Technology companies exhibiting at this year’s Consumer Electronics Show (CES) in Las Vegas are revealing just how far we’ve come with chatbots. Some of the examples on show are the most visible and tangible artificial intelligence (AI) applications in existence.
Chatbots are big business – what we call the apps of voice or text messaging platforms. Often deployed on the likes of Slack, Facebook Messenger and WhatsApp, the global chatbot market is expected to reach $1.25bn by 2025, with a compounded growth rate of 24.3%.
What’s the difference between a scripted and AI chatbot?
Chatbots come in two forms: sophisticated AI or natural language bots, also known as smart assistants, and scripted bots. Scripted bots are the ones that have been around for a long time – they are easier to build and mostly used for mobile engagement strategies.
However, AI and machine learning is becoming more common in these applications, giving chatbots the ability to learn like a human – but more effectively.
Smart assistant research has resulted in voice-activated devices, with Amazon’s Alexa and Google Assistant functionality shooting into the mainstream and our homes. Now we are talking to technology in a way we only saw possible in sci-fi programmes such as Star Trek.
Why have chatbots become so popular?
Made possible by the cloud, deep learning technologies, automatic speech recognition and natural language processing, smart assistants can provide engaging conversational experiences through voice and text that can be customised and used on mobile devices, web browsers and chat platforms.
One of the main reasons chatbots are increasing in popularity is down to app fatigue. People are fed up with having to download multiple apps and they’re only using a few applications on a regular basis.
Smart assistants solve this problem as they allow users to access multiple services via one messaging app to do to things such as chat with friends, get a takeaway, hail a taxi, buy food and check bank details.
Through AI and machine learning, a smart assistant can also learn about us and behave like a human would, which has powerful potential.
Why do chatbots matter to your business?
Like most technology that suddenly becomes trendy, the focus has very much been on chatbot use in the consumer world. But it’s easy to overlook the fact that smart assistants can have just as powerful an impact in the business world. Experts are finding that employees can be more productive and efficient if they can use smart assistants for both internal enterprise and business-to-business applications.
Analyst firm Gartner is positive about the possibilities that chatbots offer, stating that their broad appeal stems from the efficiency and ease of interaction for employees, customers and other users.
Most chatbots today are text-based but it’s enterprise smart assistants combined with voice activation that could significantly change and disrupt the way you work. Voice, for example, allows nuances in emotion and language that text just can’t replicate.
From a business point of view, this could humanise enterprises and add a layer of personalisation – we’ve seen this in the consumer world with the popularity of Amazon Echo and systems with Microsoft’s Cortana baked in. Workplace voice-activated smart assistants could allow you to multi-task – you wouldn’t need to physically interact with or see the devices you’re using.
What areas of your business could chatbots improve?
Here are a few examples of how chatbots could support your enterprise.
The automated helpdesk
Specialised chatbots could reduce the number of helpdesk workers your enterprise needs by handling routine requests via an automated chatbot response at all hours of the day.
Internally, this could help services such as IT support, while externally it could be used to respond to queries made by your customers as a replacement to point-and-click on a website.
Smart assistants could also work with your suppliers and vendors by performing actions such as answering simple queries, providing invoice information and sharing payment detail.
HR and recruitment
Your HR team’s processes could be automated with the aid of a chatbot by carrying out actions such as answering employee questions about pay and salary, or scheduling time off. You could also use smart assistants for recruitment, potentially using natural language processing to initially vet and examine candidates.
Software development work is important for many enterprises and chatbots could aid your team’s collaboration through a process called ChatOps. This is where a chatbot is modified to work with key plugins and scripts, and can perform routine tasks. Chatbots could be a great option to action tasks that are hard to trigger automatically and that need human consideration to begin their execution.
Advances in AI could allow smart assistants to converse with each other, allowing one to initiate conversations with other bots that have calendar access. Through this virtual conversation, these smart assistants could find a time that suits multiple calendars without an employee having to go through a manual process in finding a suitable time and date for a meeting.
How do you get started with chatbots?
When considering the use of chatbots and smart assistants, Gartner says you should assess the impact on resources and decide whether you want to use them to “replace” or “empower”. Replacement is easier, as you don’t need to consider integration with existing processes, while empowerment could enhance an existing process to make it better for users. Here are some tips you might want to use.
