A live process graph, with anengineering-grade AI on top.

StreamSight uses your diagrams and plant documentation to build process context, connects to the PI System through its connector for live data, and runs real analysis on top. It analyzes, diagnoses, explains, and monitors your processes alongside your engineers.

1.1

The graph, from your existing views

StreamSight uses your diagrams and documentation to build process context as nodes, edges, bindings, and sub-diagrams.
1.2

Diagnose end to end

Plain-language symptoms open an investigation that traces the graph, pulls the tag history that matters, and names every id along the way.
1.3

Computed answers

A Python sandbox executes correlations, regressions, change-point detection, and spectral analysis, whatever the question calls for.
1.4

Findings become live views

Investigations end as dashboards, with KPI tiles and trend charts wired to your tags and refreshing on their own timers.
1.5

Engineers make the call

Changes arrive as reviewable proposals, and plans open before multi-step work runs, so your engineers make the operational decisions.
1.6

Bind nodes to live data

Bind to a PI System tag, a derived expression, or a synthetic source, so the diagram updates live while development keeps moving.
1.7

Status at a glance

Profiles convert thresholds to green, amber, and red, so operators and the agent read the same state.
1.8

Drill into sub-systems

Any node opens into its own sub-diagram, so a single symbol can hold a whole skid.
1.9

A shared process library

Reusable shapes, tags, and edge types keep hundreds of diagrams consistent as the model grows.
1.10

Corrections that persist

Tell it once to use the redundant sensor after maintenance, and the lesson resurfaces whenever it is relevant.

See it on your own unit.

Design partners get a working environment against a unit of their choice, with the connector reading their PI System and their engineers reviewing what the agent proposes.