“My printer won’t connect. Who should I contact?”
“Which request should I submit for a VPN issue to avoid being rejected?”
“The knowledge base has a guide, but I can’t find the right one after searching for ages…”
These are the questions IT service desks hear dozens of times every day.
Traditional ITSM has successfully standardized service processes, but two challenges remain unresolved: knowledge bases are built but underused and quickly become outdated; increasingly complex service catalogs make users hesitant to submit requests. As a result, engineers are overwhelmed by repetitive inquiries, while minor business issues can eventually escalate into major incidents.
ITSM does not simply add a chatbot to existing systems. Instead, it introduces AI Digital Employees into the IT service operation lifecycle, enabling self-evolving knowledge management and proactive service catalog adoption. (Note: In this article, CanWay BlueWhale AI ITSM refers to the intelligent service scenarios built on CanWay BlueWhale IT Service Management Center ITSM・Jingmai (hereinafter referred to as ITSM).)
This article explores the proven business value of this approach from three operational perspectives.
01 Knowledge Operations
Most Knowledge Bases Stop Evolving After Initial Deployment
Organizations invest significant effort in building knowledge bases, creating operation manuals, and organizing FAQs. However, when users encounter issues, their actual behavior often follows this pattern:
Avoid searching documentation → Ask someone directly → Unable to find the right person → Submit a ticket and wait in line → Engineers discover it is another frequently asked question.
Knowledge remains unused in repositories, turning engineers into “human search engines.” Even worse, knowledge updates rely heavily on manual maintenance, meaning newly created content can quickly become outdated.
ITSM Knowledge Operations: A Four-Step Loop for Continuous Improvement
ITSM knowledge operations go beyond simply importing content into a system. Instead, it establishes a complete framework based on four key stages: knowledge creation, knowledge application, knowledge optimization, and operational monitoring.
The knowledge base serves as the foundation, while the four stages work together in a continuous cycle to create a complete life-cycle from knowledge creation to long-term operations.
The quality of knowledge base is measured through three core metrics: accuracy, recall, and coverage.
Coverage measures whether knowledge for key business scenarios has been incorporated into the knowledge base. Accuracy evaluates consistency between AI responses and human validation results. Recall measures whether user queries can successfully retrieve relevant knowledge.
Based on these metrics, the system automatically generates knowledge optimization tasks, enabling targeted improvements:
Automatic identification of unanswered queries — captures questions that AI cannot resolve and categorizes them as missing knowledge.
Low-confidence detection — automatically flags responses below confidence thresholds for human review.
Missing knowledge discovery — automatically triggers knowledge creation when frequently asked questions lack relevant entries.
High-frequency issue optimization — improves frequently accessed but low-satisfaction knowledge items.
Knowledge expiration alerts — automatically sends update reminders based on knowledge life-cycle attributes.
Step 1: Corpus Construction
Corpus can be integrated through two approaches:
First, external knowledge sources such as WeCom documents, Feishu documents, and customer-owned knowledge bases can be connected through the knowledge integration library, enabling one-time configuration and automatic synchronization.
Second, internal knowledge sources, including ITSM knowledge articles and historical tickets, can be directly incorporated through customized corpus repositories.
A knowledge graph is then established to structurally organize information based on business scenarios, knowledge categories, and relationships between entities, creating a searchable and traceable knowledge network.
Knowledge is further organized into employee-specific knowledge bases by scenario and assigned to different AI Digital Employees. For example, the “Intelligent Service Agent” handles FAQs, while the “Incident Diagnosis Expert” uses troubleshooting guides, allowing each AI employee to focus on its designated responsibilities.
More importantly, after AI completes ticket processing, the system can automatically extract resolution steps and outcomes, convert them into reusable knowledge, and feed them back into employee knowledge bases. This enables knowledge system to continuously grow through real-world operations.
Extend the corpus through plugins
Equip each AI agent with a tailored corpus to ensure more accurate responses
Step 2: Corpus Application
AI Digital Employees consume assigned knowledge bases across scenarios such as intelligent Q&A and ticket processing. When an employee asks in an instant messaging platform, “How do I reset my password?”, the AI service agent retrieves relevant knowledge and provides an answer. When engineers handle incident tickets, AI automatically matches troubleshooting guides based on issue descriptions and recommends diagnostic steps.
Every knowledge retrieval generates new operational data. The system continuously tracks key metrics, including accuracy, recall, and coverage, identifying which knowledge items are effective and which issues remain unresolved. These insights provide data-driven support for future knowledge optimization. Knowledge management is not completed when content is stored. Instead, knowledge is continuously validated, optimized, and improved through ongoing usage.
Automatically convert tickets into knowledge entries and sink them into the corpus
Step 3: Corpus Optimization
Corpus optimization is driven by operational metrics generated during knowledge usage, including accuracy, recall, and coverage.
