Inspiration
Modern enterprises are experimenting rapidly with AI, but most struggle to operationalize it safely.
AI insights often remain opaque, non-repeatable, and disconnected from governance and audit trails.
This project was inspired by a simple question:
How can AI-driven analytics be made enterprise-trustworthy, deterministic, and auditable—without slowing innovation?
What it does
Tableau Next Enterprise AI Orchestrator is a control-plane that coordinates AI-powered analytics workflows across Tableau Next and Salesforce.
It enables:
- AI-assisted insight generation with deterministic execution paths
- Orchestration of multi-step analytical workflows
- Governance, traceability, and audit-ready execution logs
- Human-in-the-loop checkpoints for sensitive decisions
- Enterprise-grade guardrails for AI usage
Instead of AI acting as a black box, every step is observable, explainable, and enforceable.
How we built it
The system was built as a modular enterprise kernel consisting of:
- An orchestration engine that manages workflow state, retries, idempotency, and deadlines
- AI services that generate insights, summaries, and recommendations
- Governance layers enforcing execution rules, audit logs, and approval gates
- Tableau Next dashboards as the primary analytics and visualization surface
- Salesforce as the secure orchestration and metadata backbone
AI is used where it adds value (insight generation, summarization, reasoning), while deterministic logic controls execution, compliance, and lifecycle management.
Challenges we ran into
- Preventing AI-driven workflows from becoming non-deterministic
- Designing idempotent, retry-safe orchestration patterns
- Balancing flexibility with enterprise governance requirements
- Ensuring AI outputs remain explainable and auditable
- Avoiding “demo-only” shortcuts while building a production-grade kernel
Accomplishments that we're proud of
- A fully enterprise-grade AI orchestration kernel (not a prototype)
- Deterministic workflow execution with AI augmentation
- Clear separation between AI reasoning and system control logic
- Governance-first architecture aligned with enterprise expectations
- A scalable foundation that can support multiple analytics products
What we learned
- AI becomes truly valuable in enterprises only when paired with strong orchestration
- Governance and explainability must be designed first, not added later
- Tableau Next is powerful when used as a semantic and visualization layer, not just dashboards
- “Responsible AI” is an architectural outcome, not a policy document
What's next for Tableau Next Enterprise AI Orchestrator
- Expand pre-built orchestration templates for common enterprise analytics use cases
- Deeper integration with Tableau semantic models
- Automated policy-driven approvals and risk scoring
- Multi-tenant SaaS packaging for enterprise deployment
- Advanced observability and compliance reporting
Log in or sign up for Devpost to join the conversation.