Enterprise Mesh is the secure execution layer of Smart AI. 
It connects Smart Chat requests to approved enterprise systems, executes them in real time, and returns a single, trusted answer. 

Traditional integrations rely on data replication (ETL) and point-to-point connections. 

Enterprise Mesh is query-centric: 

• No data copying or syncing 

• Queries run directly on source systems 

• Results are merged in real time 

This reduces latency, complexity, and maintenance. 

No. 
Enterprise Mesh does not store or duplicate data. It executes queries directly on source systems and returns results in real time. 

It combines several steps: 

1. Runs federated queries across systems (ERP, WMS, TMS, etc.) 

1. Performs live aggregation of responses 

1. Applies on-the-fly transformations (filters, grouping, calculations) 

1. Uses business-defined reducer logic to select the final result 

The outcome is a single, consistent answer. 

Smart Chat is the conversational interface of Smart AI that lets users interact with enterprise systems using natural language. 

You can: 

• Retrieve live operational data 

• Run approved workflows 

• Ask follow-up questions 

• Generate charts and dashboards 

Smart Chat is API-centric, not document-based. 

Yes—but only if: 

• The function is pre-approved 

• The user has the required permissions (RBAC) 

No arbitrary or unsafe actions can be executed.

A Smart Function is an approved capability that Smart AI can execute, such as: 

• Fetching data (orders, inventory, shipments) 

• Running workflows 

• Calling APIs or system commands 

Each function includes: 

• Implementation (code/script/API) 

• Metadata (inputs, outputs, description) 

Smart FX is the builder studio where teams create and manage Smart Functions. 

It provides: 

• Function definition (metadata + implementation) 

• Connections to enterprise systems 

• Evaluation tools (to test AI tool selection) 

• Governance via Git 

1. User asks a question in Smart Chat 
2. LLM extracts intent + parameters 
3. A matching Smart Function is selected 
4. Enterprise Mesh validates permissions and inputs 
5. The function executes on target systems 
6. Results are returned in a user-friendly format 

No. 

• The LLM only receives user query + metadata 

• All execution happens inside your environment 

• No raw business data is exposed externally 

For analytics: 

• Only column names (and optional small samples) may be used 

• Full datasets remain internal 

Security is built-in at multiple levels: 

• RBAC (role-based access control) 

• Approved functions only (no arbitrary execution) 

• Centralized validation and permissions 

• Secure API layer 

Users can only access what they are authorized to see and run. 

Yes (if enabled). 

Users can: 

• View executed commands or API calls 

• Understand how their request was processed 

• Troubleshoot results

Yes, Smart AI maintains session context, so users can: 

• Refine results 

• Ask incremental questions 

• Build on previous responses 

Yes, From any dataset, users can: 

• Create charts (pie, bar, line, stacked, etc.) 

• Generate dashboards automatically 

• Analyze trends without manual setup 

Any system with an API, including: 

• WMS (e.g., Blue Yonder, Manhattan) 

• ERP (e.g., SAP) 

• CRM (e.g., Salesforce) 

• Data platforms (e.g., Snowflake) 

• Custom REST / SQL / proprietary systems

Enterprise Mesh supports secure tracking, including: 

• Who made the request 

• Which function was executed 

• Execution timing and run IDs 

• Masked parameters 

Sensitive data is never logged or exposed. 

Use it when: 

• Data is spread across multiple systems 

• Real-time answers are required 

• Actions must be secure and governed 

• A unified interface is needed 

• No ETL or data duplication 

• Real-time, federated execution 

• API-driven (not document-based) 

• Built-in governance and security 

• Low-code extensibility via Smart FX