The system automatically creates optimization tasks based on metric performance, ensuring every improvement action is supported by measurable data:
When recall is insufficient, the system automatically records unresolved queries and categorizes them as missing knowledge. Frequently unresolved issues can trigger new knowledge creation tasks. When accuracy falls below expectations, low-confidence responses are automatically flagged for human review. When coverage gaps are detected, the system identifies missing scenarios and provides additional knowledge.
The system also supports targeted optimization for frequently accessed but low-satisfaction knowledge items and automatically reminds operators to update outdated content based on knowledge life-cycle information. With these metrics and optimization tasks, operators can continuously improve knowledge quality in a targeted manner.
The agent corpus dashboard displays resolution rates and usage volumes, and automatically generates optimization tasks
Step 4: Corpus Monitoring
The goal of the fourth stage is to make knowledge operations measurable and manageable.
Through the knowledge health dashboard, administrators can gain a comprehensive view of knowledge performance: Each knowledge item displays resolution rate, usage trends, and source information, including the ITSM knowledge center, AI/Dev knowledge bases, and third-party plugin. Overall knowledge performance trends are updated in real time, allowing abnormal patterns to be identified quickly. For knowledge items with low resolution rates, operators can directly initiate optimization workflows and track improvement results, creating a continuous cycle of monitoring → optimization → validation.
Knowledge operations are no longer a one-time deployment activity. Instead, they become an ongoing management model supported by measurable metrics, actionable tasks, and operational visibility.
Knowledge acquisition → knowledge application → knowledge optimization → operational monitoring together form a complete life-cycle from knowledge creation to continuous operations.
Knowledge is no longer created only through manual documentation; it continuously evolves through system operations. The more issues and tickets AI handles, the more operational experience it accumulates, continuously improving service accuracy and creating a self-enhancing intelligent operations cycle.
Corpus Health Dashboard
02 Service Catalog Adoption
(01) A More Detailed Service Catalog Does Not Always Mean Better Adoption
Employees open the ticketing system and immediately face questions:
Which service category should I select?
What information should I provide in this field?
What does this required item mean?
After spending several minutes trying to complete the form, many users eventually abandon the request and postpone the issue.
The complexity of service catalogs has become a barrier to ticket submission. The more detailed a service catalog becomes, the more training users need to understand how to use it. As a result, organizations may end up with well-designed catalogs that are rarely used, while service requests continue to arrive through phone calls, instant messaging, and emails.
(02) CanWay BlueWhale AI ITSM Approach
Users Describe Their Needs, and AI Finds the Right Service
An employee simply types: “My computer has a blue screen issue, and restarting doesn’t help.”AI automatically identifies the issue type and maps it to the appropriate service category and support team. If required information is missing, AI proactively asks follow-up questions and completes the necessary details through conversation.Once the conversation is complete, the system automatically creates the ticket, which users can review and submit.
In short: Describe your request naturally, and AI handles the rest. Users no longer need to understand complex service catalog structures or classification rules. AI handles service discovery and request routing automatically.
Seamless Transition from Q&A to Ticket Submission
A user may initially ask a simple question but later discover that a service request is required.
Traditionally, users need to switch pages and manually create a ticket. With AI-powered service interaction, the entire process can be completed within the same conversation. Conversation context is automatically converted into ticket information, eliminating duplicate input and improving request efficiency. The service catalog experience is transformed from a complex form-based interface into a familiar conversational interaction.
Seamless transition from smart Q&A to AI‑assisted ticketing
AI Workflow Automation: Improving the Entire Ticket Life-cycle
AI Digital Employees can be embedded throughout the ticket life-cycle: Automatically identify request categories and service types, Automatically complete ticket information, Automatically generate diagnostic reports, Automatically record review and processing logs. The entire workflow remains transparent and traceable rather than operating as a black box.
Engineers can reduce repetitive tasks and focus on complex troubleshooting and problem resolution.
AI agents are embedded into workflow nodes and expose execution logs for full transparency
03 Omnichannel IM Access
The intelligent Q&A and service request capabilities described above are not limited to the ITSM platform itself. Employees can access AI-powered services directly through messaging platforms without opening the ITSM system.
Enterprise Messaging Bot: The Entry for AI Digital Employees
Through enterprise messaging integration, organizations can assign different AI Digital Employees to different scenarios:
Incident Reporting → “Incident Management Expert”
Handles incident intake, assists diagnosis, and executes incident management workflows.
Access Requests → “Service Desk Agent”
Handles various service request scenarios.
Change Requests → “Change Management Expert”
Supports change submission and risk assessment.
Users do not need to log into the ITSM system. By simply mentioning the bot in messaging platforms, they can complete inquiries or submit requests through a few conversational steps.
