• Freddie McMahon, Chatot Thought Leader at DF2020


• 24.07.2018 10:00 am
undisclosed

Conversation-as-a-Service is the co-existence of chatbots and humans to deliver value through dialogue. 

A core consideration is the Conversational User Interface (CUI), which has three purposes: 

1. Orchestrate dialogue between the machine and the human 
2. Orchestrate microservices for processing events, tasks or processes 
3. Orchestrate handover for human-to-human interaction 

The CUI is an omnichannel deployment, which means the conversation can occur across all digital touchpoints that support text or voice dialogue. The conversation may continue across different touchpoints until an outcome has been reached using any combination of interaction involving: 

1. Text
2. Tap
3. Talk 

The CUI orchestrates the three types of dialogue: 

1. Natural Language Processing 
2. Single Question, Single Answer 
3. Scripts 

Natural Language Processing (NLP) and Single Question, Single Answer are underpinned by machine learning, which is suitable for some things, but not everything. 

NLP is not suitable for conversation-as-a-service in three areas: 

1. Moral & Ethics, which is a complex and sensitive subject: machine learning must not be empowered to determine say politically correct dialogue as this would be unacceptable to most corporations.  
2. Regulatory and Statutory, which is a complex and algorithmic subject: machine learning must not be empowered to rewrite or interrupt the law unbated, which would be unacceptable to all corporations.
3. Standard Operating Procedures, which is a complex and algorithmic subject that can be linked with Moral / Ethics and Regulatory / Statutory matters: machine learning must not be empowered to rewrite, or interrupt say policies unbated, which would be unacceptable to all corporations. 

Though these areas are not suitable for NLP machine learning, they are suitable for scripted dialogue, which is under human controlled learning. 

Conversation-as-a-Service is any combination of the three types of dialogue in context to the need and the controls required by a corporation. Significant benefits can be achieved through dialogue using a blend of machine controlled learning and human controlled learning to: 

1. Engage customers 
2. Empower employees
3. Optimise operations
4. Transform products 

The CUI orchestration of microservices that typically exists include: 

1. Real-time language translation 
2. Real-time text-to-voice-to-text
3. Conversational payments 
4. Adaptive Cards / Skills to perform specific tasks 
5. Robotic Process Automation 
6. Blockchain e.g. for a record of compliance automation 
7. Etc 

The CUI orchestration for a handover to a human can be during the chatbot conversation or triggered by a dialogue event or triggered by a dialogue outcome. 

The above can be achieved today with the right mix of technology.

Today's challenge is more to do with the need for the following:  

1. New business models due to the nature of the digital transformation, which is a revolution in interaction. 
2. New framework for digital governance and compliance automation. 
3. New roles and skills such as a conversational architect and conversational designer. 
4. New types of business intelligence such as accommodating the shift from lag indicators to lead indicators.
5. Treating knowledge as a working and measurable knowledge asset, which is a game changer in the world of quality management (such as ISO9001: 2015 edition), intangibles and their impact upon the balance sheet and the valuation of a business. 

These conversation-as-a-service challenges have left a vacuum that needs to be filled by business gurus, management strategists, consultants, quality accreditation firms, digital agencies, academia and training firms to take a leadership position, which is at the core of the revolution in interaction that has the potential to impact every person and every organisation across the world. 

This article originally appeared at LinkedIn