Skip to main content
Home/AI Policy

AI Policy

Effective October 23, 2026

MILCONNECT AI uses artificial intelligence in a few specific places to help service members, veterans, and military families make sense of their benefits. This policy lists every one of them, what the AI sees, who runs it, and how to tell AI output apart from everything else.

TL;DR

  • AI writes some chat answers, analyzes documents when you ask it to, compares resumes to job postings, writes the optional Daily Brief, and narrates the audiobook.
  • MILCONNECT AI runs on a commercial AI provider under terms that prohibit using your content to train AI models.
  • AI never makes decisions about you. It does not decide eligibility or ratings, and it does not approve representatives. People and official registries do that.
  • Everything AI writes is labeled where you see it, and the chat assistant tells you it is not a person.
  • AI output is informational, not legal, financial, medical, or VA-claim advice. Always check with the VA or an accredited representative.

1. Where we use AI

These are all the places the Service uses AI today. If we add one, we will add it here first.

FeatureWhen AI runsWhat the AI sees
Chat assistantSigned-in users only: when your question isn't covered by our written guides, or when you tap “Get a personalized AI answer.” Everyone else gets pre-written answers, with no AI involved.Your message, the recent conversation, and the profile details you've saved (such as branch, status, and rated conditions) so the answer fits you.
Document analysis (Record Keeper, Import)Only when you click to analyze a specific document.The contents of that document.
Resume comparisonOnly when you run a comparison.The resume and job posting you provide.
Daily BriefOnly after you turn it on. It writes one short summary a day, shown on your dashboard and optionally emailed to you.Items already in your account: deadlines, to-dos, notifications, new documents, unread messages, and saved benefits. If you connect Google, it also sees upcoming calendar and email items.
Audiobook narrationRecorded once, before publishing, with an open-source AI voice on our own computers.Only the book's text. No visitor data.
Internal team reportsWeekly and monthly reports for our own team.Site-wide totals and trends only. No individual's documents or messages.

2. How the AI is run

  • MILCONNECT AI powers the chat assistant, document analysis, resume comparison, the Daily Brief, and our internal reports. It runs on a commercial AI provider, and under that provider's terms the content we send is not used to train AI models.
  • The audiobook voice is a separate, open-source text-to-speech model. We ran it ourselves to record the audio ahead of time, so no visitor data is ever sent to it.
  • Advanced option: if you add your own AI provider API key under Profile → AI Settings, your AI requests run on your own account instead of ours.

The categories of companies that handle your data are listed in our Privacy Policy.

3. What AI does not do here

  • It makes no decisions about you. Benefit eligibility, disability ratings, and approval of representative accounts are never decided by AI. Representatives are verified by our staff and against the VA's official accreditation records.
  • It does not profile you for advertising. We run no advertising.
  • It does no facial recognition, emotion detection, or biometric identification.
  • It creates no imitations of real people's voices or likenesses.
  • It does not read your documents unless you ask it to analyze one.

4. Accuracy and human oversight

AI can be confidently wrong. We give it our knowledge base and your context so its answers are relevant, instruct it never to make up benefit amounts, and tell it to say so when it doesn't know. It can still make mistakes.

  • AI output is informational. It is not legal, financial, medical, or VA-claim advice (see Terms, section 6).
  • Before you act, check amounts, deadlines, and eligibility with the VA or an accredited representative.
  • If an AI answer is wrong or harmful, tell us through the contact form so we can fix it.

5. How AI is labeled

  • Chat: before you type anything, the chat window tells you it is an automated assistant, not a person, and which answers come from AI.
  • Daily Brief: marked “Written by AI” on your dashboard and in the email.
  • Document analysis and resume comparison: results appear only after you start the analysis yourself, and they are presented as AI analysis.
  • Audiobook: the narrator is named in the player as an AI voice.
  • Any AI-generated images or video we publish are labeled as AI-generated.

6. Your choices

  • Every calculator and free tool works without AI. Using AI is up to you.
  • The Daily Brief is off until you turn it on. To turn it off again, contact us through the contact form.
  • You can delete saved chat conversations from the chat history panel.
  • You can delete any document, along with its AI analysis, from the Record Keeper.
  • To export or delete everything, see Your rights in the Privacy Policy.

7. What we keep

  • Chat conversations of signed-in users are saved to your account until you delete them.
  • Document analysis results are stored with the document until you delete the document.
  • Daily Briefs are deleted automatically after 90 days.
  • Anonymous visitors' chats are not saved to any account.

8. Transparency laws

We tell you when you are dealing with AI, and we label AI-generated content. We built these disclosures around the transparency duties of the EU AI Act (Article 50) and of U.S. state laws on bots and AI disclosure, including California's. None of our AI features makes decisions about anyone's access to benefits, credit, employment, or essential services.

9. Changes to this policy

When we add, remove, or change an AI feature, we update this page and the effective date at the top.

10. Contact

Questions about how we use AI: privacy@milconnect.ai or the contact form.

Questions about this policy? Contact us at legal@milconnect.ai.

This page is part of the MILCONNECT AI service. See also our Privacy Policy, Terms of Service, and Contact page.