Tencent Cloud ADP (Agent Development Platform) is an enterprise-grade intelligent agent development platform launched by Tencent Cloud, designed to help developers and business personnel quickly build, evaluate, deploy, and manage large model-driven agents through a low-code/no-code approach. The platform deeply integrates Tencent Cloud's underlying capabilities in artificial intelligence, cloud computing, and big data, and leverages Tencent Hunyuan large models along with multiple mainstream third-party model services to provide a full-chain toolset from model integration, prompt orchestration, knowledge base management, workflow design, to performance evaluation and security governance.

As a core component of Tencent Cloud's AI ecosystem, ADP aims to lower the barrier to agent application development and drive AI capabilities into enterprise-grade production environments. Its core competitiveness lies in its deep integration with the Tencent ecosystem (WeCom, Tencent Docs, Tencent Meeting, etc.), enabling agents to seamlessly embed into collaborative office scenarios and achieve a closed loop from "conversation" to "action." In terms of industry position, Tencent Cloud ADP is one of the first platforms among leading domestic cloud vendors to deeply bind agent development with collaborative office tools. Leveraging Tencent's social and enterprise service DNA, it holds a unique advantage in the government and enterprise market.

• Low-code visual orchestration - Build agent workflows through a drag-and-drop canvas, supporting complex logic such as conditional branches, loops, and parallel execution without writing traditional code.
• Multi-model integration and management - Built-in Tencent Hunyuan and mainstream open-source/commercial large models (e.g., DeepSeek, Llama, etc.), supporting unified API calls and model switching, with automatic recommendation of optimal models based on tasks.
• Knowledge base and RAG enhancement - Provides document upload, automatic chunking, vectorized storage, and retrieval-augmented generation (RAG) capabilities, allowing agents to accurately answer based on enterprise private knowledge.
• Agent evaluation and tuning - Supports automated evaluation set management, multi-dimensional scoring (accuracy, coherence, safety), and generates tuning recommendations to continuously improve performance.
• Enterprise-grade security governance - Provides fine-grained permission management, content moderation, sensitive information masking, and operation auditing to ensure agent behavior is compliant and controllable.
• Seamless integration with Tencent ecosystem - One-click publishing to applications such as WeCom, Tencent Docs, and Tencent Meeting, enabling conversational collaborative office work, such as automatically creating schedules and initiating approvals.

• Registration and login - Visit the Tencent Cloud official website, register an account using WeChat/email, complete enterprise or personal real-name authentication, then search for "ADP" in the console to enter the agent development platform.
• Create an agent and configure basic information - Click "New Agent," fill in the name and description, select the target model (e.g., Hunyuan), and set the system prompt to define its personality and behavioral boundaries.
• Orchestrate functional components - Add nodes such as "Knowledge Base," "API Call," and "Conditional Judgment" on the visual canvas, connect them to form a workflow; support importing enterprise documents as a knowledge base, or connecting to WeCom through pre-built connectors to complete instruction interactions.
• Test, evaluate, and publish - Enter test questions in the debug console to view responses, use the "evaluation set" for batch validation; after passing, click "Publish," select a channel (e.g., WeCom bot) to go live, and monitor call volume and performance in the backend at any time.

Example 1: WeCom intelligent assistant - Create an "HR Policy Q&A Bot" in ADP, upload the employee handbook to the knowledge base, and configure a "Reference WeCom department information" node. Employees ask "How are annual leave days calculated?" in WeCom, and the bot retrieves from the knowledge base and answers "If your service length is at least 1 year but less than 10 years, you are entitled to 5 days of annual leave," along with a link to the original policy. The entire process requires no coding and is ready for release.

Example 2: Automated sales lead follow-up - Build a "Sales Lead Filtering Agent" that connects to a CRM database (via an API node). When a new form submission triggers, the agent automatically extracts customer needs, calls the model to assess intent level, and sends a WeCom notification to the corresponding salesperson with suggested follow-up scripts, significantly improving response speed.

Example 3: Content creation assistance - Configure a "Marketing Copy Generator" with brand-style prompt settings and connect product materials from the knowledge base. Input "Write a WeChat Moments promotion copy for a new Bluetooth headset, highlighting noise cancellation," and the agent generates complete content including persuasive copy and hashtags, with one-click copying to Tencent Docs.

