How the Knowledge Base Works
The knowledge base combines selected published pages/posts and uploaded documents in one OpenAI vector store. Selection is local intent; synchronization is the operation that makes changes available to retrieval.
SmartSite converts selected WordPress content and uploaded documents into files that OpenAI can index in one vector store. During a question, the plugin can search that store for relevant passages and give them to the active model as grounded context. Local selection, remote upload, indexing, and runtime retrieval are separate stages, so each has its own status and possible failure.
First, select published wordpress pages or posts. The useful result comes later when you after changing or deselecting content, synchronize again, because configuration values matter only when their effect can be seen in retrieval, a generated reply, visitor access, or the connected service.
What this feature does and when to use it
Section titled “What this feature does and when to use it”Use this model to understand Current, Pending, Not synced, removal, and indexing states before editing sources.
Use this feature in the following situations:
- You are deciding which site material the assistant should be able to search.
- You need to explain why a selected page is not yet available to visitor answers.
- You are planning content ownership, privacy, updates, and removals for remote AI knowledge.
Where to find it
Section titled “Where to find it”Before you begin
Section titled “Before you begin”- SmartSite Assistant is installed and activated.
- You are signed in with an account that can manage WordPress options.
Set it up step by step
Section titled “Set it up step by step”- Select published WordPress pages or posts.
- Optionally enrich a page with metadata and save it.
- Optionally upload PDF, DOCX, PPTX, or TXT files.
- Run synchronization for selected pages.
- Wait for page and file states to settle.
- Test a question and review knowledge diagnostics in Analytics.
- After changing or deselecting content, synchronize again.
Fields, controls, and important values
Section titled “Fields, controls, and important values”The controls below describe the path from a WordPress source to information the AI can retrieve during a conversation. Selecting content only marks it for use; synchronization processes it and adds it to the shared vector store. Hashes and statuses keep remote content current. Until indexing finishes, the AI may answer without that source even though it appears in WordPress.
| Field, control, or status | What SmartSite Assistant does with it | How to use it and why it matters |
|---|---|---|
| Selected pages | IDs stored in WordPress; published pages and posts are offered as sources. | Selection tells the synchronization workflow which published sources are worth preparing for the assistant. It does not expose the page to retrieval immediately; after selecting it, synchronize and ask a question whose answer is clearly present on that page. |
| Content hash | Tracks processed content and metadata so changed sources become Pending. | The hash is how the plugin notices that the words or metadata have changed since the last successful upload. When it changes, the source becomes Pending so an old remote copy is not silently treated as current. |
| Vector store | OpenAI retrieval container created for the site and reused by pages and files. | Think of the vector store as the searchable library shared by the site’s assistants. Pages and documents must reach this library before OpenAI can retrieve them for an answer; creating more assistant profiles does not create separate libraries. |
| Sync | Creates or updates changed page files, leaves current files alone, and removes stored pages no longer selected. | “Sync” helps locate where the workflow currently stands. Resolve pending or failed items before relying on them, then verify the outcome in a new conversation because earlier context can hide whether fresh content is being used. |
| Uploaded documents | Uploaded locally, sent to OpenAI, attached to the same vector store, and indexed asynchronously. | A document becomes useful only after it is uploaded, attached to the vector store, and indexed. Until processing completes, a question may be answered without the file even though its filename already appears in WordPress. |
How to confirm it is working
Section titled “How to confirm it is working”After saving How the Knowledge Base Works, repeat the final test with clean context. That distinction matters because WordPress can store a valid value even when remote processing, access rules, caching, or delivery prevents it from influencing the real experience.
Practical example
Section titled “Practical example”A policy page becomes Pending after its content or saved metadata changes; synchronize again so retrieval uses the new version.
Recommended practice
Section titled “Recommended practice”- Change one part of How the Knowledge Base Works at a time and keep a short record of the previous value and test result.
- Judge the change with a fresh question that depends on this source; an admin status cannot show whether retrieval chose it.
Important warnings
Section titled “Important warnings”Common problems and focused checks
Section titled “Common problems and focused checks”| Problem | What to check and what to do next |
|---|---|
| How the Knowledge Base Works is missing or does not match this guide. | Confirm the plugin is active and the account can manage WordPress options. Check selection, synchronization, remote processing, and retrieval evidence in that order. |
| A change on How the Knowledge Base Works does not produce the expected result. | Keep the exact notice and test case, then review the browser console and WordPress/PHP log. Check selection, synchronization, remote processing, and retrieval evidence in that order. |
Screen reference
Section titled “Screen reference”- Capture
- Open Knowledge Base after a completed sync with summary cards, one synchronized page, and one completed file visible.
- Show
- Selected-page count, uploaded-file count, vector-store status, sync health, source states
- Viewport
- Desktop, 1440 × 900
- Annotate
- Use numbered callouts only for controls referenced in the procedure.
- Redact
- OpenAI keys, tokens, secrets, personal information, private URLs, IP addresses, and conversation text