What WizField AI Does Today

Understand WizField AI Phases 0-4, including Business Brain V1, voice intake, Operator Copilot drafts, guarded send, and outcome observation.

WizField AI includes shipped Phases 0-4: AI foundation, Business Brain V1, voice intake, Operator Copilot SMS drafts, guarded human-confirmed SMS send, and read-only SMS outcome observation.

Understand the Shipped WizField AI Scope

WizField AI is a bounded product program inside WizField. Its shipped scope includes:

AI foundation with tool registry, bounded context, recommendation-run audit, and a noop model path
Business Brain V1 with a deterministic home brief on /home
voice intake foundation and a Telnyx live voice pilot
Operator Copilot SMS follow-up drafts on /calls
guarded human-confirmed customer SMS send from Copilot
read-only Copilot SMS outcome observation

Phase 5 and later product work, autonomous send, drip or sequence behavior, CRM mutation from outcomes, and self-learning claims are outside the shipped scope.

Keep WizField AI Scoped to the Active Organization

WizField AI uses the active organization from the authenticated request actor. AI context must not trust client-supplied organization IDs.

Recent calls, drafts, customer records, send behavior, and outcome hydration must remain organization-scoped. AI features must not create a shortcut around tenant isolation, membership validation, roles, or permissions.

Use Operator Copilot With Human Confirmation

Operator Copilot can generate SMS follow-up drafts for calls. Draft generation may use an LLM path when enabled, but the product can also fall back to template generation.

Sending is guarded. Copilot customer SMS send requires both calls.view and messaging.send, and the canonical send path goes through WizField messaging execution truth. AI does not bypass the product messaging service.

Read Outcomes Without Autonomous Follow-Up

WizField AI outcome tracking is read-only. Outcome observation derives reply status from messaging thread records after an outbound anchor message.

Outcome observation does not trigger auto-send, background send, drip campaigns, retry sequences, CRM mutation, causal attribution, or self-learning behavior.

Continue With Product Foundations

For Brain behavior, read How Business Brain V1 Works. For Copilot drafts, read How AI SMS Drafts Work. For explicit limits, read What WizField AI Does Not Automate.

For terminology, read AI Recommendation Run, Guarded Send, and Customer Reply Observation.

For the tenant model AI must respect, read Multi-Business Workspaces in WizField. For the broader product identity, read What Is WizField?.