A former director at Facebook is backing several Artificial Intelligence (AI) start-ups in their development of messaging apps that will allow consumers and businesses to communicate via direct messaging, according to a CNBC report.
Anand Chandrasekaran was responsible for the growth of the Facebook Messenger software developer ecosystem, which passed the 300,000 chatbot benchmark under his stewardship between July 2016 and January 2019.
For those two and a half years, his focus was on working with third-party developers to expand Facebook Messenger, telling CNBC in an interview: “We grew the platform to quite a bit of success on the back of these developers.”
But since early 2019 when he left Facebook, Chandrasekaran has instead operated as an angel investor, putting up to $50,000 at a time into tech companies that are still in their early stages of growth.
What is Artificial Intelligence?
Artificial Intelligence, or AI, covers a vast range of different areas, but in this sense it is the ability of a sophisticated computer program to mimic a human and send believable replies to direct messages.
AI chatbots are increasingly used as the first stage in customer service, essentially acting as an interactive Frequently Asked Questions page to solve common queries.
They need to be able to understand natural language and full sentence questions, rather than simple keyword searches, and to respond in a natural, human-like way too.
By doing this they can collate information from customers at the start of an online enquiry, and may either resolve the enquiry with no human intervention required, or pass that information on in a usable way to a human support agent if the enquiry is escalated.
Facebook have been working to increase the interconnectivity and completeness of their platform over the past few months and recent years, with some of the most recent developments including:
- Enhanced Facebook Marketplace shopping and direct purchases from brands and influencers on Instagram.
- Interconnectivity between Facebook Messenger, Instagram and WhatsApp so messages can be sent to contacts on any of the three networks.
- Development of a new cryptocurrency with the intention of using it to pay for goods and services on Facebook-owned platforms.
Chandrasekaran told CNBC: “At the end of the day, the relationship with customers is why people keep coming back to a brand. That’s what drives retention for companies and brands. Automating your messaging experience allows for that.”
Chandrasekaran’s investments have been split between tech firms based in his native India, and new start-ups working on innovative AI applications and particularly messaging platforms.
Some of the brands he has put money into include Avoma, Chatfuel, Fiddler.AI and Verloop, according to CNBC.
These have direct implications for the messaging apps he worked with in his former directorship – Verloop enables AI customer support via WhatsApp, while Chatfuel is a tool to create Facebook Messenger bots more easily.
“I continue to be a huge believer in those platforms,” he explained to CNBC.
As AI helper bots become increasingly sophisticated and able to resolve queries first time with just a few interactions, Chandrasekaran argues that more and more consumers will choose this method, rather than wait on hold to speak to a human helper by telephone.
He concluded: “People messaging businesses and businesses messaging back is going to become very popular.”
The other social network
While Chandrasekaran’s investments focus on Messenger and WhatsApp, the ‘other’ social network has proved to be less of a safe space for chatbots.
On March 23rd 2016, Microsoft launched Tay, an AI chatbot, on Twitter. The bot was designed to learn from the human users it interacted with; however, it soon began to repeat inflammatory language it had learned from those early tweets.
After less than 16 hours and less than 100,000 tweets, Microsoft deactivated Tay. On March 30th the account was accidentally set active again during testing, and tweeted several drug-related messages before again being taken offline.
Tay remains as a profile on Twitter but is now a ‘protected account’, which means any new followers must be manually approved before they can see the bot’s tweets.
As of September 2019, Tay had more than 125,000 approved followers.
Twitter AI bot rules
In more recent months, Twitter has consistently taken action to shut down bot accounts linked with political campaigning and trying to manipulate voter behaviour, among other things.
While automation is allowed on Twitter, it has several rules that govern the kind of interactions that are considered appropriate for bot accounts.
Tweets should not ask for too much personal information, and should direct customers to a secure platform such as the company’s own website if credit card details are needed.
Messages should not be duplicated across multiple accounts, should not spam replies or trending hashtags, and should not ‘surprise’ users by functioning differently than expected, for example by posting a tweet when a human user clicks on a link or button.
However, the rules encourage the creation of helpful bot accounts, saying: “Do build solutions that automatically broadcast helpful information in tweets; run creative campaigns that auto-reply to users who engage with your content [and] build solutions that automatically respond to users in direct messages.”
Disclaimer: The information provided here is not investment, tax or financial advice. You should consult with a licensed professional for advice concerning your specific situation.