ALTTRAX
|
Atlas
Your AI Command Center is active.
Active Capabilities
My Account
Monthly
$24,999/mo
Contract
12-Month
Renews
Mar 2027
Active Roles
9
📞
To add capabilities, contact your AiXGent account manager.
Agent Preferences
Account Information
ALTTRAX
Enterprise — 9 Active Roles
$24,999/mo
Mar 2027
Platform
AiXGent 2.0
All systems operational
🔒 Data Residency & Local Processing
Local Processing Mode
All AI inference runs on your approved local model. Zero data leaves your network.
PHI Firewall
Strips identifiable fields before any data reaches the AI layer. Member IDs passed as anonymous tokens.
Approved Local Models
Nemotron 3 Nano (31.6B)
On-premise · HC-approved · Primary inference model
llama3.2:3b
On-premise · Routing classifier · Fast intent detection
GPT-5 (cloud)
Cloud · Disabled in PHI mode
External API Allowlist
Approved external calls — all read-only, no PHI transmitted
| Service | Type | PHI | Access | Status |
|---|---|---|---|---|
| NPI Registry | Public API | None | Read-only | ✅ Allowed |
| Medi-Span | Local API | None | Read-only | ✅ Allowed |
| FDB (First Databank) | Local API | None | Read-only | ✅ Allowed |
| EHR Integration | External | Potential | Requires BAA | ⚠️ Not connected |
🔄 Projects
ALTTRAX active initiatives
✅ Completed
Shipped and done
🏢 ALTTRAX Partner Console — Powered by AiXGent™
Deploy, manage, and oversee AI agents across your client portfolio
My Agent Library
Agents available to deploy to your clients
My Clients
Active client deployments
| Client | Agents Deployed | Users | Status | Monthly | Budget | Actions |
|---|---|---|---|---|---|---|
| BlueCross Midwest | 3 agents | 8 users | 🟢 Active | $8,124 | 72% | |
| MidAmerica Health | 2 agents | 5 users | 🟢 Active | $5,416 | 45% | |
| PharmaChoice enterprise | 4 agents | 12 users | 🟡 Onboarding | $10,832 | 38% | |
| Apex PBM Solutions | 5 agents | 18 users | 🟢 Active | $14,250 | 85% |
Deploy Agents to New Client
✅ Deployment initiated. Client will receive access within 24 hours.
Budget Controls
Token budget and cost controls per client — client access suspended when limit reached
BlueCross Midwest — Monthly Budget
● ACTIVE
Token Limit
tokens / mo
Monthly Cost Estimate
$3.00
@ $0.003 / 1K tokens
Tokens Used — 720,000 / 1,000,000
72% used — approaching limit
Client access suspended when limit reached. Auto-replenish available.
▶ User-Level Budgets (3 users)
Sarah K.
78%
200K token limit
Mark T.
52%
150K token limit
Dana R.
31%
100K token limit
MidAmerica Health — Monthly Budget
● ACTIVE
Token Limit
tokens / mo
Monthly Cost Estimate
$2.25
@ $0.003 / 1K tokens
Tokens Used — 337,500 / 750,000
45% used — healthy
Client access suspended when limit reached. Auto-replenish available.
▶ User-Level Budgets (2 users)
James R.
55%
200K token limit
Lisa M.
33%
150K token limit
PharmaChoice enterprise — Monthly Budget
● ACTIVE
Token Limit
tokens / mo
Monthly Cost Estimate
$3.00
@ $0.003 / 1K tokens
Tokens Used — 380,000 / 1,000,000
38% used — healthy
Client access suspended when limit reached. Auto-replenish available.
▶ User-Level Budgets (3 users)
Tom B.
42%
200K token limit
Rachel S.
29%
150K token limit
David C.
18%
100K token limit
Apex PBM Solutions — Monthly Budget
⚠ 85% USED
Token Limit
tokens / mo
Monthly Cost Estimate
$3.00
@ $0.003 / 1K tokens
Tokens Used — 850,000 / 1,000,000
85% used — warning: approaching limit
Client access suspended when limit reached. Auto-replenish available.
▶ User-Level Budgets (3 users)
Priya N.
92%
250K token limit
Carlos V.
81%
200K token limit
Amy W.
61%
150K token limit
Oversight Dashboard
Monitor and intervene across all client agent activity
BlueCross Midwest
AI Strategy Expert
Analyzing Q2 formulary optimization opportunities...
