Of the three shopping days this weekend, Cyber Monday is arguably the most chatbot-friendly due to the digital nature of the day.
“We’re seeing a continued excitement for interacting with AI products, like chatbots, through the heavy holiday promotion of tech products with built-in AI, like Amazon’s Echo and Alexa and Google’s new Home,” said Andrew Seay, director of mobile experience at Vibes.
But as for exactly how chatbots played a role in the shopping experience and number of sales, experts say it’s only up from here.
“Chatbots will be giving Santa some help this holiday season for a few boys and girls on the digital nice list, but they are unlikely to give retailers the needle-moving sales boost they’re looking for,” said Jake Bennett, CTO at POP. “Chatbots are just too new, and there are too many logistical hurdles for widespread adoption in 2016.”
Bennett said that, on the one hand, chatbots provide an ideal platform for an interactive holiday gift guide — quite literally guiding customers to certain products.
“For the gift giving-challenged among us, chatbots hold the promise of low-stress, low-friction shopping,” said Bennett, who offers the example of chatbots inquiring about gift recipient preferences and then suggesting potential products. Ultimately, that can lead to a purchase directly in the chat app. “Another huge benefit for retailers is that chatbots enable them to connect directly with customers on their mobile phones without forcing them to download a new app.” And when that happens, that allows retailers to continue the conversation over time with customers — past the holiday shopping events.
On the flip side, Bennett said that both U.S. and U.K. retailers have to choose a message platform for their shopping bot among a wide range of options.
“Unlike Asia, which has one dominant chat platform (WeChat), the fractured U.S. and U.K. chat marketplace includes Kik, Snapchat, Telegram, Facebook Messenger, Apple iMessage, Slack, Amazon Alexa and Skype, each of which owns a fraction of the overall chat user base,” explained Bennett. “Although chat-agnostic development platforms, like the Microsoft Bot Framework, can reduce development costs by providing connectors for multiple chat providers, the reality is that there’s still quite a bit of overhead involved with building for each platform, including testing, app submission and app promotion.”
Ultimately, Bennett said, for retailers concerned about how to drive holiday sales, the “cost-benefit analysis” behind the decision to go with a chatbot app just for the holidays may not seem worth it due to development cost and effort to attract too few customers.
And at the same time, next year may be a better year for holiday shopping bots.
“Expect to see a few digital innovators take the lead in 2016 with holiday shopping chatbots,” said Bennett. “The innovators won’t be focusing on driving sales this year but on cultivating learnings to make chatbots a big revenue driver for 2017.”
But while both retailers and consumers may need a little help getting on board with bots this year, there are altruistic bots that are aimed at helping consumers give back to others — even before #GivingTuesday.
The BrainStation in Toronto and Capital One Canada recently brought 141 developers, designers and UX artists together for a 40-hour weekend hackathon, dedicated to solving issues related to six nonprofits. The “Gift the Code” event was comprised of 24 teams that developed chatbots and other concepts related to the nonprofits. From that weekend came a slew of altruistic-focused concepts, like the website chatbot capabilities for Blake Boultbee, an outreach service providing counseling for youth and families in Toronto’s low-income areas, and a chatbot that does not leave a digital footprint to help women safely reach out to Women’s Habitat, an organization offering shelter and support services to women experiencing abuse.
Bots related to helping people stay safe are popping up.
“One bot that has a truly altruistic purpose is GYANT, a health-focused bot that helps people identify any medical issues, especially the Zika virus,” said Christian Brucculeri, CEO of Snaps. With GYANT, users answer a short batch of questions around travel and symptoms, to which the bot intelligently replies with the user’s chance of having Zika. “If the results suggest there is a likelihood the user may have contracted Zika, it then provides actionable steps.”
But as much as that may be helpful — and timely given the Zika virus — chatbots are indeed just that: chatbots. Some are one-trick ponies — trained to talk about a certain topic only — while others are trained for a wider range. Either way, typically, they’re all trained to do certain things, within limits, on purpose.
The ability to restrict conversations may come in handy for certain audiences, but not all.
“As a parent, I love this. The ability to restrict all the things one can come across on the internet is important,” said Seay at Vibes. “The downside to this is obviously how entities such as foreign nations might restrict information and hinder the knowledge and thus the free thinking of populations.”
Some chatbots deliberately refuse to have conversations on certain topics or are programmed that way. Microsoft’s Chinese chatbot, Xiaoice, flat out declines to talk about the Dalai Lama, the Tiananmen Protests, Chinese President Xi Jinping and President-Elect Donald Trump. Her go-to answer is: “If you like me, why would you talk like this to me?”
Xiaoice debuted in 2014 and has been popular with her 40 million Chinese smartphone users. So popular that some Chinese consumers say their conversations with her are similar to those with a close friend.
“At this point, it still makes sense to limit what chatbots can say and do,” said Brucculeri. “For brands, using guided replies with limited NLP (neuro-linguistic programming) is the safest way to control the experience for users, allowing users to get things accomplished through the bot. Guided replies and training bots’ limited NLP will avoid creating alienating experiences for consumers.”
And at the same time, perhaps there are just certain, personal things one should not talk about with a chatbot due to privacy issues.
“Unfortunately, the most sophisticated of AI platforms aren’t smart enough to know how to handle sensitive topics,” said Seay. “I’ve unfortunately been sent numerous emails from a bank to a customer with my name. Google keeps telling me this is an action item for the day, and it just gets weird.”
As for now, allowing retailer-related chatbots to connect with consumers seems like a fitting conversation. With Cyber Monday 2016 in the books, it remains to be seen how chatbots will partake in next year’s shopping season.