The chatbot boom began in earnest one year ago, when the most popular messaging platform launched its newest gamble: Facebook Messenger Bots. However, the origin of chatbots dates back to the 60s, when the first chatbot, Eliza, was created. Since then, bots have been used in the customer service arena. But this Facebook launch was really when chatbots began to proliferate across the internet.
Today they are so popular, that according to a 2016 Imperva report, more than 50% of Internet traffic comes from bots. Fully half of these are “bad” bots.
Virtual assistants, also called chatbots or bots, have increasingly become a tool employed by companies to interact with their customers. The benefits are clear. By using Artificial Intelligence, a bot does not require full-time human resources. In turn, companies can automate a large part of their customer service, while helping customers in real time, night or day.
But it is also worth noting the problems that can occur when this technology is applied in less than optimal conditions. A company’s enthusiasm to innovate and save money can quickly turn to frustration and headache if the necessary precautions are not taken before launching a chatbot. Users are also excited when they start to chat with a bot, but when they do not receive expected answers or if the bot turns into spam, the fun evaporates. In a matter of seconds users form a negative opinion about the company, leaving a permanent digital mark.
While virtual assistants’ potential is indisputable, they must be used carefully. This is especially true when companies consider adopting virtual assistants for customer service, the most bot-friendly segment. A chatbot must be useful and transactional, i.e. it must work correctly and be able to help the the user. At a minimum, the bot must be able to solve customers’ most frequently asked questions and provide real-time solutions. If it can’t do this, it is just a simple command bot—a bot that responds to a structured flow of limited questions and answers. The main difference between these command bots and a transactional bot is that command bots do not use Artificial Intelligence in customer service.
There are a set of characteristics that bots must have to understand customer intent. This requires an initial effort on the part of the company: the bot must be to equipped with knowledge, it must be trained to understand customers, and it must be integrated with the company’s web services or CRM. Only after this process is complete can you reap the benefits: improved customer satisfaction and experience, and increased sales and reduced costs for the company.
To nip problems in the bud, find out if a virtual assistant is truly intelligent. Ask the following questions before incorporating a chatbot:
- Is this the technology that your company needs?
- Is the chatbot provider a customer service specialist?
- Does the provider have experience in your industry? Do they have clients in your industry?
- Can you see live, effective cases?
- Can you test a customized functional demo?
- Does the virtual assistant have a memory state to hold long conversations and understand the context?
- Can the virtual assistant learn from the user and evolve?
- What resources and time are needed to implement it?
- What analytical data can be extracted?
- Can you integrate the bot with other third party solutions, systems and technologies?
If your provider cannot answer these questions… keep looking!
This article was used with permission from Aivo – http://aivo.co/en/
Link to the original article – http://blog.agentbot.com/en/10-questions-to-ask-before-your-company-adds-a-chatbot