This solution simplifies the entire previously complex installation and setup process, while allowing users to command the AI Agent to compose and send messages via Zalo. This plugin has now been officially released on the OpenClaw website.
Increase accessibility of Artificial AI for users.
OpenClaw is a project that applies Agent AI technology – a type of AI capable of autonomously acting to achieve a goal, instead of just answering questions like a typical chatbot. When given a task, the AI Agent can independently plan, break down tasks, use tools (web, data analysis, or software), monitor progress, and make adjustments if necessary.
![]() |
Information about the Zalo plugin can be found on the official OpenClaw website. |
Previously, Zalo provided a Bot API allowing users to interact with this AI Agent through the familiar chat window. However, as OpenClaw gradually moved beyond being just an experimental tool for technically savvy users, simplifying the setup process became essential.
Ms. Le Thi Kim Xuyen, Director of Digital Transformation at Zalo, stated: “Operations such as creating bots, authentication, and entering API keys are still very difficult for most users, becoming a barrier to accessing this agent AI. We expect the new refinements to help users avoid having to perform these in-depth technical steps themselves, while ensuring stable connectivity as OpenClaw continues to develop.”
Zalo's new plugin has solved this problem. Instead of having to go through many separate and complicated connection steps as before, users can complete the process quickly with a simple QR code scan.
![]() |
The steps to connect and set up OpenClaw to use Zalo Bot have become simpler. |
Currently, this experience is still more suitable for users who are willing to experiment with technology and proactively set up tools to suit their personal needs. However, streamlining operations such as bot creation or manual configuration is still an important step to bridge the gap between a technology originally intended for technical professionals and the broader user base on the Zalo platform.
Laying the technical groundwork and pursuing a long-term AI strategy.
Besides changes to the setup experience, the new plugin also marks a shift in how Zalo integrates deeply with AI agents. Zalo's direct development, updating, and management of its own plugin reduces the reliance on separate integrations for connecting OpenClaw on Zalo, while also allowing the platform to be more proactive in optimizing operations, adding features, and keeping up with changes from OpenClaw.
In the long term, the significance of this step lies beyond a single plugin. Perfecting the bot system, expanding APIs, and building supporting layers like mini-apps are all part of preparing the technical foundation. The goal is to be ready for more AI agents to appear on Zalo in a way that is easily accessible and closely aligned with users' daily habits. This is a crucial part of the strategy to popularize AI among the masses and pursue the messaging app's long-term AI strategy.
Prior to Agentic AI, Zalo had invested in AI in various aspects for many years. Features such as speech-to-text conversion, voice-composed messaging, audio message reading, and translation have gradually become a natural part of how users communicate. According to statistics, by 2025, Zalo expects approximately 30% of its messaging platform users to utilize AI features each month.
Not only has Zalo integrated AI, but the company has also demonstrated its capabilities in researching and developing independent AI products. Kiki Auto, a Vietnamese in-car assistant launched by Zalo, had nearly 1.5 million installations and users in Vietnam by the end of 2025. Currently, it is estimated that one in five cars on the road in Vietnam uses Kiki Auto.
![]() |
Kiki Auto, a Vietnamese in-car assistant launched by Zalo, has nearly 1.5 million installations. |
In 2023, while large language models (LLMs) like GPT-3.5 and GPT-4 were revolutionizing the technology world, Zalo chose to pursue its own independent AI model – a more arduous and challenging journey. Instead of taking shortcuts in the LLM training process by choosing readily available fine-tuned models from abroad, the company decided to train its model from scratch.
Source: https://znews.vn/zalo-cap-nhat-phuong-thuc-ket-noi-den-openclaw-post1664688.html










