Deliver the right insights to the right people at the right time to make better decisions faster.

Personalized insights

Harmony Insights can tailor its results to each of your team members, providing highly personalized insights most relevant to their role and access level. Every user can access the same set of possible analyses and get real-time, hyper-focused insights.

Complex, enterprise-grade programs are virtually impossible for humans to properly manage. The very nature of their size and scope means you can't possibly have your eyes on every project your team is executing. Let Harmony watch your blind spots with intelligent, highly focused Insights analysis across your biggest concerns.

Watch your blind spots

Harmony's deep access to all of Conductor's data, always fully respecting access controls and permissions for the current user, gives you the superpower to keep tabs on even the most complex program. Harmony Watchtower can be your lookout, keeping a bird's eye view across your full program and acting as an early warning system to catch potential issues before they become real problems.

At Sensei Labs, customer dataprivacy and isolation are foundational to the architecture of Conductor and Harmony AI. All data is protected from unauthorized access, never used for AI training, and handled responsibly across infrastructure providers, including Microsoft Azure and Google Cloud Platform (GCP), and across AI providers including Azure OpenAI and Vertex AI Gemini.

Customer data isolation in Conductor (single-tenant)

Conductor is designed around a single-tenant data model, meaning each customer’s core application data is isolated within its own database instance.

• On Microsoft Azure, customer data is stored in logically isolated databases using Azure SQL Elastic Pools.
• For customers in the Kingdom of Saudi Arabia (KSA), Conductor is deployed on GCP in order to meet PDPL local data residency, privacy, and compliance requirements.

Across both platforms:

• Each database is logically separated and protected by role-based access controls, encryption, and audit logging.
• No customer can access another’s data.

Harmony AI (multi-tenant)

Harmony,Conductor’s AI, provides features like real-time coaching, analysis, summaries, and AI-generated insights. Harmony is implemented as a multi-tenant service, but with strict isolation and region-specific deployment strategies.

Supported AI providers

Harmony supports integration with:

· Azure OpenAI Service (Microsoft Azure)

· Vertex AI Gemini (Google Cloud)

The AI provider is selected based onthe customer’s hosting region, regulatory requirements, and configuration preferences.

Region-specific Harmony Instances

To meet compliance and residency expectations:

• Customers in KSA use a dedicated Harmony instance deployed within GCP KSA, ensuring data never leaves the country.
• All other globalcustomers use Harmony instances hosted in Azure regions closest to their Conductor deployments.

Each Harmony instance maintains strict tenant isolation at runtime, regardless of location.

Prompt handling and privacy

• Harmony processes prompts in-memory only, and never stores data unless the user explicitly chooses to save an AI-generated result.
• All inputs/outputs are ephemeral, scoped to the user’s session, and never used for model training or fine-tuning.

AI provider data handling – no model training

Harmony communicates with all underlying AI service providers via secure APIs under enterprise contracts and configurations that disable data logging and training. We ensure all AI integrations adhere to the following:

“Your prompts (inputs)and completions (outputs), your embeddings, and your training data:

• ‍are NOT available to other customers.
• are NOT available to OpenAI.
• are NOT used to improve OpenAI models.
• are NOT used to train, retrain, or improve Azure OpenAI Service foundation models.
• are NOT used to improve any Microsoft or third party products or services without your permission or instruction.

Your fine-tuned Azure OpenAI models are available exclusively for your use.”

Vertex AI Gemini

From GenerativeAI, Privacy, and Google Cloud:

“We do not use data that you provide us to train our own models withoutyour permission.”

Hosting and compliance

Sensei Labs supports flexible, region-aware hosting strategies:

Region	Platform	AI provider	Harmony hosting location
Global (default)	Microsoft Azure	Azure OpenAI	Azure-hosted Harmony
KSA	GCP (Dammam)	Vertex AI Gemini	GCP KSA–hosted Harmony instance

All deployments meet enterprise-gradecompliance requirements, including:

• SOC 2, ISO/IEC 27001, GDPR, HIPAA, and local KSAcompliance.
• Full encryption of data in transit and at rest.
• Role-based access, auditing, and secure software development practices.

Summary of Commitments

• Conductor Data Isolation: Single-tenant model with logically isolated DB per customer.
• Harmony AI Layer: Multi-tenant with strict per-tenant runtime isolation.
• KSA Residency Compliance: Dedicated Harmony instance hosted in GCP KSA region.
• AI Prompt Retention: Ephemeral unless output is explicitly saved in Conductor by user.
• AI Model Training: Neither Azure OpenAI nor Vertex Gemini use data for training.
• Hosting Infrastructure: Azure (global) or GCP (KSA), based on customer need and compliance.
• AI Provider Flexibility: Harmony dynamically routes to Azure OpenAI or Vertex Gemini.