• Deep Tencent ecosystem integration - Agents can be connected to high-frequency office tools such as WeCom and Tencent Meeting without extra development, enabling closed-loop business actions, which most competitors find hard to match.
• Full-chain low-code - From model selection to workflow orchestration to evaluation and publishing, all operations are visual, allowing business personnel to get started, greatly lowering the barrier to AI application development.
• Security and compliance DNA - Leveraging Tencent Cloud's security system, it provides content moderation, data isolation, and operation auditing to meet compliance requirements in industries like finance and government.
• Model flexibility and cost optimization - Supports calling multiple models and selecting the most cost-effective option for each task, avoiding single-model lock-in, while offering pay-as-you-go and subscription pricing for cost control.
• Robust evaluation and tuning mechanisms - Built-in professional evaluation sets and multi-dimensional scoring systems help users scientifically quantify agent quality rather than relying on intuition, enhancing production-level reliability.

Baidu AI Cloud Qianfan AppBuilder - Also offers low-code agent development, but Qianfan places more emphasis on retrieval enhancement and multimodality, while Tencent Cloud ADP is significantly stronger in collaborative office integration, and the closed-loop experience within the WeCom ecosystem is a unique advantage.

Alibaba Cloud Bailian Platform - As the development platform for Tongyi Qianwen, Bailian has advantages in model foundation, but ADP's drag-and-drop orchestration and Tencent ecosystem linkage are superior in usability and deployment depth, especially for enterprises already using Tencent infrastructure.

iFlytek Spark Agent - iFlytek has distinctive strengths in voice interaction and education, but ADP is more comprehensive in general enterprise-level features (such as permission control and evaluation systems) and cloud-native scalability, with transparent and flexible pricing.

Tencent Cloud ADP adopts a hybrid "free tier + paid upgrades" pricing model. All registered users can use the basic orchestration features for free and receive a certain amount of free call quota (e.g., 10,000 model calls per month) for development and testing experience.

• Free edition - Includes all basic features, suitable for individual developers to try, with a daily limited number of model inference calls (specific counts may vary with promotions).
• Professional edition - Monthly subscription, typically priced at several thousand RMB per month, including a higher call quota (e.g., 100,000 calls/month), full evaluation features, multi-user collaboration, and priority technical support.
• Enterprise edition - Offers private deployment support, dedicated model channels, SSO single sign-on, customized permissions and auditing; pricing is tailored based on specific requirements, usually contracted annually.
• API pay-as-you-go - Beyond the free quota, billing is based on model type and token count; basic models like Tencent Hunyuan are competitively priced, supporting pay-as-you-go usage.

Specific prices are subject to real-time announcements on the Tencent Cloud official website; the platform frequently launches special offers for new users.

• Enterprise digitalization/operations staff - Responsible for internal AI tool implementation, seeking to quickly build business assistants (e.g., intelligent customer service, approval bots) while reducing development costs and cycles.
• Software developers and architects - Want to integrate AI capabilities into existing business systems, using ADP's API and plugin mechanisms to quickly realize agent integration and reduce underlying model adaptation work.
• Product managers and business analysts - Not skilled in coding but need to design agent interaction flows, using the low-code canvas and evaluation features to independently validate product ideas and push them to launch.
• Independent developers and entrepreneurs - Utilize the free edition and Tencent ecosystem to quickly develop small-scale agent applications for specific audiences, distributed through channels like WeCom with low marginal cost.

• Intelligent customer service and ticket processing - Combined with enterprise knowledge base, automatically answer customer inquiries and integrate with CRM systems to create tickets and assign support staff, building a fully automated response chain.
• Collaborative office automation - Publish agents in WeCom to automatically handle schedule management, approval reminders, meeting minutes generation, email drafting, and other daily office tasks.
• Data insights and report interpretation - Connect database or spreadsheet data to agents, enabling natural language queries to obtain visual charts and in-depth interpretations, assisting management decision-making.
• Personalized marketing and content production - Based on product library and brand tone, batch generate marketing copy, social media posts, and video scripts, with one-click integration to Tencent Ads platform.
• Education and training, knowledge sharing - Build online teaching assistants that retrieve course materials based on learner questions, provide answers, and automatically generate practice questions and knowledge point summaries.