What This Category of Data Looks Like
Daily documents for this category include attribution rules, lead distribution tables, channel reports, and exception checklists. Data sources cover advertising platforms, CRM systems, attribution SDKs, and form systems. Most data updates run on a daily or weekly schedule. Individual documents contain fields such as channel ID, lead ID, conversion time, and assigned sales ID. These fields correspond to business identifiers from different sources, with units typically being basic business units like pieces, times, or days.
Constraints on Citation Sources and Traceability
Multiple data sources require traceability links to associate business identifiers from advertising, CRM, SDK, and other systems. This prevents mismatches between traceability results and actual channels or leads. Daily or weekly update cycles mean traceability links must be refreshed regularly to point to the latest version of original documents. Detailed fields in documents (such as channel ID and lead ownership) require precise matching of corresponding entries when citing. Broadly referencing an entire document will lead to deviations in ROI statistics or sales attribution.
Configuration Setup
This configuration is based on FastGPT v4.17.0.
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
quoteMaxToken | 3000–13000 token | Differentiate between knowledge base search and workflow scenarios to match content length requirements for different scenarios |
| Single search limit | Top 5–10 entries | Adapt to the conventional entry scale of channel data, avoiding redundant information interfering with AI generation |
responseData | Include original document links, channel ID, lead ownership fields | Match the traceability requirements of multi-source data, ensuring that cited content can be directly linked to original business documents |
detail | Enabled | Used to display specific entry information of cited documents, ensuring complete traceability content |
| Maximum context | 8000–12000 characters | Adapt to the length after splicing multi-source data, ensuring complete transmission of business information required for traceability |
Three Common Configuration Mistakes
- Phenomenon: The citation content limit returned by the knowledge base search node in the workflow is only 3000 tokens, which does not match the 13000 token limit for the knowledge base search scenario. Cause:
quoteMaxTokenis not separately configured for the workflow scenario, and the general knowledge base configuration parameters are reused directly. - Phenomenon: The AI-generated response does not include direct access links to original documents, only displaying document summaries. Cause: The original document link field is not specified in the
responseDataconfiguration, and only the internally stored document identifier is extracted. - Phenomenon: The cited original document cannot be accessed normally after jumping, or the traffic effect is poor. Cause: The public network access link of the original document is not used, and only the path of the internal system is configured.
How to Verify Correct Configuration
- Enter the workflow configuration page of FastGPT v4.17.0, check the
quoteMaxTokenparameter of the knowledge base node, and confirm that the value meets the requirements of the workflow scenario. - Generate test attribution analysis results, check the citation area at the bottom of the response, and confirm that it includes direct links to original documents and corresponding business field information.
- Import multi-source test data, run the workflow, and check the traceability results to confirm that the cited fields such as channel ID and lead ownership are consistent with the original data.
- Check the system logs to confirm that the request parameters for citation traceability include the correct configuration items for
responseDataanddetail.
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.