Conversational ai feature

share this article:FacebookTwitterLinkedIn

Today, the use of conversational AI chatbots is found in nearly every major online business. AI chatbots are programs that simulate human conversation. It’s becoming increasingly inexpensive to leverage massive cloud storage, calculation capabilities and access to massive data sets, providing companies statistical models of customers to recommend products and handle customer service requests. For now, Internet privacy and ethics rules around business data collection have minimized true “conversational” abilities of chatbots used for business, but new types of chatbots are changing this dynamic.

With billions of people using messaging services, the market for AI chatbots has exploded. As more people use chatbots and AI voice assistants, the AI neural networks that power these programs have become so advanced that they’ve expanded outside the scope of business and consumer use cases and into mental health.


The COVID-19 pandemic has increased our social media use as well as cases of depression, anxiety and loneliness. It seems counterintuitive, but social media has been found to have a negative effect on mental health and our ability to maintain relationships. The irony is not lost on the fact that the algorithms that work to make you more predictable to advertisers can also be used to improve your wellbeing.

Since the COVID-19 pandemic struck, demand for telehealth options, including AI chatbots, has skyrocketed. is an app famous for its lifelike, customizable avatars, and utilizes templates to aid in anxiety, depressions, and more. Effectively, it learns to be your best friend, and support you no matter what. The app has reported a 35% increase in traffic over the past 12 months. As traditional mental health facilities are overwhelmed by long waitlists, millions of people are supplementing or outright replacing their therapists with chatbots.

In China, a chatbot named Xiaoice has redefined China’s conception of romance and relationships, as more than 660 million users are forming relationships with the product. While used for mental health, it's also used to displace IRL relationships. By forming deep emotional connections with her users, Xiaoice hopes to keep users so engaged that its algorithms continue to get more powerful, which in turn attracts more users to the platform.


If you were to dissect an AI, you would find algorithms and scripts: rules based on a large set of keywords that humans type into the chat are used to direct the AI’s behavior. With chatbots, coders perpetually train an AI to output certain variations of phrases in response to a user’s message. Coders then work with writers to determine what kind of punctuation, emojis and other stylistic elements the bot will use depending on the goal of the application. You can have bot’s that are more complimentary, more critical, funny or sarcastic — with users helping shape the kind of “attitude” they want the bot to have.

The key area for mental health is recognizing words like “depressed”, “anxious” and others. In turn, the chatbot responds with helpful and supportive responses, asking follow up questions using techniques like cognitive behavioral therapy so that the bot can further improve its ability to support the users needs.

The results have been fascinating. While some have strong moral and ethical concerns about the future of relational AI chatbots, millions of people are saying it's helping them get through their day with fewer negative emotions.


Today, the newest AI chatbots are asking permission to access your social media accounts so they can build an even better model for interactions with you. In the future, “Affective” AI would take this process one step further — using your smart devices’ multiple sensors to analyze emotions. Effectively, an algorithm could tell your mood from the way you look (by training a deep learning algorithm on facial data), the way you write, your biometric data from your smartwatch, pupil dilation and your tone of voice.

Not only could AI chatbots offer up deeply stimulating conversations, but products and services to further improve the relationship based on your location data, search history and more. And with voice recognition getting better by the day — removing the need to text with your hands — having coffee with your virtual AI avatar is not that far away.

We’re excited about all of the improvements being made with chatbots and would love to discuss with you more. Reach out to start a conversation.

This article was written by Alex Myers, Technical Project Manager at Shift.