Private workflows stay private. Public chat stays bounded.
Kaizen is designed for daily AI use without flattening every workflow into the same risk profile. Public chat, private Max, SMS, phone-call alerts, reminders, voice, memory, dashboard operations, image generation, and automation each have their own trust boundary.
Example Split
Public Examples
These are appropriate for KaizenAI Open.
Current web questions
Ask for current weather, public news summaries, product research, or general research that benefits from current sources.
Technical explanations
Ask for a concept breakdown, code explanation, architecture comparison, or public documentation summary.
Project overview
Ask what Project Kaizen is, what KaizenAI Open can do, or how the public lane differs from the private runtime.
Private Communications Examples
These are real Kaizen surfaces for authorized users, not features exposed to anonymous public chat.
Text Max
Allowlisted contacts can text the Kaizen number and receive short Max SMS replies with memory scoped to their own profile.
Relay a message
Known opted-in contacts can ask Max to relay a short message to another known contact without letting raw phone numbers become arbitrary targets.
Set a reminder
Max can create, list, and cancel reminders that deliver through approved channels such as SMS, email, or Pushover.
Get a call alert
Approved private workflows can use voice calls as an operational notification channel when a text or email is not enough.
Call Max
Live today: an approved user calls the Kaizen number and talks to Max in real time — local speech processing, the same profile boundary and logging as every private lane, and an operator kill switch.
Ask about the stack
Authorized users can ask Max "how is the stack?" in chat or on a call and get an answer built from live service probes — not a cached status page.
Professional Workflow Example — Schedule Intelligence
The project management decision tools are a private planning workspace where the assistant works like a junior planner who can never touch the baseline.
1. Ask the schedule
Plain-language questions — what is driving the finish date, where is the open logic, how healthy is the network — answered from live reads of a deterministic local working copy, with activity IDs cited and every number labelled by its source.
2. Propose, never apply
The AI can draft typed schedule changes and stage them for review, but the platform itself refuses approve, apply, and export from any agent. A human reviews the staged change set and makes the call.
3. Simulate before commit
Proposed phases are structurally simulated first — open ends, cycles, duplicates, quality metrics — with the honest caveat that true dates and float still come from the professional scheduling tool, not the AI.
Reliability Workflow Example
Beacon workflows are the public-safe way to describe private time-sensitive notification agents.
1. Check the source
A scheduled Beacon calls or checks an external source, preserves the source evidence, and records the run in a private ledger.
2. Classify locally
Kaizen transcribes the source with local speech-to-text, classifies the result, handles replay/date edge cases, and stores the exact decision facts.
3. Notify and verify
The system sends SMS/email/call alerts, verifies downstream delivery receipts where available, escalates failures, and keeps history available to authorized Max users.
Private Examples
These remain on the protected side and are not available to public chat users.
Personal memory
Private recall, user-specific preferences, and long-running personal context are reserved for authenticated owner sessions.
Device or home operations
Automation controls, security-sensitive actions, and infrastructure operations are blocked from the public lane.
Account access
Private dashboard, iOS access, beta access, API access, and account requests go through announcements or direct approval.
SMS and notifications
Phone numbers, call routing, delivery logs, contact relay, reminder targets, and notification history stay behind authenticated user and operator boundaries.
Beacon details
Private source recordings, transcripts, decision history, run ledgers, and delivery internals are not exposed through KaizenAI Open.
Project schedules
Schedule data lives in a local, loopback-only workspace. It is never public, never leaves the private boundary, and no schedule content appears on this site.
Access Questions
If you want early access, iOS access, private access, or higher public limits, wait for public announcements on INTRAC.NET or email [email protected]. KaizenAI Open should not promise access or timelines inside chat.
