Connects to your existing VoIP call centre
Your telephony stays as it is and your numbers do not change. ArtinBot sits as a layer on top of the call centre you already run and answers incoming calls.
A large share of the calls reaching your call centre every day are the same handful of questions: opening hours, branch address, order status, warranty terms. ArtinBot sits on top of your existing VoIP call centre, recognises Persian speech, answers from your own material and logs caller details into your CRM automatically. Whenever it is out of its depth, the call goes to an agent.
A traditional phone menu asks the caller to listen to a numeric tree and press buttons: press 1 for sales, press 2 for support. The trouble is that callers rarely know which branch their problem belongs to. They go several levels deep, land in the wrong place and start over. The result is abandoned calls and agents who spend their day transferring people rather than solving problems.
A voice bot changes that: the caller speaks in their own words and gets an answer directly. What sets ArtinBot apart is where the answer comes from. It is built from your real business material — website pages, catalogue, warranty policy, price list — rather than a pre-written script. Strict mode is on by default: if the answer is not in your sources, it does not guess. It says so and hands the call to an agent.
Your telephony stays as it is and your numbers do not change. ArtinBot sits as a layer on top of the call centre you already run and answers incoming calls.
The caller's speech is recognised in Persian and turned into text, so the same engine that answers in chat can answer on the phone.
Answers are spoken in a natural voice rather than the clipped, robotic tone of older recorded-message systems.
Details exchanged during the conversation are written to your CRM automatically, with nobody keying them in after the call ends.
The same sources the website chatbot learns from — files, links and your sitemap — sit behind the phone line too. One knowledge base, two channels.
Where the bot is unsure, or the caller asks to speak to a person, the call is passed to one of your agents.
Where hundreds of calls a day arrive with a few fixed questions, an automated first layer has the biggest effect and agents get to the real calls.
Stock, pricing, returns policy and order tracking — the same questions that dominate chat also dominate the phone line.
Address, opening hours, which doctor is in and how appointments work. Administrative questions that do not need a receptionist on the line.
Warranty terms, required documents and how to file a service request, instead of an engineer repeating the same explanation each time.
Fees, syllabus, prerequisites and start dates — the questions that multiply call volume during enrolment season.
Visiting guidance, required paperwork and process steps, to cut the calls that only need one piece of information.
A fixed numeric tree the caller has to navigate.
The caller speaks; the answer comes from your own material.
No. ArtinBot sits on top of the VoIP call centre your organisation already runs rather than replacing it. Your numbers, extensions and existing structure stay where they are, and the bot is added as an answering layer.
This solution runs on a VoIP foundation. If your organisation does not yet have a network-based call centre, that infrastructure has to be in place first. ArtinBot does not sell call centre or telephony systems.
It does not guess. Strict mode is on by default, so if the answer is not in the sources you gave the bot, it says so and the call is transferred to one of your agents.
From the same sources the website chatbot is trained on: PDF, Word and Excel files, page links and your sitemap. You do not have to build and maintain a separate knowledge base for the phone.
Details raised during the conversation are written to your CRM automatically, without anyone entering them by hand once the call has ended.
Speech recognition and voice replies are built for Persian, meaning the engine works on Persian speech rather than a general model that supports Persian only at the margins.
The channel differs; the knowledge is the same. The bot that answers on your site and in Instagram direct messages draws on the same sources over the phone. The difference is the speech layer: Persian speech recognition on the way in, a spoken reply on the way out.
Tell us about your current VoIP setup and we will map out the connection and what needs to be ready beforehand.
آرتینمگ
این مطالب در آرتینمگ، رسانهی تخصصی گروه آرتین، منتشر شدهاند.

جستوجوی کلیدواژهای داخلی سازمان چرا اغلب جواب نمیدهد؟ پنج محدودیت واقعی جستوجوی سنتی و اینکه جستوجوی معنایی دقیقاً چه چیزی را متفاوت انجام میدهد.

دانش سازمان معمولاً فقط در دست چند نفر خاص است. سه مانع رایج و چهار اصل عملی برای اینکه هر کارمند، صرفنظر از سمت یا مهارت فنی، به دانش سازمان دسترسی داشته باشد.

سازمان شما هزاران سند دارد، اما آیا این یعنی دانش سازمانی هم دارید؟ تفاوت بین انباشت اسناد و دانش قابلاستفاده، و چهار نشانهای که این تفاوت را نشان میدهند.

چرا کپی-پیست اطلاعات سازمان در ChatGPT کافی نیست؟ پنج تفاوت اصلی هوش مصنوعی سازمانی با ChatGPT، از دسترسی به داده واقعی تا امنیت و ردیابی پاسخ.

دانش سازمان در ایمیل، درایو و چتهای مختلف پخش شده و پیدا کردنش سخت است؟ سه اصل عملی برای مدیریت دانش پراکنده، بدون نیاز به جابهجا کردن همهچیز.

چرا در بسیاری از سازمانها، پیدا کردن یک پاسخ ساده دقیقهها یا ساعتها طول میکشد؟ بررسی ریشه واقعی این مشکل و هزینه پنهانش برای سازمان.
شرکت هوش گستر فردا آرتین چند محصول هوش مصنوعی فارسی میسازد. هر کدام کار خودش را میکند — اینها را ببینید تا بدانید کدام به کار شما میآید.