In this video, Paul from shares an exciting update: our generative models now support sequential seed inputs, unlocking powerful new capabilities for time series simulation and forecasting.
Using the Rossmann dataset as an example, Paul demonstrates how to:
- Extend seed sequences into the future with probabilistic outcomes
- Generate multiple scenarios and visualize their likelihood distributions
- Run simulations even with limited input data (only dates, holidays, and promotions)
- Explore “what-if” interventions, like testing optimal promo schedules under budget constraints
Key highlights include:
✅ Continuation of time series with generative AI
✅ Simulation of multiple future paths with likelihoods
✅ Optimizing interventions (e.g., promo day scheduling) for maximum impact
✅ Practical applications in retail forecasting, planning, and strategy
Whether you’re working with retail sales data, customer behavior, or business planning, this feature helps you simulate, optimize, and make smarter data-driven decisions.
👉 Try it out and start simulating with Mostly AI!