Halley's AI Guide · Glossary

Contextual Data Access.

Definition — Contextual Data Access means an AI assistant or agent retrieves the records that matter for the person asking, such as their account, location, or history, and answers from those rather than from general knowledge. In Halley AI™, the assistant can reach only the sources it has been granted.

In short Right data, right moment

The assistant pulls the records that matter for this question (account, location, history) instead of answering from a generic model.

Accurate, personalized, timely

How It Works

Information tailored to the moment.

A general answer to "How many credits do I have?" is no use to the person asking. Contextual Data Access lets the assistant look up that person's own record at the moment they ask, and answer from it.

Key Features

Per-person retrieval: looks up the records tied to the person asking, not everyone's.
Recent data: reads records that are refreshed on a schedule or fetched when the question comes in.
Scoped sources: reaches only the systems and fields the assistant has been granted.
Grounded answers: responds from the retrieved record, so the answer matches what your system says.

Examples

An application dynamically querying healthcare-provider directories based on a patient's location, required specialty, and insurance acceptance.
An AI-powered support agent pulling a user's account details from a CRM upon receiving a service request.
A personal AI assistant proactively recommending nearby dealers tailored to the user's past preferences and recent searches.

Why It Matters

It is the difference between quoting the rules and answering the person. In the nurse advancement program Halley supports, the assistant connects to the hospital's learning-management and HR systems, so a nurse who asks how many credits they have toward the next clinical level gets their own current count.

Ground answers in your real context.

See contextual retrieval working on your own connected sources.

Try AI now Back to AI Guide