Training follows workforce needs
China’s 2025 vocational skills training guidance prioritises demand-led provision and better training and evaluation. It indicates a policy direction; it does not establish a software budget for individual providers.
AI Learning & Assessment Platforms
Build a learning experience around your curriculum, question bank and teaching model. Connect diagnostic quizzes, focused practice and instructor follow-up for adult learners preparing for professional exams.
Industry context and opportunity
Professional training providers already have courses, question banks and instructors. The opportunity is to connect those assets into a learning journey with clear next steps and accountable support.
Assessment tools are established
ExamStar documents daily practice tasks and reporting. Moodle supports question banks, quizzes and conditional access. A quiz engine or score dashboard alone offers limited differentiation.
Personalisation crosses roles
Area9 combines learner pathways with educator assignments and analytics. Our proposed focus is the connection between an institution’s teaching rules, response evidence and instructor action.
The opportunity to test: one subject, with answers connected to an explained recommendation, instructor approval and fresh-question reassessment. A custom build is justified only where configuration or integration falls short.
Users, needs and adoption barriers
A learner’s reason to use the product differs from an instructor’s workload or a provider’s reason to invest. Each needs its own evidence.
Adult exam learner
Know what to study next
- The current situation
- Studies through courses, question banks and group messaging around work and personal commitments.
- What adoption depends on
- Needs a useful next step, an explanation for the assignment and continuity across practice sessions.
One service. Distinct responsibilities.
LearnerRespond & practise
InstructorApprove & follow up
Content teamReview & version
Provider & ITScope & authorise
Learning evidence becomes an instructor’s next action.
Validate whether learners act on recommendations, whether instructors can complete a follow-up from the evidence, and whether the current LMS can already support the same tasks.
Established products in depth
Three useful reference routes: an institutional practice platform, an adaptive learning suite and an extensible LMS. Compare their journeys before counting features.
EXInstitutional practice and assessment
ExamStar
01Select content and mode02Set availability03Learner practises04Review and exportExplore the product analysis
Institutional practice and assessment
ExamStar
- Positioning and users
- For organisations administering learning and tests, with administration and learner experiences.
- Core capabilities
- Free practice and daily task modes, with response analysis and exports.
Task journey
- 01Select content and mode
- 02Set availability
- 03Learner practises
- 04Review and export
- Interaction model
- The official guide shows staged setup; open practice and daily tasks support different study rhythms.
- Commercial model
- An institutional procurement route. Confirm package, API and deployment terms directly.
- Fit
- A relevant baseline for standard practice assignment and reporting.
- Evaluate before adopting
- Test plan approval, version traceability and LMS connections against your requirements rather than inferring a gap from feature names.
- Design lesson
- Borrow the connection between practice mode, availability and reporting.
A9Adaptive learning suite
Area9 Rhapsode
01Organise content02Enter a learning path03Continue learning04Educator follows upExplore the product analysis
Adaptive learning suite
Area9 Rhapsode
- Positioning and users
- Learner serves employees and students; Educator serves managers and instructors.
- Core capabilities
- Individual learning paths, assignment management, analytics and risk alerts, with browser and mobile learner access.
Task journey
- 01Organise content
- 02Enter a learning path
- 03Continue learning
- 04Educator follows up
- Interaction model
- Learner and educator workspaces organise distinct tasks around the same learning process.
- Commercial model
- Enterprise consultation; confirm licensing, content adaptation and integration terms.
- Fit
- A reference for connecting learner pathways to educator analytics.
- Evaluate before adopting
- Evaluate content modelling work and the fit with existing teaching rules. Overseas product material does not establish local purchasing demand.
- Design lesson
- Design pathways and instructor actions together, including the cost of preparing content.
MOpen-source learning management
Moodle LMS
01Create course and bank02Configure a quiz03Answer and review04Unlock further activitiesExplore the product analysis
Open-source learning management
Moodle LMS
- Positioning and users
- A course-centred environment for educators, administrators and learners.
- Core capabilities
- Reusable questions, quizzes, feedback, attempt records and access conditions based on grades or completion.
Task journey
- 01Create course and bank
- 02Configure a quiz
- 03Answer and review
- 04Unlock further activities
- Interaction model
- Courses organise the journey; reusable questions remain separate from individual quizzes and their attempts.
- Commercial model
- Open-source self-hosting, MoodleCloud or a service provider. Include hosting, plugins and maintenance in total cost.
- Fit
- A strong configuration and integration baseline for providers already using an LMS.
