Add the capability your operation needs
Choose AI Workforce to carry continuing tasks across tools and people. Choose Industrial Platform to organize industrial data and calculations, develop applications or run selected professional capabilities. Industrial Platform and individually licensed applications can serve existing systems and customer AI without AI Workforce.
Existing acquisition, SCADA, production, maintenance and data platforms can continue their roles. Identify the application, records, operating relationships and output destination for the selected work. Agree meaning, units, timing, access and exception handling alongside the interface.
Integrate with your existing systems
The categories below describe the integration approach. Confirm product versions and interface scope in the solution; named compatibility requires validation of the relevant versions.
- Acquisition and SCADA
- Retain connections, monitoring and existing control. Read selected measurements, states and events through available interfaces, adding operating analysis and specialist applications.
- Data platforms and time-series databases
- Retain storage. Supply live records, historical intervals or aggregates for the application, map objects and timing, and make them available to specialist models and AI.
- Customer models and algorithms
- Retain customer models and expertise. Connect inputs, execution and results through the existing runtime, including the objects and computing environment the model needs.
- Existing AI and agent platforms
- Connect through the API/MCP interfaces confirmed for the deployment. The AI uses an authorized identity to discover objects, retrieve data, run calculations or execute workflows within the agreed scope.
Agree permissions for reading, writing and actions separately
The operating mode follows the task: analysis, recommendations for confirmation, or execution within agreed authority and limits. Connecting a data interface does not itself grant permission to change records or operate equipment.
Define the destinations, triggers, responsible parties and checks for each permitted action. Suitable existing systems may provide the interface and control capabilities.
- Read data
- Define objects, variables, historical windows, frequency and permitted scope.
- Write results and tasks
- Define the receiving system, permitted changes and result queries, reusing existing identity and access controls.
- Equipment or business actions
- Configure executable tasks, operating constraints and authorization separately. Retain required equipment interlocks and existing control responsibilities.
Choose deployment for data and operating requirements
Edge runtime performs collection, storage or computation near equipment; remote applications coordinate cross-asset information and work. Choose deployment around networks, data location requirements, model dependencies and maintenance.
Specify which data leaves the site, which calculations remain local, what continues while disconnected and how records are reconciled after reconnection. Reuse suitable customer environments.
Deliver a usable task, with continuing support
The delivery list identifies the product or application licenses, configuration, data and object mapping, interfaces, permissions, professional adaptation and handover. Check a representative task from input through calculation to the agreed output, including missing records, access failure and any human or system action.
Assign product support, method maintenance, customer configuration and site execution to their respective providers. Qualified implementation teams from the customer, partners or Be Water connect and verify supported components using established methods; specialists provide the adaptation or judgment included in scope. At handover, identify the first contact and escalation routes for configuration, product defects and method applicability. Agree support periods, updates and runtime separately from initial delivery, and assess operating value from actual work and full effort.
Use APIs, SDKs and AI-accessible capabilities
Team members’ historical industrial-platform materials and demonstrations include SDK and API foundations, API/MCP access and two-way Dify integration. These are relevant technical starting points, distinct from the Be Water products still to be launched; they do not imply a public SDK download or sandbox is available.
In the first project scope, confirm the version, materials, interfaces, licenses and delivery dates that can be provided. Identify which objects, records, specialist calculations and workflows are exposed, with their inputs, outputs, identity and error handling. The customer can retain its existing AI interface and use the selected capabilities through the agreed access route.
For your deployment
Which systems hold the records, who can authorize access and which assets may be read?
Who is responsible for data use, integration, deployment and recovery?