Phone: 0850 346 41 09 | Weekdays 10:00–18:00 · Technical support 24/7
A custom chatbot is an AI virtual assistant that answers product, order and support questions around the clock on your website, mobile app or WhatsApp. It grounds its answers in your own catalogue, FAQs and order system rather than general internet content, and hands anything it cannot resolve to a live agent, along with the conversation so far.
Let a bot handle the repeat questions. A large share of what reaches customer service is the same handful of questions: "Where's my order?", "How do I return this?", "Is this the right product for me?". Messages sent after hours wait until morning, customers queue at peak times, and your agents spend much of the day typing the same replies.
We build a custom virtual assistant that answers those questions using your own product, order and support information. It takes its language skills from large language models (GPT, Claude, Gemini, or open-source models running in-house), but what it says is decided by your data. A customer who messages "Where's my order?" at midnight, for example, gets the delivery status from your order system as soon as they give their order number; someone asking about returns is taken through your own returns policy step by step. In the morning, your team deals only with the conversations the bot has passed on.
Answers around the clock on your website, mobile app or WhatsApp Business: product details, order status, return steps and FAQs. Complaints are routed to the right team and anything unresolved goes to a live agent.
Works over company policies, process guides and technical documents, so questions like "How do I book leave?" or "How do I enter an order in the ERP?" are answered before they reach HR or IT.
Clarifies what the customer needs with a few questions, then suggests suitable products from your catalogue with a side-by-side comparison. Undecided visitors are put in touch with your sales team.
Walks users through step-by-step troubleshooting for faults or installation queries. If they upload a photo, image analysis narrows down the issue; unresolved cases become support tickets with the details already captured.
If you would rather start with a ready-made product than a custom build for your website, SmartOperator, our AI customer agent from the Smart360 family, is added to your site with two lines of code: it answers from your knowledge base and product catalogue, adds items to the cart or books appointments through the actions your site defines, and hands the conversation to your team when it is unsure. You can buy SmartOperator now.
If the assistant is purely internal and works over procedures and documents, see our enterprise LLM assistant. If customer requests mostly arrive by email, SmartMail from the Smart360 family brings shared and personal mailboxes into one panel, summarises every incoming message and sorts it by content type without you writing rules. You can buy SmartMail now.
Generic chatbots don't know your company and guess when they're unsure. We make your product catalogue, manuals, FAQs and past support records the assistant's source of truth, so it answers questions about your business correctly and leaves what it doesn't know to a person.
For more detail, read our guides on which customer service questions to leave to a chatbot, setting up a WhatsApp Business chatbot and the difference between a chatbot and an AI customer agent. More articles are collected in our AI guides for business.
What if the bot says something wrong? Its sources are limited to your catalogue, manuals and FAQs, and when it is unsure it hands over to an agent rather than guessing. Unanswered questions are reported so the knowledge base can be filled in.
What if a customer doesn't want to talk to a bot? A route to an agent can be offered at every stage of the conversation; the agent takes over with the full history, so the customer never has to start again.
What happens to our customer data? The bot connects to your order and customer systems only for the information it needs, and if data must not leave your organisation it can run on your own servers with an open-source model. Data protection obligations under Turkey's KVKK are part of the design; see our data protection compliance page for more.
Tell us the questions your customers ask most and how you answer them today, and we will scope the right bot together. Book a call or ring 0850 346 41 09.
A custom chatbot is an AI virtual assistant that answers product, order and support questions around the clock on your website, mobile app or WhatsApp. It bases its answers on your own catalogue, FAQs and order system rather than general internet content, and hands anything it cannot resolve to a live agent along with the conversation so far.
The chatbot can run in a chat window on your website or mobile app and on WhatsApp through the WhatsApp Business API. It can connect to CRM, order and support ticket systems, so it can read information such as order status directly from the source. Conversations are supported in Turkish and English.
It does not guess. It passes the conversation to a live agent with the full history, so the customer does not have to repeat themselves. Unanswered questions appear in the conversation reports and are used to extend the knowledge base.
It can use large language models such as GPT, Claude or Gemini, or open-source models running in-house. Answers are grounded in your own documents using RAG, with the documents held in a vector database such as pgvector, Chroma or Pinecone. The model is chosen to match your data privacy requirements.
No. Alongside a customer service bot, it can work as an internal assistant over company policies and process guides, a sales assistant that recommends products from your catalogue, or a technical support bot that walks users through troubleshooting. If it will only work over internal documents, an enterprise LLM assistant may suit you better.
Running a chatbot on WhatsApp requires a verified business account on Meta's WhatsApp Business Platform (the Business API) and a phone number registered to it; the WhatsApp Business app on a phone is not suited to bot integration. Opening the account, getting message templates approved and connecting the bot to your order, CRM or support systems are all planned together as part of the project.
A project starts by deciding which questions the bot will answer and which channels it will run on. Your catalogue, FAQs and past support tickets then become its knowledge source, and the bot is tested with a limited group of users. Cost depends on the number of channels, CRM or order system integration, conversation volume and whether the model runs in the cloud or on your own servers; we confirm it in a written proposal after the needs analysis.
Have a different question? Ask Us