- Evaluate before adopting
- Conditional access is not a validated mastery model. Evaluate any specialised recommendation or intervention workflow separately.
- Design lesson
- Keep question banks, quizzes and courses distinct, and test configuration before commissioning development.
Comparison and differentiation
Assessment, pathways and analytics already have established implementations. Our proposed differentiation is how a provider combines them with its own teaching service; the value still needs a real pilot.
SIDE-BY-SIDE
Compare the fit. Define the difference.
| Dimension | ExamStar | Area9 Rhapsode | Moodle | Proposed approach |
|---|---|---|---|---|
| Product centre | Institutional practice and tests | Adaptive learning | Courses and activities | One-subject support journey |
| Key experience | Assign and practise | Personal paths and educator analytics | Course-based quizzes and feedback | Explained tasks, approval and reviewable reassessment |
| Pathway mechanism | Daily task mode | Personal learning pathways | Conditional activity access | Teaching rules first; evaluate additional algorithms |
| Delivery model | Institutional terms to confirm | Enterprise licensing and implementation to confirm | Self-hosted, cloud or partner | Scoped build/integration; operating services agreed separately |
Provider-owned content
Connect question versions, explanations and rules to historical answer evidence.
Instructor-led support
Turn plan approval, question escalation and reassessment into owned tasks.
An existing-system fit
Extend the current LMS where it makes sense, with clear record ownership.
These are proposed design choices, not claims that competitors lack these capabilities. Commission only the gaps that a same-task configuration and integration trial cannot meet.
Product positioning and value
Turn your teaching model into an actionable learning service.
AI Personalised Learning & Assessment Platform
For professional exam providers with licensed content and an active teaching team. Connect diagnostics, plans, practice, reassessment and instructor support. The LearnFlow sample uses introductory data analysis to demonstrate the collaboration.
Your curriculum
EvidenceLearning tasksInstructor support
Design principles: interpret limited evidence cautiously, explain recommendations and retain instructor approval. The demo demonstrates a journey; a production pilot must establish educational and operating value.
- A clear next step
- Translate assessment evidence into tasks with reasons and goals.
- Continuous instructor context
- Keep identification, contact, reassessment and adjustment connected.
- Reviewable teaching decisions
- Version content and rules; keep objective scoring independent from generated explanations.
Product architecture
Start with one responsive web product. Learners can study on mobile browsers; instructors and content teams use desktop workspaces. Add a native app or another channel only when distribution needs justify it.
01Roles and entry points
Learner · assess / plan / practiseInstructor · approve / follow upContent team · review / version
02Shared business services
Assessment and scoringRules and study plansReassessment and support
03Assisted intelligence
Licensed-content retrievalGrounded explanationsSources and human escalation
04Data and platform services
Question versions / topicsAnswers / plans / interventionsIdentity / interfaces / audit
Understand the starting point
Look beyond the total score.
Run entry and progress assessments linked to your learning objectives. Show topic performance alongside question counts, with further assessment where evidence is limited. Give instructors a basis for action without assigning fixed ability labels.
- Entry and progress quizzes
- Topic results and evidence counts
- Answer history and question versions
- Further assessment and review
Workflow illustration
1Question version
2Learner response
3Topic evidence
4Next assessment
Outcome: a reviewable starting point
Core workflows and experience
Consider a learner who needs further percentage practice: the diagnostic informs a plan, the instructor approves it, practice produces evidence and a fresh-question reassessment informs the next decision.
EXPLORE THE WORKFLOW
Change a condition. See where the work goes.
Workflow illustration · no learner records or AI calls are created.
- AssessmentCurrent step
- Study planNext
- PracticeNext
- ReassessmentNext
STEP 1 · IN PROGRESS
Review the evidence before planning
Keep responses with their question versions and topic mappings. The evidence supports a draft, rather than a fixed ability label.
Each action advances the route above.
- 01
Assess
Establish the starting point
Keep questions, versions and answers together. Distinguish performance from the amount of evidence available.
- 02
Plan
Choose the next task
Use prerequisites, learning goals, available time and practice history to organise study.
- 03
Practise
Explain, then apply
Bring explanations and focused practice together, with a clear route to instructor support.
- 04
Review
Check with fresh questions
Retest the topic, retain both records and decide whether to revise the plan or arrange support.
Follow-up results + instructor feedback → the next study plan
Key interactions: explain the recommendation, keep saved drafts out of scores and distinguish repeated questions from fresh evidence. Route disputed content to reviewers while preserving historical versions.
