Chatbots - are they all talk?Simon Blake, October 10, 2017
If you’ve ever had a product suffer a fault, does the prospect of calling a customer support line fill you with hope… or hopelessness? I think we’re agreed the concept doesn’t appeal to most people. But this might be changing, thanks to artificial intelligence (AI).
One of the most practical and visible use cases for AI is in the creation of chatbots. Now before we explore the concept in more detail, it’s important to recognise the cautionary side of the conversation. Concerns abound in terms of the ‘devastating potential’ of artificial intelligence; Facebook recently had to shut down some automated bots because they had developed their own language, which was unintelligible to humans. That said, it’s clear we’re not yet at a ‘Terminator’ scenario of AI, where robots pose a sinister threat to humanity. So, for now, let’s look to the positive opportunities they bring.
Chatbots are already starting to infiltrate the market on a basic level, but in future, they may become our most common source of conversation. With a scalability that goes far beyond the usual headcount of a business, there are obvious commercial benefits to a digital workforce. So how are organisations going to be using these bots to change the way we interact?
Five bots we’ll get to know well
- Chatbots in customer services - Putting bots at the frontline of customer service management is perhaps the most obvious use case for the technology, and indeed it’s already in place in a number of organisations. By providing 24/7 service, bots will solve the frustration of being ‘on hold’ for minutes on end. But there’s an opportunity that extends beyond these functional outcomes; AI technology is advancing at such a rate that bots are also becoming ‘emotionally intelligent’. Companies will actually be able to adjust the emotional feedback provided by the bot, and in due course machine learning will enable the software robots to adapt their style.
- Chatbots helping international business - International business growth can be difficult if you can’t serve the full breadth of your market. While chatbots currently have a bias towards the languages of their creators, in time their linguistic capabilities will improve even further. Polyglot bots can already adapt and translate in an instant. With this sort of capability, they will help organisations to work seamlessly with customers, suppliers and organisers from across the globe.
- Chatbots that take stock- In the chaotic world of supply chain and logistics, bots will provide organisations with immediate visibility of inventory through automated alerts, while providing quality control for products leaving the warehouse. They will deliver immediate notifications when deliveries have arrived and unloaded, and when linked to ever-advancing IoT technologies, help avoid wasted journeys when orders cannot be fulfilled. Warehouse managers may simply reply on the bot in order to initiate a series of communications that update their customers, suppliers and management.
- Chatbots that make us live longer- The bots of the future will be provisioned with every piece of medical science ever documented. This will enable them to provide remarkably advanced diagnosis and prognosis services. Where access to healthcare is limited, chatbots will not just democratise well-being, but also allow people to seek medical help before it gets to crisis level. In turn this will increase capacity for higher value consultations with doctors, maximising strained hospital resources.
- Chatbots at our service - While we might not all have the benefit of a secretary at work, bots could soon give us smart automated assistance. Imagine all those diary invitations disappearing from your inbox, and your customers and colleagues delighted by your prompt and polite response. Perhaps more than any other area, bots will help us make friends in businesses when they become astute in managing our calendars and room resources.
Battle for the bots
Of course, making a chatbot strategy work will rely on more than just a superstar developer. In fact, the technology will be contingent upon a lot of elements in the enterprise infrastructure operating in synergy. Technology professionals are going to be responsible for keeping those bots chatting 24/7. Here are some of the most important priorities in the bot ecosystem:
- Bandwidth dependent bots: One of the touted benefits of chatbots is that they can work 24/7, dealing with customers and colleagues without getting tired or making mistakes. However, connected and intelligent software robotics will demand a lot of bandwidth as they gorge on data and connect with various systems within the organisation. This could mean that a lot of businesses need upgrade or refresh their infrastructure, and they’ll want analytics delivered across a robust and high speed network so that no robot analyst or automated production line arm suffers a moment’s downtime. In turn this will put extra pressure on the data centre; power and cooling will need to be highly effective and it’s likely that many organisations will need to scale up their capacity. So bringing in a robotic workforce won’t be a quick fix...
- Bots on edge: If you’ve got devices that need to act in a smart and adaptable way but are far from the main data centre hub, you’re putting limitations on the performance of those bots. Edge computing will allow fast ‘decision making’ and data processing on site, and this supports significant improvements in highly dispersed operations. Imagine a fleet of medical bots serving in remote areas, where getting an on-the-spot diagnosis could save lives. The ability to act almost autonomously opens up a new paradigm for medical treatment. The same will be true of enterprise use cases such as inventory management, where the smooth running of complex logistics could be compromised without edge locations.
- Bot brains: If an edge computing model allows a more effective transfer of data to and from a robot, the pressure is on us to make sure we can manage and use that data. And here we hit a snag; Gartner reports that many organisations don’t have the skills required to deliver an effective AI or ML strategy - simply because they can’t effectively manage the data required. Even amongst those businesses whose analytical capabilities are deemed ‘advanced’, 40% are struggling to create the right data science platform which in turn is required for chatbot machine learning. So the huge steps that have already been taken in data management will need to become a whole lot bigger before our robotic friends can truly change the workplace.
- Trust in the bots: This proliferation of sensors, edge computing hubs and network connection points must be kept secure. If bots are connected to a vast data mine of logistical data, proprietary corporate information or even confidential personal data, they become very important points to protect.As if the mobile device revolution hasn’t put enough pressure on security professionals, we’re now looking at a different league of critical challenges.
As an IT professional, chatbots may help make your day easier by dealing with challenging requests for support. At the same time as giving you that breathing room, they will make demands of the infrastructure that will increase pressure on numerous critical systems. But because of use cases from call centre augmentation to workforce productivity, bots are certain to flood our businesses. You need to be ready to chat about bots.