Artificial intelligence, big data and natural language processing (NLP) have been hot topics for a few years now. It’s understandable, as we are currently in the middle of an AI and machine learning boom. Chatbots have been around in one guise or another for quite some time, but have often failed to deliver. Unable to mimic human behavior and conversation patterns convincingly, they discouraged interaction. Recent advances in machine learning and natural language processing (NLP) techniques have allowed developers to build engaging real-time conversations into chatbots that can be deployed successfully in different use cases, yielding a whole new generation of automated customer engagement tools.
Most importantly, advances in NLP have allowed for a jump from primarily text-based to voice-based applications, opening a world of new opportunities, and greatly expanding the capabilities of this technology. In this article, we look at how chatbots have evolved, where automated customer engagement is headed, and how you can take advantage of this technology to optimize your business.
What are chatbots?
A chatbot is a communication channel that users interact with, most commonly by typing as if chatting over an instant messaging terminal with a human. The chatbot receives human responses, analyzes and interprets what the human said, and returns an appropriate response based on what’s programmed into its chat flow.
Evolution of chatbots
Historically, chatbots ran off limited canned scripts, and it didn’t take much for humans to realize they were communicating with a machine and either give up, or purposefully throw the chatbot off the “script”.
Today, advancements in AI, NLP and web development have allowed for rich applications. Products such as Microsoft Cognitive Services can be provisioned in the Cloud, making it easier than ever to build chatbot solutions that don’t run off a script and can mimic humans in a way that’s almost indistinguishable. This has paved the way for the development of “conversation as a service” solutions that are helping businesses realize many previously untapped benefits.
Earlier this year, Google released Meena, a conversational agent engineered to mimic human behavior. Meena conducts conversations unlike any previous chatbot due to its improved human evaluation metric called Sensibleness and Specificity Average (SSA). This metric analyzes important attributes in natural conversations. In a recent test, Meena’s SSA performs nearly the same as humans, meaning that Meena is capable of replicating the nuances of human speech.
Currently, Meena is being tested for safety and bias, so it is not readily available to the public – yet, although Google promises that this multi-turn, open domain chatbot will be the best one. They expect that small and medium-sized businesses – particularly service companies – will be the biggest beneficiaries of this technology.
Meena Sensibleness and Specificity Average (SSA) compared with that of humans and other open-domain chatbots
During its keynote in 2018, Google introduced Google Duplex as an AI system that books appointments over the phone on behalf of the user. The AI system is a recurrent neural network built on TensorFlow Extend, which allows the voice to understand and interact based on context. To sound more human, the system combines concatenative text to speech engine and synthesis TTS engine to add rich human mannerisms such as filler words and pauses. Current users can use Duplex within Google Assistant on Android or iOS to make reservations or schedule appointments.
The distinction between a human voice and the Google Duplex voice is shrinking rapidly. Even compared to the initial announcement in 2018, the voice today is less robotic and much more convincing. In a study from The New York Times, phone calls using Google Duplex to book reservations in local restaurants sounded eerily human. Ethics questions have arisen about the deception of robots appearing as humans. Google promises to be transparent by having these phone calls appear as “Google Assistant” on the Caller ID.
These tools – both established and insurgent – will continue to revolutionize transactions between businesses and their clients. Let’s take a closer look at how chatbots can help you transform your business.
Reduce Operation Costs
Gartner predicts that by 2024, managers’ workload will be reduced to 69% of current levels, due to advancements in artificial intelligence. If you’re a business looking to leverage some of the benefits that chatbots bring, one of the most immediate is reducing operational costs. This can take a few forms due to the flexibility and wide range of scenarios where chatbots can be deployed.
For example, if you are responsible for answering support tickets from customer requests, a chatbot can be designed to take the human’s query, parse out any common words and use these to query a back-end database of common solutions. These kinds of scenarios free up the human workforce to focus on more value-added tasks.
Improve Customer Engagement
Businesses are going to great lengths to drive conversion rates on their websites, but surfing the web remains a largely passive activity. Think of a common case scenario: a person searches for a service, finds a potential match, and may or may not decide to use the service, but doesn’t feel like taking a definitive step of contacting a business at the time.
When done well, chatbots can help businesses improve customer engagement and increase conversion rates, turning more passive visitors into customers. By asking visitors what they’re looking for after they land on your company’s web page, you’re engaging them in a helpful and non-intimidating way. Your chatbot can ask questions, advise which products or services would be best suited to the prospect, and suggest additional resources that a visitor might explore.
Structured Data Collection
Given their Q&A nature, coupled with advanced text analytics and natural language understanding, chatbots are perfect vehicles for your business to gather structured data. This information can be saved away in back-end databases, and more importantly, reports can be generated to surface valuable, actionable insights to help you make more informed business decisions.
For example, to help tighten up the data humans send your chatbot, the Microsoft Bot Framework comes with the concept of “cards”. These cards are effectively commands that have accompanying text representing options the user can click or tap (if on a mobile device). Think of this as selecting from a multiple-choice question or a drop-down menu. What this means is that as a business, you can develop a chatbot that will only accept specific inputs, making it much simpler to enforce the collection of structured data.
In the early days of chatbots, they were typically restricted to running from command prompts or terminals. As of Q4 2019, over 50% of users surf the web and access products and services using mobile devices such as smartphones or tablets. You’d be forgiven for thinking chatbots might not be suitable for mobile channels, and consequently are unsuitable for your business.
In the last 2-3 years, we’ve seen the emergence of platforms and development solutions such as the Microsoft Bot Framework that allows you to build your chatbot once, and then easily deploy to several channels that include, among others:
In this post, we’ve explored how chatbots have evolved, and how they can help optimize your business, by reducing operational costs, improving customer engagement and making reporting more efficient. We hope that this article provided insight into chatbots and technologies that underpin them.
Here at Growth Acceleration Partners, we have extensive expertise in many verticals. We can provide your organization with resources in the following areas:
- Software development for cloud and mobile applications
- Data analytics and data science
- Information systems
- Machine learning and artificial intelligence
- Predictive modeling
- QA Automation
If you’re ready to add a chatbot to your site, download our Chatbot Design Planning Template here. If you have any further questions regarding AI or automated customer engagement models, please reach out to us.
About Roy Cruz
Roy Cruz is a senior developer, data engineer and technical lead at Growth Acceleration Partners, and a leader within our Data Analytics Center of Excellence. With nearly a decade of experience, Roy builds robust web applications in both front and back-end development and creates data pipelines knowledge in Big Data. He holds a B.S. in Engineering Systems and Informatics from the Universidad Hispanoamericana in Costa Rica. You can connect with Roy on LinkedIn, or send him an email.