What the data for this category looks like
IT asset inventory data comes from asset management systems, procurement systems, financial fixed asset modules, and terminal management platforms. Data is updated monthly, quarterly, or annually. Daily output includes asset ledgers, inventory sheets, loan and return forms, and scrap applications. Core fields include asset number, name, specification model, user, storage location, purchase date, and usage status. Units are typically physical units such as units, sets, and pieces.
What constraints these characteristics impose on knowledge base retrieval and recall
Multi-source and heterogeneous data sources mean the retrieval scope must be strictly limited to asset-related knowledge base collections. This avoids interference from cross-category data. Low-frequency, batch updates mean high-frequency real-time retrieval is unnecessary. But recall results must maintain timeliness. Fixed-field document structures require accurate matching of core fields such as asset number and user during retrieval. Association logic between fields must be retained. The core account and physical verification business scenario requires retrieval results to quickly locate all associated documents of the corresponding asset. This reduces manual screening time.
How to set configurations
The following configurations apply to FastGPT v4.17.0.
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
similarity | 0.75–0.85 | Matches precise fields of asset ledgers, filters low-relevance generic content |
limit | Top 10 entries | Covers the single query requirement of a single inventory task, avoids result overload that affects manual verification |
searchMode | vector+keyword | Balances exact matching of asset numbers and semantic recall of document content, adapts to multi-field retrieval needs |
usingReRank | Enabled | Performs secondary sorting on recall results, optimizes the display order of high-weight business fields such as user and storage location |
datasetSearchUsingExtensionQuery | Disabled | The data structure of asset ledgers is fixed, no need to extend queries to generate generic content |
collectionIds | Corresponding collection ID of the asset knowledge base | Limits the retrieval scope to ensure only asset-related document data is recalled |
Three common misconfigurations
- Recommended questions generated during retrieval do not match asset-related content in the knowledge base. The root cause is that
datasetSearchUsingExtensionQueryis not disabled. The system generates generic questions that do not draw from the knowledge base. - The system prompt indicates a specified model does not exist or cannot be accessed. The root cause is one of two issues: the available model for the corresponding asset knowledge base is not bound in the configuration, or model permissions are not correctly configured.
- Recall results include non-asset documents. The root cause is that
collectionIdsis not configured to limit the retrieval collection. Cross-category data is mixed into retrieval results.
How to verify correct configuration
- Submit a known asset number for retrieval. Check that returned results only include relevant documents such as the corresponding asset's ledger and loan records. No irrelevant content should appear.
- View the sorting logic of retrieval results. Confirm that results with high matching degrees for core business fields such as user and storage location are ranked first.
- Test the recommended question generation function. Confirm all returned questions come from asset inventory-related content in the knowledge base. No generic questions should appear.
- Check the model call logs. Confirm there are no error messages indicating the specified model does not exist or access failed.
The values provided are common starting points and should be measured against the reader's own samples.
Question material comes from public community discussions. Configuration values are common starting points and should be measured against your own samples. Verified on 2026-10-19.