IT之家 AISource publication

Ant Insight 2.0 Release: Fully upgraded “capable AI assistant” breaks the single conversation limit

IT Home On October 27, Ant Group released Lingguang 2.0, enabling the AI assistant to evolve from passively responding to commands and completing single tasks to actively identifying needs and continuously managing affairs.

IT Home On October 27, Ant Group released Lingguang 2.0, enabling the AI assistant to evolve from passively responding to commands and completing single tasks to actively identifying needs and continuously managing affairs.

According to the official introduction of Ant Group, compared to traditional conversational AI and task-based Agents, one of the major changes in Lingguang 2.0 is further moving the starting point of AI involvement in tasks forward.

In the past, users had to identify problems and express needs first, after which the AI would take action. Now, with user authorization, Lingguang can combine the available information and services, actively identify potential needs, initiate analysis, monitoring, and follow-up, and provide timely reminders at points where users need attention. AI is no longer just a tool that waits for instructions; it has become an assistant capable of continuously managing tasks.

Behind this change is a comprehensive upgrade of Lingguang 2.0 in terms of cross-session memory, long-term task scheduling, and tool connection capabilities. With an independent cloud host, Lingguang can break through the limitations of a single conversation and continue running tasks after the user exits the conversation, and dynamically follow up based on new information and status changes, allowing AI to shift from one-time task execution to long-term, continuous事务 management.

In addition, Lingguang Management Services' capabilities are entering specific life scenarios.

  • In the consumption sector, after users authorize Alipay access, Lingguang can organize consumption bill records and identify potentially forgotten auto-renewal subscriptions, analyze consumption habits, and provide cost-saving suggestions, etc.;

  • In the travel industry, the completion of flight and hotel bookings does not mean the end of AI services. Lingguang can continue to monitor subsequent price changes and discover discount opportunities and notify users in time;

  • In a family scenario, after authorization, Lingguang can pay attention to the teacher notifications in the DingTalk class group, organize important information and extract tasks, and complete subsequent tasks such as document formatting.

In addition, for matters that require long-term memory and continuous follow-up, such as pet care, children’s vaccinations, family anniversaries, and vehicle annual inspections, Lingguang can proactively identify key points and arrange follow-up actions within its authorized scope.

According to the introduction, Lingguang 2.0 not only connects data and services within the Alipay ecosystem, but also supports the use of tools such as MCP and CLI in an open environment, thereby expanding cross-platform collaboration capabilities. Currently, Lingguang has built-in services for Alipay bills, memberships, travel, as well as DingTalk, Feishu, email, and Mijao. The management scope is expanding from personal information and online transactions to cross-platform workflows and smart home device control.

As agents move toward “task management,” their permission boundaries, data security, and trustworthiness of execution have also become new challenges. IT Home has learned that Lingguang 2.0 integrates with Ant APASS, which serves as an infrastructure for unified accounts and security management across agents, devices, and applications in the AI era. APASS incorporates identity authentication, intent security, and privacy protection into the agent execution process. Through mechanisms such as separating authorization keys from agents for hosting, agents can execute tasks without directly accessing the authorization keys. Users can view, adjust, or revoke authorizations at any time, allowing agents to perform tasks within their authorized scope.

Original source

IT之家 AI

Content notes

Original publication and rights belong to the source.

Machine translation · Refer to the original