the plan, where to write about a warranty. Answering them by hand every time is time that could have gone to customers with genuinely complex cases.
AI support takes on exactly this repetitive layer of inquiries. The bot answers from your company’s knowledge base rather than made-up text, and hands the conversation to a person the moment the question goes beyond the routine.
What the solution consists of
Automating incoming support
The bot takes on the main flow of routine questions, taking the load off the team for the most common topics.
Answers from the knowledge base
The bot answers based on your company's documents, instructions, and FAQs, not generic phrasing from the internet.
A chatbot FAQ format
The customer can simply pick a question from a list of popular topics, or write it in their own words - the bot recognizes the intent either way.
Tech support on Telegram
The same bot works in your Telegram support channel, if you already have a base of inquiries there.
Support on Viber
The same logic works on Viber for customers who are used to writing there.
Handing complex cases to a person
If a question needs a decision rather than reference information, the bot immediately hands off the conversation along with its context.
What happens at the setup stage
We look at which questions come up most often, and which documents the bot should base its answers on.
We sort frequent topics into scenarios and decide where the bot answers on its own, and where it hands off to a person.
Website, Telegram, Viber - depending on where your main flow of inquiries is.
We check how the bot handles actual messages, not just ideal phrasing.
The knowledge base gets expanded as new topics come up.
What this is often paired with
A narrower scenario focused specifically on working with the company's documentation and instructions.
A similar answering logic, but focused on helping choose a product or service rather than solving problems.
AI reception for business
If, beyond answering questions, you also need to take in inbound inquiries and book customers in.
What trust in our setup is built on
IntegraBAS Solution has been on the market since early 2025, and even before moving into AI, the team worked on setting up systems where data accuracy matters – integrations of BAS and 1C with CRMs.
This shapes how we build support bots too: the scenario is built not on generic phrases, but on the company’s real knowledge base, because an inaccurate answer in support costs more than no answer at all. The bot doesn’t make up information if it’s not in the knowledge base – it says it’s passing the question to a specialist.
After launch, we track which questions the bot couldn’t answer, and expand the knowledge base based on that, not just once at the start.
What people usually ask before setting up AI support
Describe where your support currently sits (website, Telegram, Viber) and which topics come up most often. We’ll estimate how many inquiries could be closed automatically and work out the cost.
AI support: how automating support removes repetitive questions from your team
At most companies, support inquiries split into two uneven parts. One part is genuinely unique situations that need to be handled individually. The other, usually larger part, is the same questions about payment, plans, timelines, and terms, just worded differently. Automating support with an AI bot takes on exactly that second part, leaving the team with what actually needs a human decision.
The bot answers customer questions from your knowledge base
The key difference from a plain button-based bot is that it phrases an answer based on the company’s documents and instructions, not generic phrases from the internet. If a customer asks about return terms, the bot answers according to the policy that actually applies at your company, not an averaged template. In this sense, a knowledge-base bot works as a first line of support: it handles routine topics and passes on to a person whatever the knowledge base doesn’t cover.
A chatbot FAQ format or free-form conversation
There are two ways to present a bot like this to customers. The first is a chatbot FAQ format, showing a list of frequent topics for the customer to pick from. The second is free text, where the customer writes the question in their own words and the bot recognizes the meaning and finds an answer. In practice, it’s more convenient to combine both: a list for those who want to find an answer quickly, and free-form input for those whose question is phrased differently from the standard list.
The same bot on Telegram and Viber
If a business has support spread across channels, a tech support bot on Telegram and a tech support bot for Viber are set up on the same answer logic, just with different technical integration for the specific messenger. It doesn’t matter to the customer where they wrote in – what matters is getting a fast, relevant answer.
If it’s not clear right now how many inquiries could realistically be closed automatically, describe in the inquiry where your customers most often write in – we’ll look at the actual topics of inquiries and suggest a configuration for your case.