Much like trying to throw an axe blindfolded, traditional opinion-based forecasting misses the mark, often with disastrous consequences. Opinion-based forecasts have low predictability and accuracy, are prone to bias and manipulation, and yield limited value to the B2B organizations that adopt them. Fortunately, artificial intelligence is infusing and enhancing B2B sales forecasting. Sales leaders can attend our upcoming B2B Summit North America session to learn how to wield these new tools skillfully.

To understand the evolution of forecasting, it helps to understand the evolution of AI. The concept of artificial intelligence is nothing new — the term was coined in the summer of 1955 at Dartmouth College. Thus began The Age of Hand-Crafted Knowledge, during which AI researchers sought to mimic human intelligence with rules-based expert systems. These expert systems dominated the world of AI until about 2007, the dawn of the Age of Statistical Learning. In this period, companies started applying machine-learning algorithms to the new “Big Data” they were capturing to build predictive models and surface insights. No one called this use of machine learning “AI” at the time, but that all changed in 2012 when a deep neural network called AlexNet won the ImageNet competition, besting humans at identifying images. The Age of Deep Learning was underway and is responsible for the renaissance the field of AI is enjoying today.

The three ages of AI are mirrored in three types of forecasting:

  • The Opinion Forecast. This is the traditional forecasting method B2B organizations employ. As the name suggests, it is largely based on the rep’s opinion and is therefore neither efficient nor consistently accurate.
  • The Augmented Forecast. The augmented forecast leverages machine learning trained on historical structured (i.e., row and column) data to build predictive models that augment sellers’ and managers’ opinions. It increases forecast accuracy and also leads to higher win rates by providing greater insight into buyers. There is still a great deal of rep input, so the predictions augment opinion.
  • The Prescriptive Forecast. This emerging type of forecast leverages deep learning on both structured and unstructured (voice, text, etc.) data to derive an even more accurate forecast. Because deep learning requires a significant volume of data to outperform classic machine-learning methods, some vendors are training models on a network of their clients’ engagement data. Here, reps’ opinions augment the prediction, and most of the human focus is on beating the number by leveraging these deeper buying signals.

In our upcoming B2B Summit North America presentation this May, Anthony McPartlin and I will dive deeper into AI-enhanced forecasting and the features and functionality that are currently available and in development. Most importantly, we’ll advise you on best practices for evolving your own forecasting practices with AI to ensure you’re hitting your targets.