Overview

Slingshot AI helps users interact with specialized data sources through a guided chat experience. From the landing screen, users select a conversation focus, enter a question, and receive targeted responses, suggested follow-up prompts, and in some cases direct links to deeper dashboards or detail views.

ℹ️ This feature is currently in beta and responses may vary.

What users see first

Users start on the Slingshot AI welcome screen, which presents conversation focus areas:

Contextual Data (Seradata): Mission-specific and organizational context
Orbital Data (SGSN): Orbital state and behavior data
Anomaly Detection (Agatha AI): Detected anomalies and operational impact
Help Desk: Platform features, troubleshooting, and support
Space Education: Space concepts, terminology, and fundamentals

How to use the experience

Choose a conversation focus based on the type of answer needed.
Enter a natural-language question in the chat box.
Review the response for key data, tables, links, and suggested follow-up prompts.
Refine the question if the assistant asks for clarification or presents multiple matches.

Best practices for users

Be specific when possible.
Include a satellite name, NORAD ID, operator, or question type.
Use follow-up prompts to drill deeper instead of starting over.
Switch conversation focus if the question belongs to a different data source.

Conversation focus guide

Contextual Data (Seradata)

Use this when the user wants mission context, object identity, ownership, or reference information.

Good example prompts

“Give me information on AMOS-3.”
“Who operates this object?”
“Find mission background for NORAD 32794.”

Expected behavior If the name is ambiguous, the assistant may return multiple matches and ask the user to choose the correct object.

Orbital Data (SGSN)

Use this when the user wants orbit parameters, state vectors, tracking status, or orbital history.

Good example prompts

“Show me the latest data for AMOS-3.”
“What orbit is AMOS-3 in?”
“Is this object currently tracked?”

Expected behavior The assistant may ask for a specific object first, then return a structured overview including orbit regime, orbital elements, tracking status, and state vector data.

Anomaly Detection (Agatha)

Use this when the user wants recent detections, anomaly scores, or operational risk summaries.

Good example prompts

“Latest detections?”
“Which object has the highest anomaly score today?”
“What anomalies were detected last week?”

Expected behavior The assistant can return a summary table with spacecraft, NORAD ID, anomaly level, and score, plus follow-up prompts and a dashboard link.

Help Desk

Use this when the user needs product support, troubleshooting, or feature guidance.

Good example prompts

“How do I reset my filters?”
“Where can I find orbit details?”
“Why am I seeing multiple matches?”

Space Education

Use this when the user wants definitions, explanations, or background learning.

Good example prompts

“What is GEO?”
“What does eccentricity mean?”
“Explain true anomaly.”

Example user journeys

Example 1: Getting anomaly detections

User action: Select Anomaly Detection and ask, “Provide the latest detections?”

Expected Result:

The assistant returns a recent anomaly summary.
A table lists affected spacecraft, NORAD IDs, anomaly levels, and scores.
The user can click through to a dashboard or use suggested follow-up prompts.

Example 2: Resolving an ambiguous object name

User action: Select Contextual Data and ask, “Get me information on Amos-3.”

Expected Result:

The assistant detects multiple matches.
It presents a choice list with identifiers and alternate names.
The user selects the intended object before proceeding.

Example 3: Viewing orbital details after narrowing the object

User action: Select Orbital Data, ask for the latest data, then specify “Amos-3.”

Expected Result:

The assistant first asks for more detail.
After the user identifies the object, it returns orbital parameters and state vectors.
The user can open detailed orbit views or ask follow-up questions about maneuvers, conjunctions, or decay rate.

Example 4: Learning basic space terminology

User action: Select Space Education and ask, “Give me space terminology.”

Expected Result:

The assistant returns a beginner-friendly list of common terms.
Definitions are short, plain-language, and easy to scan.
The response helps new users understand terms they may encounter elsewhere in the platform.