From spotting a gap to closing the loop.
Reports explain what happened. A teaching workspace also shows who will respond, what they did and what happened afterwards.
- 01
Identify
Create a follow-up from incomplete tasks, errors or learner questions.
- 02
Contact
Review the answer history and record a support plan.
- 03
Reassess
Use fresh questions after targeted study to collect further evidence.
- 04
Adjust
Agree the next tasks and keep the intervention history.
Limited assessment evidence
Request further responses or instructor review before drawing conclusions.
Disputed question or explanation
Let content authors publish a revision while preserving historical versions.
Unsupported AI explanation
Offer approved material and instructor support. Keep scoring independent.
AI supports understanding. Your team owns the standard.
Your institution maintains content, answer keys and teaching rules. Keep sources and review points with assisted outputs, and let instructors retain responsibility for plans and follow-up.
AI capabilities and controls
Use rules for deterministic decisions and AI where language assistance adds value. People remain responsible for activating plans, publishing content and closing support tasks.
EXPLORE THE AI BOUNDARY
The answer needs a source. The exception needs an owner.
Routing illustration · switches show the planned workflow, not a live model response.
Use approved material to support understanding.
Retrieve the relevant explanation, draft help with its reference, and keep a route to instructor review.
Always independent: objective scoring, study plan approval and professional certification.
FROM CONTEXT TO A REVIEWABLE RESPONSESelect a step to explore its handoff
STAGE 01
Input
The current question, topic, licensed explanation and relevant learning context, limited to what the task needs.
- INPUT
- Current question + learning context
- OUTPUT
- Only the context this task needs
Rules
In this stepAnswer-key scoring, task completion, evidence counts and state transitions need no language model.
Assistance
Explanations, error-review summaries and instructor drafts; assess correction effort, time and cost on licensed samples.
Human decisions
Approve plans, resolve disputes, publish revisions and close interventions. Autonomous agents are unnecessary for the first release.
The demo uses preset AI replies. Production models and retrieval require separate correctness, latency, cost and fallback evaluation.
Technical architecture and integration
Define boundaries around existing systems and record ownership. Production needs server-side persistence, permissions and recovery; browser storage in the demo exists to explore the journey.
Access and permissions
Responsive UI → identity and permissions API
Learners see their own records; instructors have cohort access; content roles have publication rights. Enforce tenant boundaries on the server.
Business and storage
Learning services → transactional database / file storage
Link question versions, attempt batches, plans and interventions. Handle duplicate submissions and audit significant state changes.
Models and content
Controlled retrieval → model gateway → output checks
Keep credentials server-side and apply usage limits. Model failure must not interrupt scoring or approved tasks.
Integration and operations
Identity / LMS / content → API or reviewed import
Confirm fields, rights and record ownership. Test backups, restoration, monitoring and release rollback before launch.
Record relationships
Question version → response → evidence → plan version → task → reassessment / intervention
Choose a managed or dedicated environment around the customer’s infrastructure and agreement. Confirm interfaces, retention, export access and support responsibilities before the pilot. Production identity, third-party APIs and models are not connected to this demo.
Where the platform fits
Begin with adult exam theory, then extend by question format, assessment method and support model. Each setting needs its own content and acceptance tests.
Professional qualification training
Assign topic practice after a diagnostic, follow up on repeated errors and review fresh-question performance before an exam.
- Provider type
- Providers with licensed syllabuses and question banks
- Expansion requirements
- First-release priority: one subject, with no guarantee of exam success.
THE WORK IN CONTEXT
- 01Entry diagnostic
- 02Topic practice
- 03Instructor follow-up
- 04Fresh-question review
Business value and buying decisions
Judge the investment first by whether the teaching service can run clearly, then by measurable operating outcomes. The institution owns the content and service relationship; software connects the work.
Observable mechanism
The demo links answers, tasks, reassessment and intervention records. This demonstrates a workflow, not an educational outcome.
Value hypothesis to test
Shared context may reduce repeated preparation and missed handovers. Compare handling time, omissions and instructor workload with current tools.
Outcomes requiring evidence
Renewal, learner experience, exam performance and acquisition effects require actual data. No ROI, revenue uplift or savings percentage has been established.
- The provider’s service model
- Include support in a course or offer it as an additional study service. Validate pricing and willingness to pay with actual customers.
- Investment trigger
- Evaluate a build when existing tools cannot support a distinctive teaching journey, LMS connection or content ownership requirement.
