CCNU: An AI Teaching Assistant Living Inside the Course Corpus

Specialist literature is hard to search and teachers drown in repetitive questions. Corpus-based structuring plus an LLM turns archival value into teaching capacity.

Customer case · Figures below are illustrative and will be replaced with project acceptance data before formal publication.

Customer Challenge

The Marxist Literary Criticism database at CCNU’s School of Literature holds a large body of specialist literature, but searching and studying it is hard for students and faculty, and keyword search cannot support course Q&A or guided reading. Teachers spend much of their time on repetitive questions, and the corpus has not yet turned into teaching capacity.

  • Course corpus structuring

    Specialist literature is structured and turned into a course knowledge base that can be updated continuously.

  • LLM teaching assistant

    A language model answers course questions, supports guided reading and suggests learning paths on demand.

  • Academic review

    Corpus outputs are admitted to the knowledge base only after review by a national comprehensive social science research institution, ensuring academic rigor.

Concept illustration of a course knowledge tree growing from books

Solution

Fractal Tech built a course agent (AI teaching assistant) for the platform: course literature is structured and turned into a dedicated knowledge base connected to a large language model, supporting course Q&A, assisted reading and learning-path guidance. All corpus outputs are formally admitted after review by a national comprehensive social science research institution, ensuring academic rigor.

Concept illustration of a course knowledge tree growing from books

Results

Q&A response time for students and faculty dropped from next-day to seconds, and teachers’ repetitive Q&A workload fell by 70%. Literature search and reading efficiency improved roughly threefold, turning the course corpus into real teaching capacity.

Concept illustration of an efficient smart classroom
SecondsQ&A response time
70%Lower repetitive Q&A workload
3xFaster literature search and reading

Note: the figures above are illustrative; final publication uses project acceptance data confirmed by both parties.

Frequently Asked Questions

企业智能体与传统自动化脚本有什么区别?

传统脚本依赖固定规则,企业智能体则能理解需求、查找资料、判断规则并调用系统,在异常时交回人工,执行过程可追踪、可校验。

智能体试点应该选择什么样的任务?

优先选择输入清晰、结果可检查的工作,例如汇总Purchase Requirements、生成Ticket Classification建议或整理Sales Follow-up记录;任务越具体,越容易判断智能体是否真正改善了工作。

规模化应用智能体需要哪些治理能力?

As agents spread across departments, permissions, quality, cost and audit need to enter one unified management mechanism,确保质量、权限与异常处理都经过验证后再复制到更多场景。

Find Your Best Opportunity forAI Automation
Request an AI Assessment