What this category's data looks like
This page addresses the training institution enrollment consultation conversion scenario. Materials include three categories: enrollment brochures, consultation records, and trial class schedules. Enrollment brochures come from the marketing and curriculum research departments, and contain course names, class hours, start dates, fee descriptions, and signature fields; consultation records come from consultants, and contain student names, intended courses, communication dates, and receipts; trial class schedules come from academic affairs, and contain trial class times, instructors, sign-ins, and confirmation signatures. The update cadence is: brochures are revised quarterly, consultation records are added one by one, and trial class schedules are adjusted weekly. When materials enter the vector model and indexing function, brochures and trial class schedules serve as retrievable knowledge chunks, while consultation records serve as raw data rows. Each generates vector identifiers and is written into index fields, providing supporting material for subsequent retrieval and citation trimming.
Configuration and implementation checks
This step is executed by the indexing service and vector database, and outputs vectors and index status (whether recall is possible, whether data in the same group is merged). The version ledger and change records are then completed by subsequent AI nodes, code nodes, or manual work (retrieval and citation trimming only take supporting material).
| Object | Check | Verification action |
|---|---|---|
| Enrollment brochure | Course names, class hours, start dates, fee descriptions, and signature fields must be consistent between the brochure and consultation records | Open the corresponding data row in the brochure, check whether the index field contains multiple vector identifiers, and confirm the status is indexed |
| Consultation record | Student names, intended courses, communication dates, and receipts must match across the two materials | Sample one consultation record, check whether the index field is complete, then run a retrieval to see whether data in the same group is merged |
| Trial class schedule | Trial class times, instructors, sign-ins, and confirmation signatures must be synchronized with the brochure and consultation records | Check the data range covered by the rebuild task records, and confirm that data already indexed after the model change is included |
| Index model | Must be consistent with the vector model actually used by the knowledge base, and the limit settings must match the citation budget | Check the index model in the knowledge base settings, and verify that the tokens and array count of the citation text are recorded separately |
Three common errors
- Symptom: The course name and start date for the same student do not match across the two materials. Inspect the following: After the brochure was revised, the consultation record was not updated in sync.
- Symptom: Duplicate entries appear in retrieval results, and data in the same group is not merged. Inspect the following: Not all multiple vector identifiers of one data item were written into the index field, so merging failed.
- Symptom: The trial class schedule lacks a signature or receipt, and the version number and citation are not synchronized. Inspect the following: The change record was not updated with the content, and the index still points to the old version chunk.
How to verify the setup
Feature acceptance: Sample one data row and inspect the index field to confirm that multiple vector identifiers are all present and the status is indexed; then run a retrieval to see whether data in the same group is merged and whether the score takes the highest value.
Business test: Change one piece of content and see whether the version number and citation are updated in sync; at the same time, verify that the tokens and array count of the citation text are recorded separately, and do not compare the two for equality.
- Check the index model in the knowledge base settings, confirm it is consistent with the vector model actually used by the knowledge base, and review the data range covered by the rebuild task records.
- Open a data row and check the index field to confirm that multiple vector identifiers are complete; if recall quality drops after a model change, check whether the rebuild task was triggered only for data that has already been indexed.