{"id":15193,"date":"2018-03-15T16:41:21","date_gmt":"2018-03-15T16:41:21","guid":{"rendered":"https:\/\/saffroninteractive.com\/\/?p=15193"},"modified":"2021-08-20T16:38:55","modified_gmt":"2021-08-20T16:38:55","slug":"learning-chatbot","status":"publish","type":"post","link":"https:\/\/saffroninteractive.com\/learning-chatbot\/","title":{"rendered":"Why a learning chatbot might soon be your closest colleague"},"content":{"rendered":"

We\u2019re all painfully aware that artificial intelligence<\/a> is poised to revolutionise the world of work. Learning and performance, HR, IT\u2026 every department is set for a machine-led overhaul. As the pace of change only continues to speed up, adapt to survive becomes no longer just a biological imperative, but a technological one.<\/p>\n

<\/p>\n

No business will be able to remain competitive without the power of machine learning<\/a> and all the insights it can generate from the huge amounts of data we\u2019re accumulating. One survey<\/a> of over 350 AI experts predicted that AI could even be writing a bestselling novel by 2049 (and who knows, maybe I might even have finished the first chapter of mine by then).<\/p>\n

What will the outcome for people be? And how does a business need to respond? In a world where things change so quickly but humans don\u2019t, learning chatbots could be at the forefront of these transformations. But not in the way you might think.<\/p>\n

AI\u2019s ever-evolving capabilities will see it shift away from fulfilling esoteric use-cases and behind-the-scenes content curation towards something more recognisably adaptable, more recognisably creative \u2013 more recognisably human.<\/p>\n

Picking up chatter<\/h2>\n

But before we get ahead of ourselves, we should probably consider the definition of a chatbot. Broadly speaking, a chatbot is a service powered by rules, code, and often a sprinkling of AI, that you interact with through a chat interface. These interactions can serve a variety of purposes, whether transactional, entertaining, or informative.<\/p>\n

The thinking behind chatbots isn\u2019t new. Ever since computers surfaced midway through the 20th<\/sup> century, people have been envisioning having heart-to-CPUs with them. Closely linked to early imaginings of AI, the idea of a talking robot has been culturally ever-present. From the sinister HAL 9000<\/a> and his Simpsons reimagining<\/a>, to KITT in Knight Rider, to the recent intimate anthropotechnic relationships seen in Her<\/a> and Blade Runner 2049<\/a>, we\u2019re fascinated with the idea of digital sentience and our interactions with it.<\/p>\n

Perhaps the earliest example of attempting to make these ideas a reality was proto-chatbot ELIZA<\/a>, built in the 1960s by MIT professor Joseph Weizenbaum to hold simple psychotherapeutic chats with users. Think an awful lot of \u201cHmm\u2026 tell me more.\u201d<\/p>\n

Later, as things began to go digital, we started seeing desktop escapades of primitive chatbots being found in programs such as MSN Messenger, as the Gen Y\u2019ers amongst you might recall. They then started to appear on the web, with chatbot services for weather forecasts or talking to Miss Piggy<\/a>. Things went a little quiet as everyone\u2019s attention became focused on the mobile app rush, but it seems that we\u2019re finally at the point where the chatbot wave is ready to break.<\/p>\n

Riding the chatbot wave<\/h2>\n

The technology is finally catching up. AI is now reaching a high-water mark, one where chatbot interactions can be meaningful enough to hold real Turing-standard conversations, rather than just illusory interactions delivering pre-programmed responses.<\/p>\n

Our everyday tech is now much more capable of hosting a chatbot, too. They now permeate digital society and are available to the general consumer. You\u2019ve probably already got one \u2013 or several \u2013 in your pocket. Most of us are familiar with Siri (or equivalent), who\u2019s able to answer the vast majority of practical questions we ask, with that ratio only increasing every day.<\/p>\n

We take it for granted, but having an AI assistant in your pocket is actually pretty futuristic stuff. There\u2019s no doubt these chatbots are going to play a greater and greater role in how users interact with technology and proliferate further still. Indeed, ask Siri if it has kids, and it answers “Artificial intelligence does not have the capability to reproduce. (long pause) Yet.” Creepy.<\/p>\n

The imminent ascension of our robot overlords aside, the spread of chatbots isn\u2019t just limited to phones. Think about the recent surge in sales of the Amazon Echo or Google Home. These are basically internet-of-things-enabled chatbots that are given a permanent fixture in your home.<\/p>\n

