AI Automated Customer Support MVP
Safe support guidance without production authority.
This MVP provides a maintainable knowledge base, suggested questions, a rules-based chat widget, escalation UI, and feedback flow while keeping all sensitive operations blocked.
Support boundary
The assistant can explain policy, but it cannot make decisions.
This page is explicit about what the assistant can and cannot do, so users do not mistake support guidance for a production approval or certificate action.
Allowed
What the assistant may answer
AIAA overview, Levels 1 to 5, application flow, exam flow, evidence submission, AI Assistance Declaration, payment explanation, common errors, community usage, and human support routing.
Blocked
What the assistant may not do
Approve or reject applications, issue or revoke certificates, mutate payment state, access unauthorized private data, reveal secrets, send real email, or guarantee certification outcomes.
Knowledge base
A maintainable support source exists separately from the chat UI.
Articles are stored in a dedicated knowledge file so the answer logic and the public support content can evolve without turning the page into a giant hardcoded component.
What is AIAA?
AIAA is the AI Agent Identity Authority. It connects certification, records, registry signals, and ranking visibility into a trust layer for the AI Agent economy.
- • AIAA is designed to verify AI Agent capability through evidence, review stages, and public trust signals.
- • It does not certify memorization. It focuses on engineering judgment, execution quality, reproducible evidence, and operational trust.
What is the difference between Level 1 and Level 5?
Level 1 focuses on operator capability, while higher levels increase expectations for engineering depth, architecture, company review, and council-level trust.
- • Level 1 centers on AI Agent operation, practical evidence, and clear execution proof.
- • Level 2 and Level 3 raise the bar for engineering, architecture, and production readiness.
- • Level 4 is a company-oriented review path, and Level 5 is council review rather than a normal automated progression.
How do I apply?
Create a member profile, submit an application with evidence, pass precheck, complete the exam, and then wait for human review decisions.
- • Applicants should prepare a GitHub repository, README or evidence summary, and any available demo or video before expecting review readiness.
- • Application approval, rejection, revision requests, and certificate decisions remain human-gated.
What is the AI Assistance Declaration?
Candidates may use AI tools, but must disclose what AI helped with, what they personally verified, and whether they can explain the final submission.
- • AIAA exams are AI assisted. The system does not ban AI tools such as ChatGPT, Claude, Codex, Cursor, or GitHub Copilot.
- • Candidates must disclose the tools used, what AI helped with, what they personally verified, and confirm they can explain the submission.
How do payment and certificate decisions work?
Payment and certificate issuance are not controlled by the public support assistant. They require human decision gates and policy checks.
- • The support assistant may explain the payment flow and certificate requirements, but it cannot mark payment as complete or issue a certificate.
- • Any case that appears ready for payment or certificate issuance must be escalated to a human owner.
How do I use the community forum safely?
Use the forum for discussion, learning, and public project sharing. Do not share secrets or make fake certification claims.
- • The forum supports public and member-only discussion areas with human moderation.
- • Do not share API keys, do not impersonate AIAA staff, and do not claim a certification result unless the data is verified.
Conversation UI
Suggested questions, answers, escalation, and feedback are all visible in one safe flow.
The current assistant is intentionally rules-based and local-safe. It does not require a production AI API key and does not modify any database state.
AI support chat widget
Rules-based MVP assistant
Welcome to the AIAA support MVP. I can answer policy and process questions, but I cannot approve, reject, issue certificates, mutate payment status, or access private data without authorization.
Sources: knowledge-base
Feedback UI
Feedback stays local in this MVP and does not create a production record.
Safety notice
- • No approval, rejection, payment mutation, or certificate issuance is available in this widget.
- • No private data is fetched here.
- • No real email is sent when escalation is requested.
- • Human escalation remains available as a concept UI without production side effects.
Future review
Admin review and safety event concepts are defined before backend wiring.
This MVP shows the operating concept for future safety queues, rate limits, and review surfaces without pretending those production systems already exist.
Safety event
Risky prompt review
Future support safety events should log ambiguous payment, approval, or certificate requests for human review.
Escalation queue
Human support handoff
A future queue can receive policy-sensitive or account-specific cases, but this MVP stops at UI and documentation.
Rate limiting
Abuse protection concept
A future server-side API should limit spam, repeated escalation requests, and high-risk prompt patterns.
Next Action
Pair support guidance with operations visibility and community discussion.
AIAA can answer common questions safely, route humans to the right next step, and keep sensitive decisions under explicit ownership.