The demark between AI and ML is simple – AI uses deep learning where for example a neural network could auto-determine new features for classification for example. Artificial Intelligence is the broader concept of machines being able to carry out tasks in a way that we would consider “smart” and Machine Learning is a current application of techniques based around the idea that we should really just be able to give machines access to data and let them “learn” for themselves.
Popular examples of artificial intelligence are Self-driving cars, voice recognition, natural language processing and image recognition. You must have seen or heard about Waymo’s ( part of Google ) self driving cars which have driven more than 8 million miles in real world working condition in autonomous mode. They use Lidar for building a image of their surroundings, radar for figuring out how far away objects are and their speed and on-board hi-resolution cameras for visual information such as whether a traffic signal is Red or Green. All these pieces of information are combined to form a singular image which is used to make decisions and the system is learning from every mile driven.
Apple uses machine learning for making its products better understand what we want and better able to supply what we want in a form we can understand. This technology concept is very important in customer service, summarizing business intelligence, and is a great interface for business data.
Virtual Agents are evolving simple chatbot based applications to intelligent systems that can network with various applications and systems to provide a seamless context aware network with humans. Alexa from Amazon is a good example of this. By enabling various skills, you can provide many useful pieces of information in a conversational way. In the recent Google I/O conference, Sundar Pitchai demonstrated how a digital assistant can setup an appointment with a series of interaction and with a near human like voice.
Google in particular uses deep learning to analyze images and identify objects and tag them correctly and they have built valuable deep learning algorithms and trained it with a huge data set.
There are many good examples of AI technology in everyday use and this is has largely made possible by having powerful and increasingly cost-effective GPU’s available to train data (which is a very expensive operation) and then make inference available at scale via the cloud. In addition the process and technology stacks available today have made collecting and processing data from a variety of sources easier There are even ways to make inference work on mobile devices via some types of models..
Now these technologies can finally be harnessed in business applications to help users with highly accurate recommendations and automating a wide variety of actions which can be performed. . The first generation of apps focused on capturing the data, the second generation focused on improving the business process, the third generation focused on better management reporting and business intelligence. It is time to build the fourth generation of business applications which deliver up-to-date accurate real-time recommendations for the user and easy execution of those recommendations.
What do you need to build this fourth generation of applications ?
1. Large and Disparate data sets
It is all about the data. All the answers an application user needs is available in the data and we must apply AI to get the nuggets in the data and this creates major competitive advantage. By having access to the best data, even in a highly competitive industry environment with similar algorithms, the best data wins and provides the best recommendation.
Here is an approach to create the best data to get the competitive advantage.
- First, get access to large volumes of comprehensive data and not the just a sliver of the data set.
- Second, create compelling business experiences to allow the users to provide information which were never captured – for example all the communications across multiple channels.
- Three, bring together social and fast moving market news e.g Twitter, Facebook, LinkedIn, Business News, Political news…and finally, bring curated data from select third party vendors to augment the data.
2. Artificial Intelligence engine
AI can help with identifying with unique insights and data correlations which cannot easily done by humans. Finding these correlations are the most important to finding the advantage to win new business or sustain existing business and provide the best end user and customer experience. Why is AI is important for business applications.
- First, the application of AI techniques allows for the analysis of large data sets and and multiple layers of data.
- Second, the learning algorithms can continuously improve the accuracy of the predictions.
- Third, the model itself can be adapted through manual and automated techniques.
All the ingredients needed for the next generation of business application are there – namely – large and disparate data, vast amount of enterprise data, the compute power to analyze all this data in a timely fashion and most importantly the ability to get these services with reasonable pricing. In a future post, we will cover more about how these can be applied to Sales applications.