AI Search — Frequently Asked Questions
This article compiles the questions uConnect admin users most commonly receive — from students, colleagues, faculty, and institutional administrators — about AI Search. Use it as a reference when questions come up.
About the Tool
- Can AI Search be disabled or turned off?
No. AI Search is a core feature of the uConnect platform going forward. It is not an optional add-on, and there is no toggle to revert to the previous keyword search experience. Once it is live on your VCC, it is the search experience for all users.
- Is AI Search connected to ChatGPT or the open internet?
No. AI Search operates entirely within your institution's VCC environment. It does not have access to the internet and does not use ChatGPT or any general-purpose public AI. Every result it returns comes from content your career center has published on the platform.
- Does AI Search learn from or train on our institution's data?
No. Student queries and interactions are not used to train any AI model. Results are personalized based on a logged-in student's profile attributes (role, graduation year, community affiliations), but this personalization is scoped to result ranking — it does not involve the AI learning from student data over time.
- How is AI Search different from a chatbot?
AI Search is a discovery tool, not a conversational assistant. It does not engage in open-ended dialogue, remember context across sessions (beyond search history), or provide general-knowledge answers. It is designed to do one thing well: help students find relevant content from your career center's content ecosystem. Think of it as a smart search engine, not a chat companion.
- What AI model powers AI Search?
AI Search is built in-house by uConnect using a technology called Retrieval-Augmented Generation (RAG). It is not a white-labeled version of any third-party AI product.
- Will AI Search be turned on automatically?
Yes. AI Search activates for all VCCs at General Availability — no configuration is required to turn it on. Your role is to prepare your content and communicate the change to students; the technology takes care of the rest.
About Content & Results
- Can AI Search read the contents of a PDF?
Not at launch. AI Search can surface PDFs based on their titles and descriptions, but it cannot read the text inside the document. Full-text PDF ingestion is planned as a Fast Follow on the roadmap. In the meantime, consider migrating important information from frequently used PDFs into native platform content.
- Do we need to add keywords or tags to our content?
Technically, no. AI Search uses semantic retrieval, which means it understands the meaning and context of your content rather than matching exact keywords. Your existing content does not need to be modified, retagged, or optimized in any way for AI Search to index and surface it effectively. However, tagging is still very valuable and may improve results in the future, so please make sure to always tag your content from a content management perspective!
- What happens if a student asks about something we don't have content for?
AI Search acknowledges that it couldn't find a relevant result and encourages the student to reach out to career services directly. It will not fabricate a response or surface unrelated content. See Article 12 for more detail on fallback behavior.
- Can AI Search surface content from other institutions?
No. Each institution's VCC is strictly isolated — your content is never accessible to other institutions, and your students will never see content from another school. This tenant isolation is a foundational security principle of how the platform is built.
- How soon after I publish content will it appear in AI Search?
For most content types (resources, blog posts, community pages), content typically appears in AI Search within approximately one minute of being published. Job postings may take longer depending on your integration's sync schedule.
- Is there a limit on how many resources AI Search returns in a response?
There is no hard cap, but in practice you'll typically see up to approximately 10 resources returned in a single response. AI Search is designed to return the most relevant results, not an exhaustive list — quality over quantity.
About Configuration
- Can we customize the suggested prompts?
At launch, you can select from a curated set of pre-populated prompts aligned to your enabled modules.
- Can we provide background information about our institution to shape AI Search responses?
Some foundational context can be added on the backend in partnership with your CSM. The ability for institutions to self-configure this directly is on the product roadmap. Talk to your CSM about what's possible for your institution right now.
- Can we hide certain content from appearing in AI Search results?
If you don't add it into your uConnect platform, it won't show up. But if it's published content, it will. An "Exclude from Search" feature is on the roadmap. In the meantime, SSO-protected content is automatically excluded from results for unauthenticated users, which provides some degree of control over what guests can discover.
About Data & Privacy
Please refer to our Privacy Policy: https://support.gouconnect.com/articles/uconnect-privacy-policy
- How does AI Search handle student data?
uConnect maintains SOC 2 compliance with rigorous governance frameworks and continuous monitoring. AI Search operates within your institution's existing authentication and permission boundaries. Student data is not shared across institutions and is never used to train AI models.
- How does SSO-protected content work with AI Search?
SSO-protected content is only visible to authenticated users with the appropriate access rights. If a student is not logged in, they will not see protected content in their AI Search results — the system always prefers returning no result over surfacing unauthorized content.
- Will we be able to see what students are searching for?
Not at launch. Aggregate analytics, search trend data, and query-level visibility are all on the analytics roadmap. When these features become available, they'll appear in the Analyze section of your admin dashboard.
- Will staff be able to see individual student search histories?
Not at launch. Individual student search history visibility for staff is on the roadmap. Students can already see and delete their own search history. If this is a priority use case for your team, share the context with your CSM so it can inform roadmap prioritization.
- How long is search history stored?
Search history is stored indefinitely. Students can delete their own history at any time. If a student deletes their account, their associated search history is also removed.