Traditional forecasting uses past observation to predict the future to improve customer engagement, increase sales, reduce inventory and improve productivity. While it is simple to understand, the challenge is that this method too often neglects the precision and agility needed when making important decisions. The past does not always equal the future as environments are complex and chaotic. Shape trend lines lose value as they force moving averages and smooth out variability until we have an over-simplified the world that may not be an actual reflection of reality. Luckily, there are some more sophisticated tools today that produce future insights through the power of predictive simulation with time-based KPIs to enhance well-informed decision making. Moving away from over-simplified assumptions helps companies develop models that are complex, probabilistic and can incorporate potential failure events.
Scenario planning provides a framework to develop policies in the face of uncertainty. The scenario is merely an account of the reasonable future. Scenario planning contrasts different scenarios to explore the uncertain future consequences of a given decision. It is vital in scenario planning to have a clear purpose. Necessary steps include assembling a diverse group of participants to collect, discuss and analyze scenarios. When artificial intelligence is incorporated in the tool, these diverse group of participants should be used in the early stages to help define the inputs into the program. Companies across the world are using these tools today to make decisions based on what they think will happen in the future. The benefits of using scenario planning techniques include increasing understanding of important uncertainties, incorporation of alternative perspectives into planning and greater resilience around decisions to surprise.
For example, IBM uses a scenario building tool that uses artificial intelligence to support risk management activities in the areas like security and finance. The tool can generate alternative scenarios of the future and predict outcomes including both likely and unlikely futures. Both structured and unstructured data like relevant news, social media trends and domain knowledge can be inputted to paint the current state that can generate scenarios explaining the key drivers and related futures.
Another example includes the Stena Line that just announced the use of artificial intelligence on its ferries. The model predicts the most fuel-efficient way to operate a vessel on a route taking into account alternative perspectives of the future. Jan Sjostrom shares, “Planning a trip and handling a vessel in a safe and, at the same time, fuel-efficient way is craftsmanship. Practice makes perfect, but when assisted by AI a new captain or officer could learn how to fuel optimize quicker. In return, this contributes to a more sustainable journey.”
From transportation to healthcare, companies are leveraging machine learning engines that analyze a wide-variety of data and present a range of scenarios for optimization. With sophisticated tools that can learn and improve, there is a shift away from unfocused, historical data mining, to iterative, time decision analysis that can provide more useful insights for the future.
#PredictiveModeling #AI #ScenarioPlanning #MachineLearning