Intelligent CIO Africa Issue 65 | Page 39

CIO OPINION was designed to answer customer questions , and to do it in a conversational way – just like a human agent would . The real breakthrough with these info bots was their ability to converse in natural language and to learn from a company knowledge base and ongoing customer responses . This allowed them to improve their ability to predict the right information to answer a specific question .
For most customers , these info bots were positioned poorly . Rather than being offered as a ‘ smart librarian ’ that could help find answers to specific questions , they were positioned as the digital go-to for customer sales and support queries . This resulted in customer ’ s engaging with chatbots as if they were virtual agents and being bitterly disappointed . What they found was a very limited virtual agent that seldom offered the right level of information , let alone got any query resolved for them . And as a result , all roads still led to a live agent , either via live chat or a call back .
To enhance these first-generation virtual agents ’ capabilities , companies tried to get their info bots to perform actions , not simply find information . To do this , they upgraded them to transaction bots that could work directly or indirectly with relevant back-office systems . This meant that basic transactional requests like ‘ top up my data by 100GB ’ could be handled by the virtual agent . If they needed more information , they could surface a web form , and could then pass the information through to the relevant systems for processing . This worked as long as the customer had a clear instruction . Unfortunately , they soon found that most customers have contextually rich service queries , issues or complaints . They want help first working out what to do .
And as a result , most digital engagements with these virtual agents still ended up with a live agent .
This situation is changing with the emergence of service bots , designed to handle all known queries , issues and complaints without having to hand over to a live agent . These service bots are far more capable than the early generation info and transactional chatbots . They can handle the complex advisory-level logic required to clarify , analyse , resolve and action customer service queries , in context and in line with business and regulatory rules .
As a result , these third-generation virtual agents are reducing the calls coming through to human agents .
The simplest form of virtual agents is a chatbot , or info bot . This early generation of front-office digital worker was designed to answer customer questions , and to do it in a conversational way – just like a human agent would .
www . intelligentcio . com INTELLIGENTCIO AFRICA 39