Natural Language Processing Helps Chatbots Get Smarter

Chatbot

Chatbots have only been around as an accessible technology feature for businesses since early 2015. But, for the business, existing isn’t enough, technology has to move on, do more, and work better. NLP is one of the technologies vendors and providers are banking on to make chatbots more accommodating.

In the non-stop world of technology evolution, chatbots are rapidly moving from script-based design to all-singing AI constructs, enabled with natural language processing skills and other refinements. The likes of California’s Cognitive Code whose Silvia conversational intelligence platform is about to launch or Massachusetts Semantic Machines are all focused on bringing AI and NLP to services, solutions, games and other use cases.

Get On The Chatbot Bus

The truth is, while all of this news sounds very exciting, the huge majority of chatbots still happily use scripts and do their jobs perfectly well. A script can handle the basic interaction needs of any business from a plumber to an IT reseller just as well. So, before investing in these exciting projects, be aware that most businesses can get by perfectly well with a script-based chatbot and perhaps expand their services as the provider slowly moves up the chatbot gears with these new technologies.

One example is SnatchBot, they recently added simple-to-use NLP interactions to their script-based bots, so customers and users can gradually adopt these features, build them into their existing bots without having to worry about a major overhaul. While many companies are only too happy to charge cloud-service rates and growing premiums on volume, SnatchBot is totally free with analytics and publishing across a range of services from Facebook Messenger, websites to iOS apps, making it among the best platforms to use for rapid deployment of a chatbot.

Going Full AI

For brands or businesses that want to explore the full possibilities of AI-based chatbots and agents, the road ahead is an exciting one. Silvia, mentioned above is a local agent who doesn’t rely on the cloud, meaning she can be bolted into offline apps, built into games and so on. Semantic Machines has set itself the task of moving from today’s command-based systems like most chatbots, Siri and Alexa to conversational-based models, allowing people to express complex ideas and for the machine to understand them.

By linking conversation engines and deep learning, we are on a path to better bots, agents and AIs. While there is lots of talk of paradigm shifts and other gibberish (as always when business and technology are involved). Where do you go to see such AI in action? For now, they are mostly limited to enterprise applications or drones and smart home gadgets, as you can see in this tech roundup. When it comes to business, technologies like artificial intelligence and predictive analytics are being used to drive decision-making processes.

NLP is a poster child for chatbots, translation tools as well as appearing in human resources packages to help with hiring the right people or for business knowledge tools. By scanning conversations, resumes or texts, they can extract subtle underlying information to come up with useful information, better understanding, or pointing out a better candidate.

NLP-powered Talla is billed as a knowledge base that thinks for itself.  I mention this company largely because their last blog post was Testing Talla Against Our Own A.I. Bullshit Detector which services as a valuable reminder that a lot talked about AI and similar technologies is still in the realm of hype and not what many typical businesses need.

Regardless of hype when it comes to conversational platforms and chatbots, providers need to focus on the needs of the business, and those of their end users, to ensure the product works at that scale. One fancy use case shown in a demo is unlikely to resonate with any real company working usual business hours and with pressing workload demands.

As a technology buyer, any company needs to look at the products available and find one with proven examples that do a similar job to meet their needs. Many verticals have chatbots dedicated to meet their needs, while successful chatbots also have large example libraries across a range of business types that can help companies create their own bots with ease.

Yes, AI and NLP will be important in the months and years to come, with 2020 likely a tipping point for super-smart AIs appearing in our cars, phones and other devices. But alongside them, less sexy chatbots will be helping businesses save time and money.

Chris Knight

Chris writes about augmented and virtual reality, chatbots, games and anywhere the miracle of technology will take us next.

  • Ashish K Jain

    That true Chris. One of the most exciting things about this chatbot technology evolution is their use of AI  — especially machine learning — to mass-accomplish tasks that are repetitive in nature. It can improve the customer operational support process including a substantial reduction in operational and service costs. We at Engati http://www.engati.com have started this evolutionary journey. Do visit us and provide us with your valuable feedback as we strive for more mainstream acceptance of this technology of the future.