That\u2019s not to mention the growing population of chatbots in the digital communications channels we use every day. Channels like Facebook Messenger, Skype, and Slack are all overtaking traditional social channels<\/a>, and chatbots allow organisations a presence in them.<\/p>\n

We know that consumer trends such as mobile access, curation, and personalisation all trickle down to the business world, and this one will be no different, with chat becoming a more and more important way for organisations to communicate both externally and \u2013 importantly for learning and performance \u2013 internally.<\/p>\n

So the point is, chatbots are all over the place, soon to be everywhere. The pressing question is: how can we use them in business, and particularly in learning, to facilitate change?<\/p>\n

\"hand<\/p>\n

Chatting down to business<\/h2>\n

Businesses are using chatbots to keep up with their consumers. Learning and performance need to use them to keep up with their learners \u2013 and with the business.<\/p>\n

We\u2019ve already seen the huge potential in customer-facing support roles. Would you rather go through one of those agonisingly slow monotone \u201cPress 1 for all queries related to the exact opposite of yours\u201d or get a direct answer to your question? The same applies to internal learning and performance support.<\/p>\n

Don\u2019t make your users trawl through endless piles of \u201cresources\u201d to no avail, then end up turning to a busy colleague or a search engine. We need to get rid of Google<\/a> and provide our own point-of-need responsive support to learners so that they can access standardised organisational knowledge on the job.<\/p>\n

Whether that\u2019s through performance support tools<\/a> or chatbots, the key is to reach the learner at the point of need<\/a>. That means workflow integration, presence in internal social channels, and constant availability. That doesn\u2019t<\/strong><\/em> mean endless suggestions of microlearning content that will never get read, or sitting users down at the end of the day to plough through a two-hour course.<\/p>\n

Workflow integration, presence in internal social channels, and constant availability. Ring any chatbot-shaped bells?<\/p>\n

There are already some great examples of performance chatbots in their early stages out there.\u00a0Jamie Good\u2019s<\/a>\u00a0SMS-based chatbot, for instance, uses natural-sounding conversations to provide conference attendees with key takeaways and reminders to act on them.\u00a0 There\u2019s also Woebot<\/a>, a Messenger-based chatbot that\u2019s having great success helping to shore up mental resilience (admittedly not just in a workplace setting). We\u2019re starting to see digital learning providers wake up to the potential, too, but they\u2019re a long way from fulfilling it\u2026 although watch this space for our own!<\/p>\n

Businesses are using chatbots to keep up with their consumers. Learning and performance need to use them to keep up with their learners \u2013 and with the business.<\/p>\n

Tweet
\n<\/a><\/p><\/blockquote>\n

The buzz of conversation<\/h2>\n

Current learning chatbot offerings only function according to pre-programmed rules and limited AI. They\u2019re essentially glorified search engines. Their responses and your input are limited, and the bot is only as smart as it\u2019s programmed to be. We need to bring in machine learning to forge a chatbot that really understands your learning and your workflow, not just your commands.<\/p>\n

With this kind of learning chatbot, we could provide not only instant, constant feedback when a user asks for help, but also pre-emptive guidance based around their current task before they even start it. That kind of point-of-need technology is available<\/a>, but the next generation of chatbots will be able to provide it as part of an overall virtual assistant package, and to a greater degree.<\/p>\n

Consider an AI assistant that really understands<\/em> our current work process. One that can can also access and analyse all the reams of data from previous users\u2019 actions and results to make its own predictions and suggestions. That could revolutionise decision-making support.<\/p>\n

It would also be far more agile than the current learning trope of prescriptive content that\u2019s out-of-date almost as soon as it\u2019s written. Adaptive machine learning could keep up with situations where the requisite workplace skills and information are constantly changing. That\u2019s going to be more and more often.<\/p>\n

Package that all with advanced speech recognition and conversation simulation and learning chatbots could really make the experience more human and engaging. We know L&D departments struggle with internal perception and achieving the status of trusted advisor. Imagine if they were represented across the business by an avatar that people found truly helpful. That people actually wanted<\/em> to engage with.<\/p>\n

We\u2019re getting into the realm of AI taking over everyone\u2019s jobs again, but you get my point. The potential here is huge. I\u2019ll leave it there by going by going back to Spike Jonze\u2019s wonderful Her<\/em>.<\/p>\n