Imagine one living model of the whole city: where the heat, water, traffic and energy are right now, and where they are heading. It goes one step beyond the digital twin: not just a mirror of the city, but a model that knows its state and can play it forward.
An urban world model, fed by every sensor. AI agents and robots can read that state and act through Yggio, under the same rights and rules as everyone else.
Foundation models like Chronos-2 already forecast time series they have never seen.
Next, they learn to read a rough description too, like "return temp, substation 4", and understand what they are forecasting. Urban foundation models put every series together, and you get the model from the top of the page.
Much of Yggio’s AI started in funded research with universities, industry and public partners. The work continues in Horizon Europe and Chips JU projects on edge AI and resilient infrastructure.
An engineer frames the task and its acceptance criteria. AI coding agents write the change together with its tests. Fifteen required automated checks qualify it: build, static analysis, a complexity ceiling, licences, a coverage requirement and thirteen test suites run against real services. An AI review reads the whole change and its callers first, flagging authorisation gaps and scale regressions for the human reviewer; it advises, it never merges. A named engineer then reviews, signs off and remains accountable for the code that ships.
All Yggio code lives in one GitHub repository, the single source of truth. Every change is made on a feature branch and merged through a pull request that another engineer approves. The main branch is protected: nothing lands directly, and a merge queue re-runs the full suite before anything reaches it. A change is done when its tests exist and pass, not merely when the feature works, and each new test stays in the suite for good.
A commercial Yggio release comes every 8–12 weeks, after four weeks of staged quality assurance (staging, beta, public, private and on-premises), with API compatibility, authorisation and vulnerability checks before rollout.
We use AI to move faster, and keep customer trust and data protection non-negotiable. Our AI policy sets the rules for everyone at Sensative:
Looking for an industrial partner for AI on real infrastructure data, such as sensors, buildings, networks and city systems? We take part in Horizon Europe, Chips JU and Vinnova projects.