Analysis of System, Content and AI of an Informationized Corporate University-A case study of 100TAL.com (NYSE:XRS) a leading K12 tutorial school

Online learning is not a brand new topic. But, due to the special circumstances of 2020, inevitably, online-learning's pace of development and expansion has been accelerated.

Then, where should a company start with online-learning? And how does it work to achieve desired results?

Please follow me to take a look at what 100TAL.com (NYSE:XRS)(TAL hereafter) , a leading K12 after school tutorial education group, does on their own company training and development of talents.

As a technology and education company, from the design of management structure to the development of specific training products, the TAL training system has fully considered the needs of decentralization and online application since its initial stage.

In order to ensure the growth and development of employees, TAL has established a vertical and horizontal training system - horizontally, leadership training is its main focus and it creates a "three new systems" for new employees, new managers and new directors and improves the knowledge and behavior of employees at key stages. Vertically, it has built six professional colleges of academic products, Internet products, human resources, finance, technology and design to serve professional communities.

At TAL, the training function is decentralized into HRBP, teacher selection and training teams, and business functions, with TAL University only performs learning data management.

With block-chain technology as the core of the training system, TAL University has built a set of highly collaborative learning management system including live-streaming system, training record system, online learning platform and offline learning record system, as well as a productized basic content production mode.

Part I System Support: Create a highly collaborative learning management system

In terms of building online training systems, TAL faces two main challenges -

How to solve the problem of internal multi-platform coexistence？

How to solve the problem of turning offline training into online training？

Therefore, putting the block-chain technology platform at the center, TAL has built a highly adaptable learning solution system to solve these two problems at a low cost.

Proprietary platform for accessing online learning data

TAL’s training scenarios are very diverse, which has led to the coexistence of several different online learning tools internally. Some of these platforms are bought from external suppliers, while others are based on learning platforms or functional modules (live systems, new employee training systems, etc.) developed by the company's in-house technology.

Based on the above situations, TAL University and the block-chain team have jointly built a special data platform to access all online learning data without changing the actual training condition of the current business.

In order to get through the online data, the existing structured training data must be retained. In this process, how to digitize offline training scenarios becomes a challenge.

In order to solve the challenge, TAL created a training record template in the internal IM system, which is specifically used to record and deposit offline learning data. The training documents break down the training process according to the general training process and maximize coverage by allowing the trainer to complete the data recording while organizing the training. For the training organizer, the training documents are more like a note-taking list for the training organization.

At the same time, the block-chain system provides a front-end recording function that allows the instructor to evaluate the participants at any time during the training process. The combination of these two functions completes the offline training from the organization of the process to the data recording of the process.

Coordinating data operations, you need both light and heavy strategies.

At TAL, all internal training data including online training, offline training, and blended training will be collected. Based on this data, TAL University will further design two types of operation strategies: light and heavy.

Light Operation Strategy 

Light operational strategies are generally used in conjunction with empowerment-based courses. The goal of this type of course is to assist employees in self-growth and address real business needs, and it is suitable for learners with strong motivation to learn.

As a result, the data from this type of course is primarily used for continuous iterations of the course product itself - courses with low learning numbers are taken off the shelf and archived, and are no longer displayed.

For courses with moderate learning data but insufficient growth, in-depth user interviews are conducted with learning data filtering to understand the strengths and weaknesses of the courses, and course iterations are conducted.

Heavy Operation Strategy

It mainly serves learning courses that need to be mastered by all employees, such as newly released rules and regulations, corporate culture, and mandatory courses for promotion.

Compared with the light operation strategy, the heavy operation strategy will collect the course learning data and publish the data in a specific range, and show the completion status among various divisions and student learning exams directly to the corresponding department heads to promote further learning in the divisions.

It is worth mentioning that, under the heavy operation strategy, TAL University and specific training managers can also use the internal IM system developed by TAL to send learning alerts to remind relevant trainees to complete their studies on time.

Part of the learning data will flow directly to the talent data warehouse, where the system will tag the trainees with whether they have completed their learning task or not, and then be applied in conjunction with the company's internal promotion and performance regulations. This further promotes the learning and growth of employees.

Trusted storage supports learning outcomes management

Based on the untamperable nature of the block-chain, and based on the existing system, TAL University is further exploring students' self-operated learning model, i.e. abandoning process management and only managing results.

For example, online classes will be disassembled to form exams, no longer focusing on the learning process of the students, and only managing the exam data of the students.

