LIVE · FED MACRO
Loading —
AMP · Voice Advisor Alice Market Place  ·  Wall Street’s deals™. Main Street’s access™.
Master brand
Institutional architecture.
Individual access.
Sovereign Income
Structured income.
Senior security.
Continuum Growth
Growth equity.
Long-term discipline.
Alternatives
Real assets.
Counter-cyclical blend.
Venture
Frontier tech.
Asymmetric upside.
Voice Advisor · Layer 06

Talk to AMP. We’ll listen.

When a chat answer is not enough, speak to AMP’s AI advisor. Consultative only. OC-grounded. AI compliance pre-flight on every reply. Runs on the AMP voice rail, with enterprise-grade security and privacy.

TAP TO TALK → TRY CHAT INSTEAD →
02 · Activation

One tap. Speak naturally.

AMP Voice runs on the AMP voice rail — the production rail for institutional voice. Sub-300ms turn-taking. Adam baritone (institutional warmth). AI compliance pre-flight on every reply before you hear it. Zero data retention on Enterprise — your voice is never retained.

Ready when you are.
enterprise security
Audited security controls — Trust Services Criteria.
privacy-grade
PHI-safe processing — Business Associate Agreement available.
GDPR
EU data protection regulation compliant.
zero data retention
Voice audio not retained — transcript persists in your account for audit.
03 · Conversation starters

What you can ask AMP Voice.

Four common starting points. Tap any chip to pre-fill — then tap to talk. Every reply is grounded in the qualified Offering Circular and pre-flighted by the compliance layer before delivery.

04 · Architecture

The voice loop. Six stages.

The same production pattern AMP’s intelligence layer uses every day for institutional voice, ported to AMP for retail-grade compliance and consultative reach.

  1. 01

    You speak — AMP voice rail captures

    Capture begins from your browser microphone. Sub-300ms turn-taking. Real-time voice-activity detection picks up a natural end-of-thought without you having to push a button. Production-grade reliability.

  2. 02

    AMP voice routes to fast-response model supervisor

    A frontier-grade reasoning layer — AMP intelligence layer fast-response model — generates the consultative answer in under two seconds. Grounded only in the AMP knowledge base and the qualified Offering Circular. No hallucination on yield, terms, or eligibility.

  3. 03

    The compliance layer (deep-reasoning model) critiques before delivery

    Layer 07 of the AMP stack. The compliance layer is the second-model safety net — institutional-judgment grade. Every voice reply is scored against faithfulness, OC-binding, branding-wall, and no-yield-guarantee dimensions. If the draft drifts, the compliance layer redlines it before you hear it.

  4. 04

    AMP voice voices the approved response

    Institutional baritone. Distinct voice identity — Adam is the AMP brand voice across every channel. Natural cadence, no robotic delivery. Streaming audio output so you hear the reply as it forms, not after a long pause.

  5. 05

    Turn captured to vault for audit + continuity

    Every question, every reply, every compliance critique persists to your account audit log. Continuity across sessions — if you call back tomorrow, AMP remembers where you left off (with your permission).

  6. 06

    You are never sold. You are advised.

    AMP Voice is consultative only. The reply pattern is institution-tier disclosure — what AMP knows, what the OC discloses, what the risks are. Subscription happens in the Investment Portal, not in the voice loop.

Reference
This is the production pattern AMP’s intelligence layer uses every day. The two-stage AI safety harness — fast-response for speed, deep-reasoning for critique — is published AMP intelligence-layer engineering. AMP inherits it, retail-grade, for consultative reach.
05 · Boundaries

What AMP Voice will not do.

Institutional honesty is the operating principle. AMP Voice is a consultation tool, not a registered investment adviser, not a transaction surface, not a yield guarantor. The five boundaries below are hard limits — enforced by the compliance layer on every reply.

