Machine learning powered conversational interfaces
We use the same technology powering Amazon Alexa to build intelligent chatbots which can assist customers without human interaction.
Our flexible bot framework allows chatbots to integrate with your other software, e.g. booking systems or knowledge bases.
We work closely with our clients to develop a chatbot script which is well-researched, tailored to users’ needs and appropriate in tone. We take a data-driven approach to developing the chatbot responses, based on surveys and mock interactions. By doing this, we can be sure that the bot is going to do a great job before any money is invested on development.
Interacting with a chatbot can be daunting for users who don’t know what the bot is capable of. That’s why the scripts we develop are carefully crafted to keep the conversation flow as intuitive and understandable as possible.
Our chatbot framework uses AWS Lex as its natural language processing – the same technology behind Amazon Alexa. Lex uses advanced deep learning to understand what the user is asking, and then to match it to the correct response we scripted.
When we develop scripts for a chatbot, we “train” the bot using example user phrases – we consider all the ways a user can ask a question, including slang, regional differences and typos. Lex can take the examples we provide and fill in the gaps – learning to respond well no matter how the user’s message varies.
Our framework has detailed monitoring and reporting, which lets us quickly spot when a user’s message hasn’t been recognised. This feedback means we can quickly train our bots with new phrases, allowing us to continually improve the bot’s performance.
Our chatbot framework can easily be integrated with other services. Want to build a chatbot that lets customers book meetings? We can integrate it with your calendar or booking system. Need the chatbot to search your existing knowledge base? We can build it.
We are integration experts – we’ve worked with dozens of different services and API types. We can make sure the bot works seamlessly with ordering systems, CRM software, contact centres and more.
We can provide regular reports about how the chatbot is being used, including the number of unique interactions and users sessions, the questions customers are sending to the bot, and which messages users are sending which aren’t being handled by the chatbot. This allows us to improve the chatbot over time, and focus the chatbot’s development on the interactions which are most useful.