Bill-of-materials generator
DoesDrafts a BOM from the completed LLD, mapping every signal path to equipment and cost estimate.
Does notNever orders equipment. Never commits spend.
Intelligence Layer
The AI specialists behind the platform.
Every ShowOps.AI module ships with purpose-built AI specialists — each trained on one discipline, each operating under human approval. No one specialist does everything. No specialist mutates without a human. No black boxes.
Every specialist follows the same four rules
Propose, don't commitCite every claimStay inside your roleShow the signals
Domain 01
Infrastructure
InfrastructureSpecialists who validate the technical substrate every show is built on. They catch a multicast collision on paper, not in the truck compound.
Bandwidth planner
DoesValidates per-link capacity against SMPTE 2110 flows. Flags oversubscribed trunks before truck roll.
Does notNever reshapes the topology without human approval.
Signal-flow verifier
DoesTraces flows end-to-end for completeness. Flags orphan inputs, missing audio returns, untagged outputs.
Does notNever patches the flow itself.
Domain 02
Production
Produce · Staffing · DispatchSpecialists who keep schedules, crew, and vendor work-orders coordinated. They surface the overlap between the rigging window and the sound check before anyone else does.
Schedule optimizer
DoesProposes timeline adjustments when a conflict is detected — reflows load-in, rigging, and sound-check windows.
Does notNever shifts a locked milestone.
Conflict scanner
DoesCross-references departments, venues, schedules, crew. Surfaces overlap before it reaches the floor.
Does notNever cancels or reschedules without human approval.
Call-sheet generator
DoesBuilds per-department and per-individual call sheets from the live schedule, with distribution preview.
Does notDoes not send — a human hits distribute.
Domain 03
Rehearsal
RehearseSpecialists who turn a show structure into a running rehearsal programme. Auto-generate the schedule, then keep it honest as rooms and times shift.
Grid auto-generator
DoesBuilds the rehearsal schedule from the show structure — rooms, slots, speaker blocks.
Does notDoes not self-book speakers.
Risk scanner
DoesSurfaces scheduling conflicts, double-booked rooms, tight turnarounds before show day.
Does notNever reshuffles the grid without human approval.
Speaker nudge generator
DoesDrafts speaker-facing reminders and confirmations based on programme status.
Does notDoes not send — a human reviews and sends.
Domain 04
Commercial
Sales · Budget · SuppliersSpecialists who keep the money and pipeline visible — and never commit either without human sign-off.
Pipeline optimizer
DoesSurfaces stalled deals, stage bottlenecks, missing contacts. Every flag cites the underlying data.
Does notNever moves a deal stage.
Margin advisor
DoesProjects margin, variance risk, change-order stalls, vendor overruns. Always cites the line items driving the call.
Does notNever commits budget, signs SOWs, or approves change orders.
Vendor scorer
DoesRanks suppliers for a category by reliability signals — win rate, on-time delivery, dispute history. Exposes every signal behind the score.
Does notNever picks the vendor.
Report generator
DoesPulls section data from the live operation (budget actuals, production recap, crew deployment) and drafts the AI executive summary. Surfaces everything as a proposal in the queue — the human reviews, edits, and approves before any share link is created.
Does notNever sends to the client. Never publishes without explicit human approval.
Every specialist. Every time.
Four rules every specialist operates under.
These aren't marketing promises. They're contract tests, RBAC boundaries, and proposal queues. Enforced in code, not in decks.
Propose, don't commit
Every consequential action creates a proposal for human review — never a direct mutation.
Cite every claim
Every output links back to the source data it referenced. No black boxes. No "trust me."
Stay inside your role
RBAC-scoped context, least-privilege access. No cross-domain data fishing.
Show the signals
Ranking features expose the features behind the score — not just a scalar. You can override with full context.
Want to see the enforcement in code? Read the Safety deep-dive — ten commitments, each with a PROOF line citing the migration, library, or test that enforces it.
See the Intelligence Layer at Work
Want to see the AI specialists at work?
A demo walks you through one live phase — you'll see the proposal queue, the citations, the override buttons, and every migration that backs a safety claim.