1. Evaluate chatbots for possible use in the enterprise
Understand what your business requirements are, how chatbots and smart assistants could apply to business tasks you have, and what they offer your employees when it comes to efficiency and ease of use. What are your current pain points in the business applications you’re using?
2. Deploy at least one chatbot internally and actively develop bots to assess use
There are already a fair number of enterprise chatbot-building platforms available for you to look at, many of which can be developed without any coding knowledge or maintenance from your technical teams.
Smart assistants run on widely available services such as Skype for Business and Slack, though integrating AI functionality will require more technical know-how.
3. Understand your budget
Chatbot and smart assistant technology and pricing can vary widely, depending on how complex a system you want to create. A text-based customer Q&A will be cheaper than an AI-powered calendar or booking system for example. You also need to examine whether you’re using an in-house development team, self-service platforms or managed service solutions.
4. Watch what your users are doing
By closely watching the behaviour of people using computers in your workplace, you’ll be able to understand what kind of business processes would work with the use of chatbots to create effectiveness and efficiency. You need to make sure the chatbots are sophisticated enough to deal with your business requirements.
5. Have the right security practices in place
If they’re doing their job properly, your chatbots will need to collect and store data, which demands you have the right security processes and practices in place. Poorly secured chatbots could put anything from financial information to intellectual property at risk.
What are the ethical considerations of working with AI?
AI is evolving faster than many of us have expected and the fact it can surpass decision-making in certain situations means it has major applications in the business world. High-profile examples include driverless cars, drones and manufacturing robots. In finance, AI could be a great aid in fraud detection, conversational banking and even the stock market with algorithmic trading.
However, this power is causing some alarm, with SpaceX founder Elon Musk claiming AI is “humanity’s greatest existential threat”. We shouldn’t be thinking so much about “military robots taking over the world” but more around how automation might increase the speed of issues we already have such as pollution and resource exhaustion.
In finance, it’s easy to imagine trading houses with super-intelligent technology making themselves richer, distorting the markets and “amplifying financial shocks”.
Until recently, it may have seemed strange to think about the ethical concerns of AI but the topic is more pressing – the Institute of Electrical and Electronics Engineers (IEEE) for example, has been working on establishing ethical standards around autonomous and intelligent technologies.
If you’re working with AI technology, following a set of standards will help:
- Minimise legal risk and the threat of a PR disaster. With GDPR guidelines for personal data use, you’ll come into scrutiny on how the technology you develop is applied in the market.
- Become more transparent around your technology. With AI being so powerful, you may be required to be more open about it. Doing so now will prevent future customer trust issues and avoid the need to update your systems in the future at high cost.
- Build trust with privacy-conscious customers. Increasingly, consumers want to understand systems that are using their data and targeting them with algorithms, meaning more demand for clarity and ethical standards.
How to work responsibly with AI
Kriti Sharma is vice president of AI at Sage. With an increasing number of enterprises working with AI, she says we need to take a step back and ensure the work we do is ethical and responsible.
Here are five values she believes members of the tech community must adopt to develop accountable and fit-for-purpose AI applications.
1. AI should reflect the diversity we try to build in our own organisations
We must make sure the algorithms we develop filters bias and negative sentiment that can result in AI-created stereotypes.
2. AI and its users must be held to account
Users tend to trust AI after a few meaningful interactions, which means just like humans, AI needs to be held accountable for bad behaviour.
3. Reward AI for “showing its workings” and good behaviour
Developers should think about reinforcement learning measures, based on what AI or robots do to achieve an outcome and how they align with human values to achieve that result.
4. AI should level the playing field
It should provide solutions such as voice technology and social robots that can be used by physically disadvantaged people.
5. AI should create as well as replace
This means we must take advantage of the new opportunities for humans offered by increased automation.
AI has huge potential to change our personal and business lives for the better. But transformative technology also has its risks if it doesn’t have the right regulations and safeguards. So let’s make sure we’re thinking about the future today, as it’s already here…
Are you thinking of implementing chatbots and smart assistants within your business or are you already using them? Let us know in the comments below.