In this special guest feature, Dmitry Gritsenko, CEO of the Master of Code Global, suggests that as interest in the commercial use of conversation design continues to rise, it’s a great time to look at the technologies that are making all of this possible as well as its impact on users today and in the future. Dmitry offers expert content on the biggest AI trends of 2022. And his company partners with the world’s leading brands to design and develop conversational AI experiences across mobile, web, and AI-powered chatbots.
Conversation design and artificial intelligence have a long and storied past but it’s only in the last six years that they have become cost-effective enough for small-medium-enterprises (SMEs) to deploy them. As interest in the commercial use of conversation design continues to rise, we think it’s a great time to look at the technologies that are making all of this possible as well as its impact on users today and in the future.
What is Conversation Design?
There is a lot of ambiguity around the term Conversation Design so let’s address that first. Just a few years ago, building a chatbot (or any other form of virtual agent) meant two things: first, writing the application itself, and second, writing hundreds and thousands of responses that the chatbot would use to communicate. Usually, both of these tasks were handled by the same person/team.
However, with recent advancements in machine learning and particularly natural-language understanding (NLU), developers no longer need to write thousands of lines of code themselves, they simply need to point the machine learning engine to a suitable dataset and it stores all of that information in a “knowledge base” to use in the future.
Now since the chatbot already has all the information, developers are only concerned with ensuring it uses the information in a way that satisfies the end-user. This has led to the re-emergence of a field called conversation design.
At a high level, conversation design is the process of teaching computers how to communicate more like humans. Designers do this by applying a series of technologies, principles, and best practices.
How Machine Learning Has Changed Conversation Design
We’re using machine learning in the broader sense of the term as there are numerous distinct technologies involved in conversation design. That said, the most important of these technologies is the language model which is responsible for Language Understanding, Dialogue Management, and Language Generation.
Usually, the bigger the model, the better (of course, there are nuances to this) but this is a common rule of thumb. That is why conversation designers are seeking larger language models, to make virtual assistants more comprehensive. A few popular language models are BERT (developed by Google) and OpenAI’s GPT and GPT-2 models. In addition to being extremely versatile, these NLP models are also “pretrained” which saves a significant amount of time and effort.
To fully utilize the greater processing power available to designers today, they are adding additional variables such as bot personas, user intents, and contexts that give chatbots more human-like speech. Some of the things made possible with these variables include:
- using the knowledge base, context, and intent to triangulate answers to new/unique questions
- handling two and even three part-questions (clauses)
- taking into consideration dialogue history and user preferences
- identifying the correct intent as well as when the user intent has been satisfied
Together, these capabilities allow chatbots to understand an exponentially wider variety of questions while taking an active role in the conversation, that is not waiting around for the user to give instructions.
All of this brings us to the actual “creation process” of building chatbots and that too, has changed for the better. Designers can use these language models on their own but due to the massive commercial potential tied up in building conversation flow, major cloud-based NLU platforms have emerged.
The fact that every single major cloud vendor (and Facebook) is pouring significant resources into conversation design is a good indication of the commercial value it provides, which brings to our final point of discussion – the current and future commercial uses of conversation design.
Making Robots More Useful
Right now, the most popular implementation of conversation design is in customer-facing virtual assistants or chatbots. According to a 2021 report, the chatbot industry will be worth $10.5 billion by 2026 at a CAGR of 23.5%. This is partly because being able to offer personalized and instant customer service every hour of the week is a tremendous advantage, especially for companies that cannot compete on price or choose to focus more on customer loyalty.
That said, we do believe that the commercial uses of conversation design are not limited to customer-facing positions. In fact, the technology is already more than advanced enough for applications in other areas of internal business operations such as:
- Virtual assistance: An AI chatbot would make getting information about company news, project updates, and general alerts much simpler and faster for everyone, from the factory floor to the C-suite.
- Interactive employee handbook: A conversation design can be used to replace traditional employee handbooks with an AI chatbot that is easier to use, engaging, and accessible from any smartphone.
- Document management: Document management is one of the most mundane and time-consuming tasks in any company that can be improved significantly with AI chatbots that can find, share, and receive documents with a simple voice command.
- Business intelligence: Translating live progress, market updates, and overall company performance from quantitative figures into natural human speech would save decision-makers a lot of time while at the same time, making data even more accessible.
Wrapping up…
All speculation aside, two things are certain, the advancement and accessibility of NLU technologies are making conversation design a critical business investment for modern companies. And second, consumer behavior trends shifting towards personalization and increased AI automation in businesses are both strong indicators of a future where robots are more widespread, smarter, and helpful.
Sign up for the free insideBIGDATA newsletter.
Join us on Twitter: https://twitter.com/InsideBigData1
Join us on LinkedIn: https://www.linkedin.com/company/insidebigdata/
Join us on Facebook: https://www.facebook.com/insideBIGDATANOW
Speak Your Mind