Training your virtual assistant is essential for short-term improvements. Most companies either don't have the right tools or haven't budgeted resources for it. Because to accomplish this task, you need at least one employee or even an entire team of intent/content managers, depending on how many conversations the virtual assistant needs to handle. In addition, the team must be trained and allocated time or resources to deal with the optimization of the bots on a daily or weekly basis.

How do virtual assistants work and what is chatbot training?

Each chatbot is made up of intents that describe the various intentions of the user. For example, an intent could be that a user wants to know the weather forecast. Intents consist of training phrases and one or more responses. Training phrases, also known as utterances, are example questions. Training phrases for the "weather forecast" intent could therefore be: "What will the weather be like tomorrow?" or "Could you tell me whether it will rain at the weekend?" An answer could be: "The weather will be sunny tomorrow with 20 degrees". Most services that are used to process and understand natural language work like this. When implementing a chatbot, it is advisable to set up an intent with around 15-25 training phrases. This gives the system enough data to match the user requests with the correct intent. The higher the number of question examples, the better the virtual assistant's ability to answer the user's questions correctly.

Establish processes to train and optimize your virtual assistant 

As soon as the chatbot goes live, training should have already taken place. For this, it is crucial to already have a defined process that specifies who and when reviews and monitors the conversation progress between bot and User:in. This is a continuous process that initially requires more effort and time as the virtual assistant improves.

The main task is to see if the bot is able to match the correct intent with the user's question. If it does, the intent manager must confirm the bot's action. Then the user's question is entered into the intent's list of training phrases, if it does not already exist there. If the request was mismatched, it is necessary to select the correct intent so that the bot enters the question into the list of training phrases. Often chatbots send out error messages indicating that they could not find a matching intent to the query. In this case, it is important to evaluate whether a new intent should be created to add more knowledge to the system. So the next time a similar question comes up, the bot now knows the answer. This is how chatbot training works.

Use an effective chatbot training tool  

To accomplish this task, the team needs the right tools with specific functionalities. This should be an interface that displays the conversation history and the intents matched with the incoming questions. It should be easy to correct or confirm the bot's decision. Also, it is very helpful to have filtering options to list only conversations in a certain time period, with certain intents, or with a certain confidence score. The confidence score is a metric from 0 to 1 that indicates how confident the system is in terms of intent detection. For example, one can set the natural language understanding (NLP) system to a score of 0.6. This means that the system will send the response from the taken intent as soon as it is at least 60% sure. If the chatbot is only 50% sure that the intent matches the request, the user would receive an error message. Generally, the higher the confidence score, the more error messages will occur. The lower the confidence score, the higher the probability of hitting wrong intents. In terms of training, it is then very useful to be able to filter for intents that have been matched with a very low confidence score, as it is more likely to improve the chatbot there. This is especially true for virtual assistants that receive more requests than the team can review. In this case, it is recommended to filter the conversation history to work more efficiently.

The Conversational Middleware Platform BOTfriends X provides an interface for training in intent management. Learn more about BOTfriends X.

Intent Management Training example

Optimize your virtual assistant for the long term with a strategic and content roadmap

If you want to optimize your chatbot, consider long-term improvements. Consider what features or more use cases can be integrated.

For example, the current channel may not be receiving as much traffic or users may not be interacting with the chatbot as expected. Then you should consider whether it makes sense to switch channels. For example, from Facebook Messenger to the website. You might also need to revise the overall structure of the conversation by adding new training phrases or removing content.

You also need a roadmap from the start that illustrates your "Conversational AI Journey." It makes sense to start with a POC, but after that, scaling is critical to reach more users. You can do rollouts to other countries by adding more languages to the bot. Sometimes it's useful to evaluate whether the chatbot has the potential to act as a voice assistant or vice versa. Or there are new technologies like NLP services on the market that better fit the needs of your business. In general, many companies start with a simple FAQ bot that answers the most important questions. After evaluating call data, they plan to add more complex use cases or processes within the chat that require integration with enterprise systems such as SAP, databases or other software. In summary, these enhancements need to be defined in the long-term roadmap to pursue an overall "ConversationalStrategy".

If you would like to find out more about training your chatbot, you can download our free white paper "Chatbot Operations".

Download Whitepaper