Conversational AI is a form of Interactive Artificial Intelligence that enables a humanlike dialogue between a person and computer system. It could be a customer service chatbot, home control system, assistant for vehicles, Employee self service etc… It’s found across multiple channels such as web, social media, mobile apps, telephone and smart speakers using text or voice.
There are many aspects to a conversation that we use without realising it. We use memory, context, disambiguation, language, slang and tone without thinking. For a computer, human interaction is very complicated.
Try having a conversation with a friend without using your memories of them?
Think about this sentence: “I saw a dog on a hill with a telescope”. Can you spot the 5 different meanings?
Understanding what someone means is only part of the challenge. In the business context you also need to respond or take action. A human agent has access to other systems to make decisions, retrieve or update information. All of this is impossible to automate without the right platform and Conversational AI expertise.
If you get it right the potential return on investment (ROI) is huge. But if you get it wrong, performance targets (KPI’s) will not be met, customer satisfaction (CSAT) will be low, and eventually the business will shut it down.
The entire landscape usually involves a Conversational AI platform such as Microsoft Luis, Teneo, Google Dialogflow, Cognigy etc to learn and manage user intent recognition, build the dialogues and integrate with data sources. Then there is the technology of the chosen channel(s) for the dialogue. This is combined with machine learned models and/or linguistic rules to understand the user through the entire interaction.
There are many technical resources across the globe who can write lines of code to connect the platform and implement a scripted conversation. However, the art of Conversational Design requires Computational Linguists who passionately understand the construct of language and automated conversational techniques. This is needed to provide demonstrable value to both parties in the conversation. It is very, very hard to find experienced Computational Linguists who have worked extensively with Conversational AI.
To be successful, a CAI solution must:
- Focus on topics that drive the most value for the business, and carry a conversation to completion without human intervention.
- Provide a User Experience that is more awesome than traditional channels. In this context, User Experience includes a wide number of areas, including look & feel, text/content (think tone of voice / brand style), intelligence of the CAI solution, dialogue design, etc.
- Be able to complete the user request by integrating the solution with data sources and transactional systems. Examples are customer sales records, order entry systems, delivery status, etc. Otherwise this amazing technology is just a sign-post, or mildly improved FAQ web page!
- Capture and easily report on dialogue data for measurement, monitoring and improvement.
- Provide an ROI that warrants meaningful corporate commitment from the top down.
- Be built in an agile way that can scale and adapt based on real user feedback and business events.
- Focus on continuous KPI driven investment and improvement.
The world is short of Conversational AI experts; these are people who are educated and experienced in Computational Linguistics and Data Science. They have a flair and a passion for the construct of automated human language interaction. Consequently, many thousands of projects fail, do not deliver on the AI promise and leave a business unimpressed with the results.
With the right expertise, these challenges can be overcome to provide a Conversational AI solution that excels and delights the business and their customers. This is where The CAI Company comes in. We provide highly experienced senior resources that will drive successful results from a new or revitalised Conversational AI solution.
