Security & Privacy Practices for Elements AI‑Powered Features

Last updated: July 10, 2025

Elements incorporates artificial‑intelligence (AI) components to help advisors surface insights faster and automate routine analysis. The safeguards below ensure these capabilities meet the same security and confidentiality standards as the rest of the Elements platform.

---

1. What data are the models trained on?

• No in‑house fine‑tuning. Elements does not fine‑tune its own large‑language models (LLMs) at this time.
• Pre‑trained only. All models in use were pre‑trained on publicly available and/or properly licensed datasets.
• Zero customer training data. No customer data is ever used to create or improve those underlying models.

2 . How your data is handled during an AI request

De‑identification

• Before any data leaves Elements, we strip or hash all personally identifiable information (PII)—names, birthdates, email addresses, phone numbers, account numbers, etc.
• Only numeric scorecard values and other non‑PII context required for the calculation are included in the request payload.

Model invocation

• We call closed or enterprise‑tier models that contractually prohibit training on prompt data and prevent it from appearing in other users’ sessions.
• Open, consumer‑grade endpoints are never used.

Response handling

• Returned results are stored and displayed only inside Elements.
• They inherit the same retention schedule and access controls as any other data in Elements.

3.  Frequently Asked Questions

Q: Will my clients’ financial data be sent to a public model?
A: No. All requests go to vetted, closed‑model endpoints. Data is de‑identified before transmission, and the provider is contractually barred from re‑using it for training.

Q: Am I required to use the AI features?
A: No. Advisors can still use the assessment feature and populate information bythemselves without using the AI feature. 

Q: Does AI output constitute investment advice?
A: AI responses are intended as informational aids. Advisors remain fully responsible for final recommendations in line with their fiduciary duties.

Related to

• assessment
• security
• AI
• AI Assessment
• privacy