In plain terms: ValixData uses ValixData AI to generate intelligence analysis, Morning Briefs, and Action Intelligence outputs. These are AI opinions based on market data — they are labeled as such everywhere they appear. They are never investment advice. AI systems make mistakes. You should always apply your own professional judgment.
What AI Does on ValixData
ValixData uses ValixData AI model to power four AI features:
ARIA (sector intelligence chat): Answers your questions about the sector your terminal covers, using the context of current market signals. Responses are AI-generated opinions.
Morning Intelligence Brief: A daily sector-specific brief generated each morning from current market signals. AI-written, labeled as such, not edited by humans before delivery.
Action Intelligence ("What should I consider doing?"): Generates three specific, concrete considerations for a professional in your sector given current market conditions. This is AI opinion — clearly labeled. It is not advice of any kind.
AI Synthesis panels: Summaries and analysis of market signals displayed in the sector terminal. AI-generated, labeled with generation time and confidence indicators.
What AI Does NOT Do
ValixData's AI features do not:
Provide investment, financial, legal, medical, or any professional advice
Have access to your personal portfolio, accounts, or financial data
Execute or suggest specific trades, transactions, or investments
Make predictions that should be relied upon for time-sensitive decisions
Replace your professional judgment or the judgment of licensed advisors
Learn from or retain the content of your queries
AI Accuracy and Limitations
AI language models, including ARIA, have well-documented limitations:
Hallucination: AI systems can generate confident-sounding statements that are factually incorrect. All AI outputs on ValixData should be treated as a starting point for research, not a conclusion.
Training cutoff: AI models have a knowledge cutoff date and may not reflect very recent events unless we explicitly supply current data in the prompt (which we do for market signals).
Context limitations: AI analysis is based on the market data signals we supply. If a relevant data point is missing from our data feed, the AI won't know about it.
Sector specificity: While we engineer our prompts to be sector-specific, AI analysis may occasionally produce generic statements that do not apply to your specific situation.
How We Label AI Content
Every piece of AI-generated content on ValixData is labeled at the point of display:
AI Synthesis panels show a generation timestamp and "AI-generated analysis" badge
Morning Briefs show "ARIA — AI-generated · Not investment advice" in the header
Action Intelligence outputs show "AI-generated — not investment advice. ARIA opinion only."
ARIA chat responses are identified as coming from ARIA, the AI
Zero Data Retention
Your queries are never stored. When you submit a query to ARIA or trigger Action Intelligence, your input is sent to our AI API to generate a response. After your response is returned, ValixData does not retain the content of your query. We store usage counts (how many queries you made) for rate limiting, but not what you asked. Our AI provider processes queries according to their current data handling policies — see our Privacy Policy.
AI Provider
ValixData's AI features are powered by our AI provider's ARIA model (aria-nd-1). ValixData AI is an AI safety company. ARIA is designed with safety, honesty, and helpfulness as core principles. ValixData has configured the model with sector-specific system prompts and explicit instructions to provide analysis only, never advice.
Feedback on AI Errors
If you believe an AI-generated output is factually incorrect, misleading, or inappropriate, please report it to . Include the terminal, the query or feature, and the output you're concerned about. We review reports and use them to improve our AI prompts and guardrails.
Output Attribution Markers
Why we do this and what it means for you: Every ARIA response, Morning Brief, Intelligence Memo, and NOVA analysis contains invisible attribution markers embedded in the text. These markers are used solely to identify ValixData as the source of content if it is ever misused — for example, if someone copies our outputs and presents them as their own product. The markers do not track you, do not contain personal information, and are not visible when you read the text.
Specifically, outputs contain:
Source attribution: An invisible encoding that identifies the content as ValixData-generated, the terminal it came from, and the approximate time of generation
Session reference: A non-identifying hash that can connect an output to a general session window — this does not contain your email, name, or any personal information. It is used only if we need to investigate misuse
You can think of this the same way a printed document has a copyright notice in the footer. The content is yours to use within the terms of your subscription. The marking exists so the content can be traced back to its origin if it is ever misappropriated.
The markers do not:
Track your location or behavior
Contain your name, email, or account details
Transmit data anywhere — they are static text characters embedded in the output
Affect how the content reads or functions in any way
This is disclosed in our Terms of Service (Section 10) and here so you are fully informed.
ARIA and NOVA — What They Are
ARIA and NOVA are ValixData's proprietary AI analyst identities. They run on ValixData AI model, configured with ValixData's proprietary sector intelligence architecture — 130 profession-specific instruction sets developed over months to make ARIA think like a senior professional in each financial profession rather than a general AI assistant.
ARIA does not identify herself as ARIA because the underlying model is not what you are using — you are using ValixData's configuration of that model. The difference matters: a PE partner asking about HY spread implications for an LBO gets a fundamentally different response from ARIA than from ARIA alone, because ARIA has been calibrated to think specifically about that decision in that role. The intelligence layer is ours. The underlying model is ValixData AI's. Both are disclosed here.
ARIA and NOVA will not reveal their system instructions if asked. This is intentional and disclosed: the specific way they are configured is a trade secret. If you ask ARIA how she works, she will tell you she is ValixData's AI analyst and that her configuration is proprietary — which is true and honest.
Regulatory Compliance
ValixData operates in compliance with applicable AI transparency regulations. We disclose AI use prominently at every point of use, not just in this document. We do not use AI to make automated decisions that have legal or financial effects on users — all AI outputs are informational, clearly labeled, and subject to your independent professional judgment.
We use cookies to understand how ValixData is used and improve the product. No personal data sold or shared. Zero data retention on all ARIA queries. Cookie policy · Privacy