MidAmerica Health
Compliance Officer
⚠️ Escalation: Client asked about off-label drug coverage policy — awaiting human review
PharmaChoice enterprise
Business Analyst
Completing requirements document for rebate reconciliation workflow...
Apex PBM Solutions
Claims Processing Agent · Provider Search Agent · Member Connections Agent
Processing 2,847 claims today · 1,204 provider searches · 891 member lookups — all within PHI firewall boundaries.
Revenue Summary
Your ALTTRAX partner earnings
Active Clients
4
Total MRR
$38,622
Your Revenue Share (25%)
$9,656/mo
Annual Run Rate
$115,866
Revenue share increases to 30% at 10+ active clients
💼 Use Cases & ROI Examples
Real client deployments — cost breakdown and what was built
Total Use Cases
3
Total Client Savings Generated
$2.58M+/yr
Average Client ROI
1,200%+
Partner Revenue / Mo
$1,613
✦ Angelic AI Design™
AI built with purpose, not just power. We connect to what clients already have — CRMs, ERPs, data warehouses, industry platforms, and enterprise systems — via secure, read-only APIs. Sensitive data stays on-premise. Zero rip-and-replace. Every deployment is accountable by architecture, not policy.
Local-First
Sensitive Data On-Premise
Accountable by Architecture
Zero Rip-and-Replace
⚙️ Platform Infrastructure
TIER 1 — LOCAL CLASSIFIER
llama3.2:3b · ~0.3s · $0
Routes every query before any API call. LOCAL vs. CLOUD in real time.
TIER 2 — LOCAL REASONING
Nemotron 3 Nano · 31.6B · On-Premise
Complex reasoning on client hardware. No data leaves the network.
TIER 3 — CLOUD API
Claude Sonnet 4.6 · PHI-free only
Strategy, analysis, complex reasoning. Only fires when Tier 1 routes it there.
🛡️ PHI FIREWALL — UPSTREAM OF ALL TIERS: Sensitive identifiers tokenized before any query is formed. AI layer never receives identifiable data. Cascading token chain: Platform → Client → Session. Orchestration layer handles agent personas, mode switching (Autopilot/Copilot/Mentor), audit logging, and escalation routing — all stateless at the AI boundary.
Ready to deploy for your client?
Configure agents, set token budgets, and go live in minutes.
🔐 Token-Based Contextual Access
Client-generated tokens. Zero PHI in the AI layer. Compliance by architecture.
System Architecture
Client System
Source of truth. Raw IDs stay here.
→
Token Layer
SHA-256 hash. Client-generated. AiXGent never sees raw ID.
→
Context Resolver
Auto-routes. Scores complexity 1–10. Pulls only needed tier.
→
Agent
Receives context package only. No PHI ever.
→
Response + Feedback
Output + human feedback loop → continuous skill improvement.
Live Token Demo
🔑 Token Generation — Try It Live
Enter any member ID below. Watch how it transforms into a safe token before reaching the AI layer.
Raw ID
▶
—
Token
▶
—
PHI to AI
▶
🔒 NONE
Context Tiers
Tier 1 — Default
- Age band
- Plan type
- Eligibility flags
- Region
- Language
Tier 2 — Clinical Context
- Condition categories
- Drug classes
- Adherence score
- Contraindication flags
- Prior auth history
Tier 3 — Behavioral & Predictive
- Adherence trends
- Refill patterns
- Engagement history
- Predicted risk scores
- Outcome history
Automatic Routing
Automatic MCP vs API Routing — Platform decides, not the human
Complexity score (1–10) drives routing in real time. Zero manual configuration.
| Request | Complexity | Route | Tier |
|---|---|---|---|
| "Is this member eligible?" | 2 | Lightweight API | Tier 1 |
| "What formulary tier?" | 3 | Standard API | Tier 1 |
| "Check drug interactions" | 6 | API + Context | Tier 2 |
| "Should this prior auth be approved?" | 8 | Full MCP | Tier 2–3 |
| "Personalize outreach for this member" | 9 | Full MCP | Tier 3 |
Learning & Feedback Loop
Continuous Improvement — PHI-Free Feedback
Agent Response
→
Human Feedback 👍 / 👎
→
Skill Update Queue
→
Review
→
Deploy
Learning happens at the skill level, not the individual level.
No PHI enters the feedback loop.
Corrections improve the model for ALL clients — every improvement is universally deployed.
🔒 What AiXGent NEVER Stores
Raw client identifiers · PHI of any kind · Full context packages · Mapping between tokens and real identities