Foresight is a predictive modeling platform: models that forecast how places, networks, and markets will change, the data that powers them, and the infrastructure that serves them to your systems. Atlas AI scientists and engineers work alongside your team, fitting each forecast to your decision and your conditions.
FIG. 1Forecasting demand: monthly air travel to 2030, with its likely range.
An illustrative example on public data from the US Bureau of Transportation Statistics.
A port expansion reprices freight a border away. A drought moves people, and demand for power, housing, and clinics moves with them. Standard models treat places as independent, so they miss how change spreads.
Demand and risk move along roads, grids, and supply chains. Forecasting them takes models built for networks.
Local statistics arrive late and at coarse scales, in every market. Foresight can start from Pulse’s record of how each place has changed.
What a place will look like in 1, 3, or 10 years (population, demand, the built environment, and exposure), with a likely range to size capital decisions against.
Where to place, route, and reposition towers, clinics, depots, and substations. The models keep running, so the plan updates as conditions on the ground change.
Scanning imagery and sensor data around your sites and routes for construction, clearing, demolition, and other change that adds risk.
What a new plant, road, or power line changed for the people around it, separated from everything else that happened at the same time.
FIG. 2Measuring impact: settlements newly connected to a mini-grid in the eastern Democratic Republic of the Congo gained 0.22 points of asset wealth by 2021.
Redrawn from Itad (2024), CC BY 4.0: an evaluation for FCDO designed in partnership with Atlas AI.
Outbreak forecasts with their likely range, and the settlements that health-facility catchments miss, across every district of the state. With Gavi and UNDP.
Bottom-up, route-level passenger-demand forecasts built with Airbus’s Global Markets Forecast team, to inform supply-chain and commercial strategy.
Evaluations of a hydropower mini-grid (Fig. 2) and of broadband fiber in the Democratic Republic of the Congo, designed by Itad with Atlas AI, for FCDO and BII.
Predictions of where households are at risk of missing the health and nutrition services essential to child survival, to target the programs that prevent malnutrition.
Used across five sectors
Foresight runs on Aperture®, the engine that also powers Pulse.
Data
Every source, on one grid
Models
Trained, then tuned to you
Infrastructure
Runs and updates every model
Delivery
Into your systems
Team
Atlas AI scientists and engineers, working with your team at every step
Our method for forecasting change across places received a US patent in 2026. How we govern our models →
Weeks 1–4
Define the decision, the geography, and the data you hold. We match the decision to models that already exist and agree on what success looks like.
By weeks 8–12
Models tuned to your conditions and joined to your data, served to your systems through our APIs and infrastructure.
After launch
New use cases, places, and teams build on the same models, data, and infrastructure, with new models where a decision needs them.