What is AI (Artificial Intelligence) anyway? Help Center May 04, 2022 19:22 Updated Index: But what is AI in Blip anyway? How to use AI? Have you published your AI model? Artificial intelligence is a branch of computer science concerned with building machines that simulate intelligent behavior. The topic has gained great notoriety in the last 10 years due to two main factors. First is the significant increase of available data, mainly unstructured data (images, videos, text, etc.). Second is the increased availability of computing resources in the cloud. This set of factors boosted investments in machine learning technology by several companies. When talking about machine learning, problem solving techniques fall into two main categories: supervised and unsupervised learning. In general, supervised learning requires that we have the data and further examples to train the model. Unsupervised learning only requires that we know the data for the model to work. A topic of great importance when it comes to AI is Natural Language Processing (NLP). This discipline studies how machines can understand human language. This is particularly important when we talk about chatbots, because language is the foundation of this type of solution. But what is AI in Blip anyway? Blip offers integration with AI providers that perform natural language processing (NLP) used to understand text messages sent by the user to the so-called "smart contact". In this case, we are talking about supervised learning algorithms. Therefore, this artificial intelligence needs to be set up and trained by the users of the platform. Made for those who want to make their chatbot smarter, AI in Blip helps to create automated services that can reduce potential communication mistakes and deliver faster and more fluid conversations. By using AI to create automated customer service that understand how your customer talks, you can have the following benefits: More natural and fluid conversations between your brand and your customer. Effortless delivery of information, without the need to create several menus. Decrease in communication mistakes that lead to the abandonment of conversations. Knowledge of communication patterns of those who talk to your brand. How to use AI? To set up your AI, first you need to choose an AI provider. It delivers an artificial intelligence computing resource to the cloud, that is, it is a company that will host your AI. Blip currently provides native integration with 3 different AI providers. There are performance differences between them, depending on the basis used for the model. The main impact for a business though is the price that each provider charges. The platform's proposal is to allow the customer to change providers without difficulties. Integrated providers are: Google Dialog Flow (learn how to set up Dialog Flow on Blip), IBM Watson (learn how to set up IBM Watson on Blip) Microsoft LUIS (learn how to set up LUIS on Blip) Once you choose your AI provider, you have to create some "intentions'' for your AI model (learn how to creat intentions). An intention is a task or an action that the user wants to perform, for example "issue a ticket". In other words, they are the objectives expressed in an input message (i.e., text) from the customer. Creating an intention involves creating some examples of sentences (samples) that express the same idea. In the case of the "ticket" example, other sentences that express the same idea would be: "generate a ticket" or "create a ticket". To complement your model, you can create "entities" that give your AI greater reach and capillarity (learn to create entities).Entities are parts of a sentence that represent an object or a value. Creating an entity involves mapping products, services and contexts within the business. If you are a more advanced user and want to use a combination of intentions and entities to interpret different contexts, your to-go tool is the content assistant (learn how to use the content assistant). With this assistant, it is possible to identify several sets of intentions and entities to give an even more assertive answer to your client. Once your AI model is set up and ready, you must train it and publish it (learn how to train and publish your model). At this point, you also have at your disposal an AI testing window that allows you to validate if intentions and entities are being correctly identified within example sentences (learn how to test your AI model). Have you published your AI model? Now we have to set up your builder flow so that our smart contact can begin to better understand the users. To do so, we have several options of how to apply AI in the flow. But there's an important question here: the evolution of your AI. That is because in order to develop a good model, one that spans several options that your user can type, we have to evolve our model over time. Blip currently has 2 important tools to help evolve your model. An improvement tool, which allows you to identify the sentences sent to your AI and add them as examples in your intentions (learn how to use the improvement tool). In addition, you can use a test file or your customers' examples to evaluate your model's performance and coverage (Evaluate your model using the model's analysis tool). With the tools described above, it is possible to create and set up your first AI, apply it to your intelligent contact in an effective way, according to your business, and evolve it to get the most of the AI power. That way, you can improve the relationship between your brand and your customers. For more information, visit the discussion on the subject in our community or the videos on our channel. 😃 Related articles How to configure Dialogflow as your AI provider Creating entities and intents Using the AI tool inside Builder How to incorporate your AI model to your chatbot How to verify the AI provider settings