Runs itself · ConstructINS
AI Insights
Scenario modeling, forecast accuracy and the reasoning behind the assistant’s recommendations.
What changed in AI Insights recently
Improved · Release 1.99 · 3 November 2026
The decision queue’s confidence is worked out from the evidence each recommendation holds instead of a typed-in figure, and the confidence breakdown no longer derives its parts from the headline number. AI Insights shows recorded outcomes and how often they came true, instead of figures that were not measured.
More in Runs itself
- AI ControlWhich models run, what they may read, and exactly what they are allowed to do.
- AI AgentsOne agent for each of the 47 job roles: it runs the job’s checks, drafts its paperwork, and prepares the decisions a person takes.
- Autonomous OpsWhat the system decided and did on its own, and how it re-routed delivery when something went wrong.