Chat with Alba on
Instant answers, wherever you message.
Channels · Processes & API
Assistance inside processes and systems.
Contextual chat in Smart Processes and web API endpoints for third-party applications.
Smart Processes
Assistance inside the work itself.
Contextual chat reaches the process and the specific text field.
A process carries more context than a general conversation. It has a purpose, a current activity and information already entered by the employee. Smart Mind brings assistance into Smart Processes through integrated chat, including support connected to specific text fields, so the employee can ask for help where the content is being prepared.
Use the context of the current activity
An employee drafting inside a process should not have to reconstruct the situation in a separate chat. The integrated experience uses the process context supplied to it, with the knowledge and actions assigned to the agent. A field-specific request focuses on the exact text being developed.
Keep ownership of the process clear
Smart Processes remains the system in which structured work is managed. The employee reviews suggested content and accepts it when appropriate; the process keeps its own workflow and responsibilities.
Process-based assistance
Current activity and field
Relevant knowledge and agent
Suggested content for acceptance
Web API endpoints
Connect agents to the systems you use.
A programmatic entry point extends assistance beyond the chat interface.
Some requests originate in another application. A business system may need help interpreting text, preparing an explanation or invoking a configured agent capability. Web API endpoints provide a route for third-party systems to communicate with Smart Mind agents, with the requesting system supplying the context needed for the task.
Two directions of integration
An external application can call an agent through a Smart Mind endpoint. Separately, an agent can invoke an action exposed through the Service Repository. The first brings work into the agent; the second gives the agent a capability to carry it out.
Design around a concrete exchange
A useful integration starts with a specific request and a clear expected result: the context sent, the identity and permissions involved, how the response is handled, and how errors and incomplete inputs are treated. These are agreed per project.