Google CEO Sundar Pichai recently claimed that «A.I. is more important than fire or electricity». He is one of many leading technology experts who are convinced that AI will transform both our daily lives as well as the economy across practically every imaginable industry. We have discussed the hype and the added value of AI in our latest blog. But while the potential of AI is being hyped all around the world, it can be very challenging to break through this buzz and figure out where to begin. In this blog, we want to give you some tips and tricks to get started with AI in practice!
Create value with out-of-the-box models
One of the most important ways to create momentum when introducing a new technology is showcasing the added value as soon as possible. In AI this can easily be done by leveraging out-of-the-box models. No need to reinvent the wheel for common use cases, since there are plenty of models available that have been pre-built by AI experts. For example, if you want to perform image recognition or face detection, you can use the Google Vision or Amazon Rekognition API to get access to state-of-the-art models within minutes. With just a few lines of code, you can embed AI into your applications. Another example is Einstein AI, through which Salesforce offers plenty of out-of-the-box models to its customers. With Einstein AI, businesses can build truly intelligent CRM systems by incorporating features such as lead scoring and product recommendation systems.
Customise to fit your unique needs
After you have discovered the possibilities and added value of AI, you can start looking at building your own models to cover specific use cases. Don’t worry if you don’t have a PhD in Data Science or Statistics, there are some great tools available to get started without being an expert in coding machine learning algorithms. Amazon Machine Learning for example, is a service that enables you to explore your data and build custom models using a graphical interface. On top of that, it let’s you easily deploy those models into your application without managing any underlying servers. Salesforce offers similar tools to customise your models through MyEinstein. With MyEinstein you can build predictive models and machine learning apps that are not provided out-of-the-box, such as customer retention models or chatbots.
The drag-and-drop tools mentioned above are great to get started building custom machine learning models. However, they do not offer full customisation in terms of the types of data you can use and the algorithms you can build. If you want 100% flexibility you will need to write code yourself. In AI and machine learning, open source programming languages have gained massively in popularity. In Python for example, you can access many specific AI frameworks, such as TensorFlow and MXNet. If you are new to all of this, I would recommend starting of with Keras, in which you don’t need to worry about the smallest of details.
Start your AI journey
Building these custom models from scratch can be quite challenging. At 4C we have plenty of experience in this domain. Our latest development TellMi, is an AI platform that analyses unstructured text data. TellMi can for example be trained to handle cases more efficiently or discover what people are telling you in customer satisfaction surveys. So, if you ever get stuck on your AI journey or simply want a head start, we are here to help you!