Configure WeCom smart bots and attach AI agents to enable intelligent Q&A and ticketing within WeCom
AI-to-Human Handoff: Seamless Escalation with Full Context Preservation
AI service agents are not designed to handle every situation. When encountering complex incidents, sensitive access requests, or user actively requested manual support, the system supports seamless escalation to human agents.
During the transfer process, AI automatically transfers the complete interaction context, including: Original user request details, Troubleshooting steps already performed, Key information collected during the conversation, AI analysis results and recommendations.
Human engineers do not need to ask users to repeat their issues and can immediately continue troubleshooting. After resolution, the solution can also be captured back into the knowledge base, creating a continuous improvement cycle: AI Service → Human Expertise → Knowledge Capture → Improved AI Capability
One Time Configure, Deploy Across Multiple Channels
AI Digital Employees, tools, and knowledge bases configured in the AI Employee Center can be simultaneously deployed to both the ITSM Xiaojing conversational interface and enterprise messaging bots. This is not two separate systems with duplicated logic. It is the same AI Digital Employee operating across multiple service channels. A single configuration update is applied across all channels, significantly reducing maintenance fees.
Users can ask questions and submit tickets directly within the WeCom chat with the smart bot
04 Core Design Philosophy of CanWay BlueWhale AI ITSM
The knowledge operations and service catalog capabilities described above are powered by the same underlying platform: Canway ITSM AI Digital Employee System.
A key design principle is: All AI scenarios are configured through visual interfaces rather than requiring custom software development.
Business administrators can independently configure: AI Employee creation, Knowledge base assignment, Workflow orchestration, Channel integration. No coding is required, no dedicated development resources are needed, and deployment does not depend on lengthy development cycles.
This enables:
Rapid AI Employee Deployment: An AI Employee can be configured and deployed within minutes.
Rapid Business Adaptation: When business requirements change, administrators can update configurations independently and apply changes immediately.
Lower Experimentation Costs: New business scenarios can be tested through configuration instead of lengthy development cycles.
ITSM does not provide only a set of predefined AI features. Instead, it provides a complete framework for building AI Digital Employees.
The framework includes:
Role-based AI Digital Employee definition
AI Tool Set
Knowledge Plugin Library
Employee Knowledge Bases
AI Workflow Nodes
Hybrid Search Capability
Omnichannel access capabilities
These capability modules work like building blocks, enabling organizations to quickly create AI Digital Employee teams tailored to their IT service scenarios. A Complete AI ITSM Capability Framework Covering the Entire IT Service Life-cycle.
ITSM provides a comprehensive AI ITSM solution covering end-to-end IT service management scenarios.
Key capability modules include:
Role-Based AI Digital Employees-Each AI employee has clearly defined responsibilities, objectives, and execution paths, similar to a real organizational role, ensuring accurate and controllable service delivery.
AI Tool Set-Integrates with CMDB, automation platforms, monitoring systems, and other IT tools, enabling AI to not only answer questions but also perform queries, analysis, and operational actions.
Knowledge Plugin Library-Connects external knowledge sources such as enterprise documents and third-party knowledge bases through centralized configuration and automatic synchronization.
Employee Knowledge Base-Organizes knowledge by business scenarios, assigns relevant content to specific AI employees, and supports dynamic updates.
AI Workflow Nodes-Embedded throughout ticket workflows to automate classification, information completion, diagnosis, review, and other standard tasks with transparent execution logs.
Omnichannel Reuse-The ITSM Xiaojing assistant and enterprise messaging bots share the same AI employee capabilities, enabling single configuration and multi-channel deployment.
05 Conclusion
We believe successful AI adoption is not about showcasing technology, but about solving real business challenges.
CanWay BlueWhale AI ITSM's AI Digital Employees are not simply an additional chatbot interface layered onto existing systems. From no-code role-based AI employee configuration to hybrid retrieval-powered intelligent Q&A, conversational service requests, and omnichannel IM integration, each capability is designed to address real challenges in IT service management.
Addressing Limited Service Capacity: Provides 24/7 support for repetitive inquiries and standardized service requests, reducing frontline workload and allowing engineers to focus on complex issues.
Improving User Experience: Transforms complex service request processes into natural conversations, enabling employees to complete inquiries and submit requests without learning complicated service catalog structures.
Activating Underutilized Knowledge: Unifies internal and external knowledge sources, automatically captures service resolution experience, and enables continuous knowledge accumulation and reuse.
Simplifying AI Implementation: With no-code configuration, compatibility with existing ITSM workflows, and multi-channel deployment, organizations can reduce implementation complexity without lengthy development cycles.
AI-powered IT services should not remain a future vision; they should become practical productivity tools that deliver value today.
If your organization is facing challenges such as slow response times, complicated service requests, underutilized knowledge bases, or engineers spending too much time on repetitive tasks, contact us for a personalized demonstration and explore how AI Digital Employees can help transform your IT service operations.