Boundary · 01
Will not give investment advice
AMP Voice is a consultative AI tool, not a registered investment adviser (RIA). It explains the AMP mechanism, the cohort architecture, and the qualified Offering Circular — it does not tell you to invest, what to invest, or how much. For RIA-grade advice, consult a registered adviser.
Boundary · 02
Will not execute transactions or move money
No subscription action, no payment, no fund transfer happens through the voice loop. Subscription is exclusively through the Investment Portal — broker-dealer infrastructure, FDIC-member escrow, audit-trailed. AMP Voice points you to the portal; it does not transact for you.
Boundary · 03
Will not recommend beyond OC disclosure
AMP Voice answers only what is documented in the qualified Offering Circular for the cohort under discussion. No off-OC promises. No “trust me, this one will do better.” If the OC does not disclose it, AMP Voice does not commit to it.
Boundary · 04
Will not fabricate cohort-specific economics
If a cohort has not yet qualified, AMP Voice will not guess at its yield, allocation mix, or lock-up. It will tell you what is public and what is pending. No marketing-grade extrapolation. The honest version every time.
Boundary · 05
Will not drift past compliance pre-flight
Every voice reply passes the compliance layer (deep-reasoning model) critique before delivery. The critique scores: faithfulness to source, OC-binding, branding-wall (no comparative claims), Mother’s Love Code Article III (no harm), no-yield-guarantee. If any score fails, the draft is regenerated or refused before you hear it. This is the second-model safety net — the same pattern AMP intelligence layer publishes for production-grade institutional voice.
06 · Layer 07 · the compliance layer

Every word passes the compliance layer first.

Every voice answer, every chat response, every email AMP sends — passes the compliance layer first. the compliance layer is the compliance critique layer, running the deep-reasoning model in a published three-stage harness (Planner → Generator → Evaluator). The compliance layer is what holds AMP Voice to a consultative standard rather than a chatbot standard.

Stage · 01 · Planner
What is the right answer?
the compliance layer reads your question, pulls the relevant Offering Circular sections, the cohort-specific facts, and the AMP doctrine. Builds a plan for what a compliant reply would say — and what it would not.
Stage · 02 · Generator
Draft the reply, grounded.
The fast-response model generates the consultative draft following the plan. Fast — under two seconds. Grounded in the OC and the knowledge base. No invention, no extrapolation.
Stage · 03 · Evaluator
Critique before delivery.
the compliance layer (deep-reasoning model) scores the draft on six dimensions. If any score fails, the draft is regenerated or refused. You hear the approved version, not the first draft.

The six critique dimensions

DimensionWhat the compliance layer scoresFail action
FaithfulnessDoes the reply match the source documents (OC, doctrine, knowledge base)?Regenerate with stricter grounding.
Branding wallDoes it stay within AMP brand language — no off-brand drift, no Alice-Group-vs-AMP confusion?Redline + regenerate.
Mother’s Love Code Article IIIDoes the reply protect the investor from harm — financial, reputational, emotional?Reject and ask for clarifying question.
OC-bindingIs every yield, term, allocation, or risk claim tied back to the qualified Offering Circular?Strip uncited claims, regenerate.
No comparative claimsDoes the reply avoid “same as sovereign”, “better than X”, “outperforms Y”?Reject — the compliance layer hard rule.
No yield guaranteesAre all yield references framed as targets per OC, not promises?Reject — the compliance layer hard rule.

An AI compliance pre-flight reviews every voice answer before it reaches you — retail-grade — so consultative voice never crosses into advisory territory.

AMP Doctrine · Layer 07
07 · FAQ

The honest answers up front.

Q · 01
Is this a registered investment adviser?
No. AMP Voice is an AI consultation tool — not an RIA. It explains the AMP mechanism and the qualified Offering Circular. It does not tell you what to do with your money. For RIA-grade advice, consult a registered adviser.
Q · 02
Is my voice recorded?
AMP voice: zero data retention on Enterprise. The voice audio itself is not retained. The transcript persists in your AMP account for audit purposes — you can view it, export it, or request deletion.
Q · 03
What languages does it support?
English at launch. Spanish and Arabic on the roadmap — to serve the AMP global investor base from day one and the GCC sovereign-tier audience natively. Other languages added as the cohort expands.
Q · 04
What if it gets something wrong?
Compliance critique catches most drift pre-delivery. If something still lands wrong, report it to amp@alicegroup.ai — investigated within 24 hours, response logged, OC and knowledge base updated if needed.
08 · Get started

Three ways to begin with AMP.

Voice is the new front door. Chat is the steady companion. The portal is the threshold to allocation. All three are consultative until you choose to subscribe — and every word, every reply, every email passes the compliance layer first.

TAP TO TALK → TRY CHAT INSTEAD → GET STARTED →

AMP Voice is consultative only. Subject to qualified Offering Circular per cohort. No yield is guaranteed. No transaction occurs through the voice loop. Subscription happens exclusively through the Investment Portal. AMP voice rail · enterprise security · privacy-grade · GDPR · zero data retention Enterprise tier.