In another example, when carrying out distributed learning task design, action learning, reading tasks and other related learning, the students themselves upload various photos documenting the learning process, enter the block-chain system, and the system will combine AI functions for basic discrimination, and certify the students to pass intelligent learning tasks.

In TAL University, all the learning data of employees are credibly stored through the block-chain. This ensures the quality of all data and forces all data generators (lecturers, students) to take the data they generate seriously.

And finally, all the learning outcome data is then connected to the group's talent data warehouse, generating talent learning data tags, and completing the closed-loop application of data.

With the support of the system, TAL University is able to manage learning through learning projects and learning data, achieving a lightweight operation and management.

Part II Content development, let good products speak for themselves

By digitization, TAL University has achieved an organizational structure that is as light as possible. But how do you ensure the quality of program execution? How do you drive a steady stream of good product content? This relates to TAL University's product design model.

Tapping into demand? Leave it to the TAL Training Alliance.

A good product will talk by itself and if you develop a product that meets the needs of users, the user will naturally follow the logic of the product. The starting point for achieving this is the mastery of training needs.

In TAL, there is an online training needs collection community - "TAL Internal Training Alliance". This community is based on TAL's internal IM tool "Zhingyinlou", which allows users to initiate a variety of synchronized training information methods such as live streaming, online meetings and document sharing. The community adopts an open-access group format to support anyone with training needs. In the community, everyone can put forward their own training needs, and group members will actively share relevant course resources and instructors, so as to achieve mutual benefit. In response to the high quality and frequency of these needs, TAL University will organize and focus its research and development efforts.

Similarly, the quality of these courses is also evaluated through data. The system will score the courses based on the learning data collected (number of viewers, positive ratings, and numbers of recommendations are the three main dimensions)

For courses with lower overall scores, TAL University will take them down. 

For courses where some details are problematic, iterative updates are performed, such as re-editing, re-recording and so on.

For courses that are highly rated overall, the group's award program are added directly, such as material and honorary titles.

Developing an online course with 3 ways to work together

The model of learning resources provided by frontline staff not only facilitates the iteration of courses, but promotes the production of new knowledge as well. However, in the face of the massive demand for online training, this model alone cannot fully meet the demand.

In order to fill the demand gap, TAL University has three other ways to improve content output in the enterprise: the self-preservation model, extraction technology promotion and live archiving.

Self-preservation Model

Online course modules are divided according to different professional groups, functional departments and business units. For example, TAL will set up professional modules on the homepage of the internal learning platform for all employees to learn, and regularly publish learning data and the satisfaction level of the training needs of their own ethnic groups, so as to inspire business units to produce more and better courses.

Extraction Technology

Combined with the sharing tradition of TAL, the effective experience of the business line is extracted into micro-sharing and micro-courses, which are disseminated on a small scale to form an effective knowledge output for a certain group of people.

Archiving Live Broadcast  

Applicable to a relatively large range of groups. Through TAL self-developed live broadcast system, anchors and viewers can realize strong interaction in the live broadcast process. This good experience has formed a user habit of using the live broadcast system for policy announcements and experience sharing within TAL.

By classifying and archiving a large amount of live content according to three categories: management, general and business, the online course library system of TAL is further enriched.

Through the combination of system, technology and offline operational actions, TAL University's online course resources reached 3,500 within two years and continued to grow at an accelerated pace. This has further contributed to the formation of an online learning atmosphere within the company.

Part III Driving a paradigm shift in learning with the assistance by AI

On the basis of achieving basic distributed and productized operations, TAL University is exploring the use of micro-learning to drive changes in the learning model within the company.

Micro-learning is a new learning model that places more emphasis on the combination of fragmentation and scenarios in the learning process. For example, in a work scenario, the system understands the learning needs of employees by retrieving their work plans, daily reports and other documents. When an employee mentions certain specific words, the system automatically pops up a mini-guided course for the employee to learn.

The advantage of this learning mode is that it fits well with the learning needs of adults. When an employee generates a task demand, a corresponding course is provided to meet the employee's learning needs in a timely basis. At the same time, by processing knowledge in fragments, employees are able to learn only specific modules of a particular type of knowledge to maximize learning efficiency.

However, to achieve this learning goal, two prerequisites need to be in place: the development of a basic online learning plan for employees, and a continuous content production model.

At TAL University, the foundation for both of these has been initially in place through the system. The next issues that need to be addressed are specific language associations, recommendations, and intelligent segmentation of courses based on the learning needs of learners, all of which need to be further supported by AI technology. This is also the next step in the system construction direction and goal of TAL University.