- Forecasting relies heavily on opaque ML/AI models
- Policy domains are politically sensitive and legally constrained
- Outputs are difficult to interpret and act on
Curve Intelligence
Human-centric AI for complex policy challenges
Curvint builds transparent, interpretable forecasting tools that put experts at the centre, embedding domain knowledge and rigorous theory to keep policymakers in the loop, so humanitarian response and policy planning can be driven by evidence you can trust.
The challenge
A new kind of policy foresight
Policy forecasting is increasingly data-driven, yet many models remain opaque, politically sensitive, and hard for policymakers to interpret with the theory-informed accuracy needed for actionable decisions.
- Data is rarely AI-ready in a theory-informed way
- Governance and transparency are often afterthoughts
- Policymakers lack tools they can validate and refine
- A human-centric, theory-informed AI framework
- Core policy drivers embedded with expert knowledge
- Built for transparency, trust, and human oversight
The platform
AI-powered decision support
An interactive platform that unifies diverse datasets, combines state-of-the-art forecasting with explainable models, and keeps human expertise at the centre of every prediction.
01
Integrate
Diverse datasets brought into a unified forecasting framework, AI-ready and governance-aware.
02
Forecast
Temporal Fusion Transformers paired with explainable models for transparent, interpretable predictions.
03
Validate
Human-in-the-loop: policymakers and practitioners validate, refine, and contextualise every output.
04
Decide
Interactive decision support that promotes accountability and informed planning across organisations and governments.
Adaptable to your priorities
The model reflects your policy priorities and operational needs, with expert knowledge embedded directly into the forecasting process.
Rigour meets usability
Transformer architectures for accurate, explainable forecasts, combining academic rigour with practical usability for diverse stakeholders.
Secure and self-contained
Operates independently of external APIs for greater control and data security. Continuously incorporates user expertise over time.
Stakeholder engagement
Built with policymakers, for policymakers
We work in active collaboration with government directorates, international organisations and the humanitarian sector, focusing on how policymakers use and interpret AI insights in practice.
- Integrating expert judgement through human-in-the-loop design
- Building systems that are transparent, trustworthy, and practical
- Developing tools that work for real-world policy environments
Contact
Let's talk
Interested in collaboration, a demo, or learning more about Curvint? Drop us a line. Your message reaches the full project team.