Augmented intelligence is all about bringing together the power and strengths of AI with those of humans by integrating AI systems into the day-to-day work of people to help them make better decisions. While augmented intelligence is easy to understand in theory, many organizations struggle to implement it in practice and at scale — here are three real-word examples of augmented intelligence.
1. Next-Best Action Systems
A perfect example of augmented intelligence are next-best action systems, or machine learning-based recommendation engines that are surfaced to front-line people (whether customer service representatives, claims handlers, financial advisors, sales associates, etc.) to enhance their ability to help customers. For example, imagine a system that provides personalized recommendations for products he or she is most likely to be interested in for a sales associate to recommend while on the phone with a customer.
In the video below, Jeff McMillan — Chief Analytics and Data Officer at Morgan Stanley — discusses their next-best action system, which takes thousands of data points and uses those to recommend possible content for the financial advisors to provide to customers (the power of machines), yet ultimately allowing the advisors themselves to decide what makes sense and provide the best possible and most personalized experience (the power of humans).
2. AI-Powered Targeting for Marketers
As with next-best action systems, business experts (like marketers) can greatly benefit from augmented intelligence systems that combine the power of machines to determine who campaigns should target at a massive scale plus the power of humans to determine exactly what the right message is for that audience.
For example, Showroomprivé (an e-commerce retailer specialized in flash sale) leverages Dataiku to build machine learning-based targeting. Their marketers then use this targeting to build campaigns that are 2.5x more effective.
3. Automated Email or Case Triage
When it comes to customer service, though AI has come a long way, nothing beats the ability to talk to a human who will solve the problem. The case for augmented intelligence here is the ability to better triage emails or cases (whether from customers or for an internal department, like IT) with machine learning so that issues get resolved faster. Again, the idea is not that the machine replaces the human, but that humans become even more powerful and able to make good decisions when augmented with machine intelligence.
For example, both Rabobank (in the IT department) and Etihad Airways (in the customer service department) use machine learning-powered automated triage to enhance the problem solving ability and speed of those representatives.
While augmented intelligence is similar to the concept of human-centric AI, it’s slightly different than human-in-the-loop machine learning, which is more focused on infusing human intelligence back into machine learning models. In other words, while the goal of augmented intelligence is to use machines to enhance humans, human-in-the-loop machine learning is sort of the reverse. Of course, a well-rounded AI strategy includes efforts to implement both augmented intelligence and human-in-the-loop systems.
Augmented Intelligence + Everyday AI
Everyday AI is about making the use of data pedestrian — AI that is so ingrained and intertwined with the workings of the day-to-day that it’s just part of the business (not only being used or developed by one central team). Everyday AI goes hand in hand with augmented intelligence because the best augmented intelligence systems are deeply integrated with the business at this level.