Purpose of the AI Assistant
The HolonIQ AI assistant is a guided support tool designed to help users understand and navigate the QS Analytics Platform.
It does not function like a general-purpose AI or search engine. Instead, it operates within a defined set of sources and rules to ensure accuracy, consistency, and responsible use of data.
What the AI uses as sources
The AI assistant only draws from:
HolonIQ help and documentation hosted in the Intercom knowledge base
(https://help.holoniq.com/en/)QS support and product documentation
(https://support.qs.com/hc/en-gb)Approved internal product descriptions, definitions, and methodologies aligned to the four QS Analytics products
The AI does not:
Search the open internet
Read news articles, blogs, or external reports
Access university websites or third-party labour market sources
Pull data from live dashboards or raw databases
All responses are constrained to the documentation and knowledge explicitly provided to it.
What the AI is good at
The AI performs best when helping users:
Understand what data exists in the platform
Identify which product answers their question
Navigate to the correct dashboards, views, or reports
Understand definitions, metrics, and methodologies
Interpret trends at a high level (directional, not numeric)
Translate real-world questions into platform actions
Its strength is guidance and explanation, not data delivery.
What the AI cannot provide
The AI cannot:
Provide raw numbers, tables, or exports
Quote institution-specific or country-specific statistics
Give exact rankings, scores, headcounts, or growth rates
Reproduce charts or dashboards in chat
Estimate or infer missing data
Even if the information exists in the platform, the AI will redirect users to the appropriate location rather than restating the data.
How the AI should respond to data-specific questions
When users ask for specific figures, the AI should:
Acknowledge the user’s intent
Clearly explain that it cannot provide raw numbers
Point to the exact product, page, or workflow where the data lives
Offer to help interpret the data once the user has accessed it
Example bot response:
“I can’t provide exact figures in chat, but I can help you find them.
You’ll find this data in [Product] → [Dashboard / Report].
If you’d like, I can explain how to read or interpret what you see there.”
Limitations users should be aware of
The AI:
Is limited to the documentation provided in the HolonIQ and QS support knowledge bases
Does not know what access level or entitlements a user has
Cannot validate numbers taken from the platform
Cannot answer questions that fall outside the scope of the four products
If a question cannot be clearly mapped to one product, the AI should ask for clarification rather than combine information across products.
How to get the most value from the AI
The AI is most helpful when users ask:
“Where can I find…”
“How does this metric work…”
“Which product should I use for…”
“How should I interpret this trend…”
These questions align with the AI’s design and lead to clearer, more reliable answers.