The first chatbot was built at MIT in 1966 and it certainly has a rich history, which you can read about in The history of the chatbot: Where it was and where it’s going. But, over the last ten years they’ve become increasingly visible; both to the customer and within organisations.
Chatbots fall under the umbrella of artificial intelligence (AI) and automation. But, in order to truly enhance the user experience (UX) and build intelligent workflows, Johan Toll — executive director of transformations at IPsoft — argues that enterprises need to tap into the power of a robust cognitive agent (or a digital colleague).
This digital colleague can deliver more business value than a low-level chatbot, he argues. Increasingly, these digital colleagues will take on repetitive, mundane tasks by combining their intuitive user interface with robotic process automation (RPA).
For example, instead of a HR manager having to make individual access restriction requests to the IT support team when off-boarding workers, the manager can make the request to the virtual assistant (using natural language) and the assistant completes the process. For one IPsoft client, automating the off-boarding process for roughly 200 workers each month in this way has eliminated the need to make more than 2,400 individual requests; and the mean time to resolution is down 81%.
HR, chatbots and robotic process automation: key for digital transformation
Organisations must modernise their HR function to succeed in digital transformation efforts. How? Success relies on software automation — chatbots and robotic process automation — and continued human involvement. Read here
Okay, introduction over..
The whispering agent
Chatbots, primarily, have been used to enhance the customer experience. However, their presence is increasingly being felt inside the enterprise — and the chatbot transition to colleague has begun.
Toll points to Amelia, who is used “by our clients like a whispering agent or colleague which they can co-deliver services with.”
Amelia houses the collective knowledge of teams so that they can consult with her at any given time. Similar to a chatbot, this digital colleague can then provide “very correct and strict answers back to questions, which then can convey back to the client,” says Toll. This, importantly, does not remove the human element of the process. “The human,” explains Toll, “still has to handle all the aspects of the client being satisfied or not, where you need to alter your responses.”
Digital colleagues can handle the routine tasks of a virtual assistant or chatbot, while sitting alongside human colleagues to observe best practices “and those individuals that have the talent to provide the response and capture that for further reviews”
It’s all about conversation
With chatbots you’re “merely chatting,” says Toll. Or, [I would argue] providing enhanced customer service — conversation is not super important, it’s all about solving a problem or answering a question as efficiently as possible. But, to become a real digital colleague the technology needs to have conversational capabilities — which is not a requisite of a successful chatbot.
Why chatbots fail?
The human is by definition a very unpredictable species — we change our minds, make mistakes and do not speak in scripted dialogue. It’s very important for the technology that forms the foundation of the digital colleague to be able to have a human-like conversation and be able to respond to the nuances of the imperfect human.
From an enterprise perspective, the machine that handles all the data from HR, Finance and IT needs to be able to cope with all those different areas and go back and forth between them and different humans in a different department.
“The chatbot doesn’t have those capabilities today but the digital colleague has, and that’s why we focus so much on the aspect of being able to have a natural conversation,” says Toll. “It’s not the digital colleague’s knowledge which is key, it’s the digital colleague’s abilities.”
“For example, I can read an HR policy but it’s not until I understood it and transform it into knowledge that I can actually make use of it and provide something which helps a client. And chatbots are usually a question-answering technology, but with a digital colleague you can actually do complex business processes and change between them: go back and forth and still figure out everything that came into the conversation.”
Laugh with me / Cry with me
Digital colleagues, similar to chatbots, need the ability to understand both the emotions, moods and personality of the end user — no small task, even by human standards!
In certain scenarios, employees — just the same as customers — will need the technology to understand what is being said and be able to provide an appropriate emotional response to that; in either its tone or text.
Just like humans, the digital colleague will need to filter every response it gives back by running it through its social engine; in order to figure out if it needs to prefix any response back with either formal, informal, apologetic or neutral responses.
“This ability is still in its infancy,” says Toll. “Currently most of the solutions within this field are based on chat and voice, not body language or tone.”
Chatbot implementation: Best practice tips from Avanade
Chatbots will evolve into digital colleagues in the enterprise, and digital customer service agents for the consumer. That is Toll’s firm belief.
If you have a really capable chatbot solution or virtual assistant that really understands the concept of self-awareness and has the ability to empathise, understand the client, be able to understand the context and really behave in a human-like manner, then “I would define that as a digital colleague,” says Toll.
This capability is based on language and the technology’s perception of it. Everything in society is based on language and language is not related to any kind of vertical or market or business. This is what separates the bot from a digital colleague.
“I cannot see why any business organisation or society should be exempt from automation. So I think this is something we’re going to see everywhere” — Toll
Train, don’t programme
Automation isn’t going to impact certain jobs or processes, it’s going to impact every aspect of the enterprise and smaller firms. This realisation has brought a change of attitude, practice and development.
When talking about digital colleagues, “we train them, we don’t program them,” says Toll. It’s all about training the cognitive capabilities of the technology — which means that it’s supposed to have a concept of observational learning.
Colleagues learn from each other. You get better by using the knowledge of your colleagues. You get better by sharing what you know with them. The same thing with digital colleagues. You need to train them and the good thing it has to be trained by humans
In this autonomous era, “instead of using a just-in-case model where you bring something to the client, we now bring a just-in-time model, or at least a combination of both, which captures the client’s utterances and goals,” explains Toll.
This means companies, such as IPsoft, can train the digital colleagues to the needs of the clients. And, it provides a great opportunity to have the customers define the capacity and variations of the service that the digital colleague provides.
What does this achieve? Well, a very personalised service that is not presumptive, and responds to the needs of the client. “That to me is a real digital colleague,” says Toll.