Use cases for assistants that talk, remember, and get things done.
Assistant Core fits products where an assistant has to answer from your data, remember each user, take action through tools, and reach people on the website, in messaging apps, by voice, or on hardware.
- Support & sales on chat apps
- Internal knowledge assistant
- Voice on ESP32 devices
- Assistants deployed in production
- 50+
- Messaging channels, plus web, apps, voice, and devices
- 6
- To first audio in Realtime voice
- ~0.7s
Where teams use Assistant Core
Every use case follows one pattern: the assistant answers from trusted data, keeps context, and acts through tools you allow, on the channels your users already use.
Customer support & sales on chat apps
Answer product questions, policies, and order status on Zalo OA, Messenger, Instagram, WhatsApp, Telegram, TikTok Shop, and your website, all from one knowledge base.
- Six messaging channels plus the website widget
- Order and ticket lookup through your REST APIs
- Conversation history and feedback in the dashboard
Developers & SaaS
Add an assistant to landing pages, apps, dashboards, docs, or admin consoles without building the AI runtime yourself.
- Website widget, Chat API, and Voice API
- Origin allowlist for every assistant
- Tools that call your product APIs
Internal knowledge
Build assistants for HR, sales, onboarding, training, or knowledge management on your company's private documents.
- Roles and permissions per assistant
- Shared documents plus each user's private files
- Memory of each user's context
Personal assistants & experts
Assistants that remember each person, talk by voice, and run tasks on a schedule, or package an expert's knowledge for students, communities, and customers.
- Long-term memory per user
- Scheduled tasks, with optional email results
- Knowledge base from your own material
Smart devices & IoT
Bring voice assistants to smart homes, learning hardware, and compact ESP32 devices. Robots and smart toys are on the roadmap.
- Firmware for ESP32 and ESP32-S3 boards
- 6-character activation and over-the-air updates
- Device MCP to read sensors and control hardware
Workflow automation
Assistants can do more than answer: they can look up and update records in CRMs, ticketing, internal tools, or IoT systems, with every call traced.
- Custom REST tools with encrypted credentials
- MCP servers for external systems
- Tools enabled per assistant and per channel
How to choose a use case
A strong use case has a clear user, trusted data, a primary channel, and a bounded set of actions the assistant is allowed to take.
- 01
Pick the user and the channel
Start with one touchpoint: the website widget, a messaging channel such as Zalo OA, browser voice, an app, or a device.
- 02
Prepare trusted knowledge
Add documents, websites, FAQs, or business data to the knowledge base.
- 03
Allow only the actions it needs
Connect the tools and MCP servers for this use case and turn the rest off for the assistant.
- 04
Measure and expand
Track conversations, feedback, tool errors, and LLM traces, then add more channels once the use case works.
Use case FAQ
Explore next
Which use case should we start with?
Start with one that has clear data and measurable value, such as answering customer questions on your website and Zalo OA, an internal FAQ assistant, or a voice assistant for one device.
Can one assistant answer on our website and on Zalo at the same time?
Yes. Connect the channels to the same assistant and it answers everywhere with the same knowledge, memory, and tools.
Do we need a separate assistant for every use case?
Use separate assistants when the data, prompt, permissions, or domain differ, so each stays isolated. Several channels for one experience can share one assistant.
Can an assistant both chat and control devices?
Yes. The same assistant can chat on the web and call device tools to act on connected ESP32 hardware.
Hear it before you build it.
Talk to an Assistant Core assistant right now, then create your own in a few minutes.
Free plan available.