Oswald provides you with all the tools you need to train your chatbots NLP. In the background, it uses Oswald NLP, Watson Assistant or Microsoft LUIS. It's up to you!
Using the Oswald scenario builder, you can build advanced conversations in a visual user interface. You can even integrate your chatbot's conversation with other services.
You can integrate your chatbot with messaging platforms, voice assistants, text messages, regular voice calls, robots, cars... You can integrate with just about anything.
When your chatbot is live for your users, use Oswald built-in tools to test and analyse your chatbot. You can use these test results and analytics to extend and improve your chatbot.
Forward the conversation to a human agent, if the chatbot is unable to help.
Create one chatbot in multiple languages, like English, Dutch, French, German...
Collaborate with your colleagues and assign the right role to every collaborator.
Export your intents and entities for easy translation outside the platform.
Easily manage and switch between different versions of your chatbot.
Let the system automatically detect emails, numbers, currencies...
Take a look at Oswald in action! The scalability of our platform ensures the processing of a large number of messages and sessions, divided over many chatbots.
Oswald is used today by many customers, and by their customers. Take a look at our cases, built with Oswald.
NLP or natural language processing is a subdomain of artificial intelligence. Natural language processing is the technique to let a computer interpret and understand natural language, just like humans do. In the context of chatbots, this means the chatbot can recognise the intent of the user, based on its input. Also, the chatbot extracts some important entities in the input. When the intent and all entities are correctly understood by the chatbot, a response is triggered, executed and sent to the user.
Oswald is an all-in-one solution to train, build, integrate and maintain chatbots. This is different from other NLP services, like Watson Assistant, Microsoft LUIS or Google Dialogflow. These services focus mainly on the NLP, the recognition of intents and extraction of entities. Oswald includes many more features, like a visual scenario builder, human takeover, multilingual chatbots, advanced analytics, regression tests, a code editor, automatic context management... Furthermore, you can even swap the default Oswald NLP with your favourite supported NLP service like Watson Assistant or Microsoft LUIS. This means you can enjoy all the Oswald features described above, with the NLP offloaded to another NLP service.
Yes, you can! You can assign multiple supported languages to your Oswald chatbot, so you don't need to set up a separate chatbot evironment per language. Then, you can train one chatbot in these multiple languages. The structure of the chatbot is the same for all languages, like the scenario flows, the intents and the entities. The actual content, like training sentences, entity values, responses..., need to be implemented per language.
Oswald supports four languages:
Other languages can be added to our language models. Do you need a chatbot in another language? Contact us for more information.
All features in Oswald use an easy-to-use, clean user interface. You do not need a technical background to build an Oswald chatbot. However, if you want to integrate your chatbot with for example your CRM or other internal or external services, you can write some basic Python code in the Oswald code editor. But don't get too scared: in a major part of your chatbot this is not needed and if this is needed, it's nothing more than some basic HTTP API requests. If you don't know what that means, remember that it's nothing too fancy ;)
You can integrate your chatbot with your own CRM or even other internal or external services. The sole requirement is that the service has an API or application programming interface. If that is the case (for most services this will be the case), you can use the Oswald code editor in the scenario builder to call those API's.
Take a look at our code documentation.
The chatbot does not adapt its conversations or improve itself automatically over time. After all, the data to improve itself comes from your users and you don't want to give this responsibility to your users. Therefore, Oswald makes it as easy as possible to improve your chatbot in a controlled environment. You get an overview of unanswered questions to train and retrain your chatbot, execute your tests and publish it again. That way, you know and you control when your chatbot is adapted and improved, not your users.