- Total cost
- Separate content preparation, development/integration, training, hosting, inference, maintenance and migration. Agree project and recurring charges separately.
When does a custom build make sense?
- Standard quizzes, practice and score exports
- Compare established assessment SaaS first.
- An existing LMS needs a better study pathway
- Assess configuration and integration options.
- Proprietary teaching rules and instructor services
- Evaluate a custom journey around your content and workflow.
Implementation and delivery
Complete one subject and one end-to-end journey before expanding. Scope, schedule and cost depend on content, interfaces and evaluation samples.
First release: one adult exam subject, reviewed questions, topic mappings, diagnostics, instructor-approved plans, practice, fresh-question reassessment, support records, content versions and basic permissions. Start with approved explanations; connect a model after sample evaluation.
- 01
Map content and teaching
Confirm content rights, topic structure, explanation quality and the learner-to-instructor journey.
- 02
Define the rules
Specify assessment evidence, task assignment, approval and exceptions. Check the rules against sample scenarios.
- 03
Integrate and pilot
Choose mobile web, an existing app or another suitable channel. Validate identity, learning records and assisted explanations.
- 04
Review the evidence
Use task completion, intervention outcomes, follow-up results and operating costs to guide the next release.
Product and implementation
Agreed workflows, roles, data definitions, interfaces and business services; source and deployment configuration as contracted.
Content and connections
Import specification, content review and version rules, field mappings and integration test records.
Acceptance and operations
Workflow and permission tests, AI sample evaluation, user guidance, training, backup and monitoring documentation.
Accept the work against real tasks.
- Reproduce scores from the same responses; retain insufficient-evidence prompts.
- Require instructor approval and reconcile task completion with response batches.
- Link fresh-question reassessment to support records and retain historical question versions.
- Test role and tenant access on the server, exports, audit and restoration.
- Evaluate sources, error types, human correction, latency and per-request costs against thresholds agreed before the pilot.
Dependencies: licensed material, curriculum reviewers, a pilot cohort, identity/LMS access and model service terms. Reduce the pilot scope when a critical dependency is unavailable.
Review content revisions, task completion, intervention closure, reassessment evidence and service cost. Reassess content, permissions and models before adding subjects, subjective tasks or channels.
Product and industry sources
Sources reviewed: 3 November 2026
Start with your teaching service. A question bank may be enough for quizzes, practice and score exports. A broader platform becomes relevant when assessment, task planning, reassessment and instructor support need to operate together. Existing systems may only need configuration or integration.
Professional exam workflows can be built around the relevant syllabus and question formats. Practical competence requires additional evidence, such as work samples, observation and assessor review. Essays, speaking and other subjective tasks need their own scoring and review design.
Confirm rights to questions and explanations, topic mappings, teaching objectives, scoring and assignment rules, instructor processes and available integrations. One subject and a reviewed sample set are enough to begin scoping a pilot.
Objective questions use defined answer keys. Recommendations consider response evidence and teaching rules, while AI assists with explanations and summaries. Limited evidence requires further assessment or review. The platform does not certify professional competence or guarantee exam outcomes.
The demo uses original sample questions and fictional learners to show assessment, plans, practice, reassessment, instructor follow-up and content revision. AI replies are preset simulations and data stays in the current browser. Production identity, licensed content, models and integrations are scoped separately.
Assess the identity, course, question and response interfaces first. Keep ownership of core records clear and retain the responsibilities of existing systems. A reviewed import-and-export workflow can support an initial pilot where interfaces are unavailable.
Let's Talk
Build around the way you teach.
Tell us about your learners, content and instructor service. Together we’ll define the learning journey, integrations and acceptance criteria.
Business email
contact@boilingwater.cnOffice
10F, South Block, Kingkey Yujing Times, Longgang District, Shenzhen
Related solutions
More Wavesteam solutions
AI, capital-markets docs, OCR, vision, IoT and membership operations — composable for your industry.
Enterprise Knowledge Base RAG
Connect enterprise documents and business knowledge with source citations and access controls. Start with a model API, or deploy processing, retrieval, and inference on an appliance for fully internal use.
- AI
- RAG
- On-Premise
Capital Markets Doc AI
Connect terminology checks, version comparisons, human review and project handoffs for bilingual financial documents, with traceable decisions and review records.
- AI
- Capital Markets
- Automation
Handwritten Order OCR
Turn handwritten orders into structured records through field extraction, validation, and human review, ready for integration with your order management workflow.
- AI
- OCR
- ERP Integration