Demystifying AI – Think big, start small

2018 March 9th

Many organizations and businesses overestimate the power of AI – perceiving it as that special kind of magic dust you spread all over your processes and everything will work much smarter than before. That is the one thing AI is not… However, given the right effort and focus AI can, and will, add enormous value as an integrated element of solutions, products and processes. So, how do you turn AI from religion into a genuine value-adding tool?

When working in the areas of business transformation, digitalization, disruption or the like, you need to take AI into account. This goes without saying. However, you also need to have a realistic picture about what AI is and exactly where it can add value.

AI has long been a much-hyped buzzword and it is easy to focus on the negative impact of AI in the future: The potential loss of jobs, the existential dangers of robots becoming too intelligent and the more philosophical discussions around ethics. But let’s turn it around and focus on killing the myths around AI and start benefitting from the many advantages AI can deliver in organizations!

The myths around AI

In my humble opinion, there are two main reasons why AI is still immature in most Danish organizations. They are, believe it or not, not technical – they are psychological and cultural boundaries to a technical solution.

First, we still do not fully grasp the concept of AI. We lack a common understanding of what AI is which is the reason behind the fundamental resistance towards AI in many organizations. The fear of losing jobs, control or identity is falsely perceived as a real and present threat when choosing to go ahead with AI. “The robots are coming – and they are taking our jobs” is not only used as a political statement, it forms and shapes business strategies every day.

A recent study made by Gartner says that AI will create more jobs than it will remove by 2020, with the public sector, healthcare and education as the big winners. Moreover, Thomas Davenport and Rajeev Ronanki published a survey in Harvard Business Review, interviewing 250 executives where more than half of them answered that their main objective with AI is to make existing products better – not to reduce head counts. So maybe we are just afraid of ghosts.

Secondly, we need to understand that many AI solutions are actually so-called “blended AI”, i.e. solutions dependent on both humans and artificial intelligence. We must find the right balance and learn to accept the fact that some routine jobs are actually better done by robots. This implies a huge change process. A change in culture and a change in the routines and tasks we perform today, by letting a more effective computer do the more tedious and repetitive jobs for us. If we were to be replaced by another human being, the resistance would be much less. The fact that we are being “outperformed” by robots is much more difficult to cope with.

Blended AI – How and why?

When talking about blended AI, there are two main types of solutions that come to mind: Process Automation and Cognitive Engagement. Let me give you give two examples of solutions that could make a real difference in your organization.

  1. Reading emails (process automation)

Today, organizations and businesses receive a huge amount of email communication from customers, partners and the like – emails that all need answering or that somehow tell us to perform a task. For instance, a membership organization receives daily email communication from members with questions related to their membership. In this case, you need subject matter experts to answer these emails on the basis of vast and broad experience. Or do you?

In many cases, AI can do this for us. Today, bots are capable of reading and understanding emails, even in Danish, and then make rules that transfer the content of the email into a specific action – all based on artificial intelligence. The specific action could then be a person calling the member and having and actual conversation. Or maybe the specific action could be an automated email reply with answers to the questions, without involving human beings at all. You can choose to automate only certain steps of a process, or all of them, depending on the level of complexity and need for human interaction.

  1. Chatbots (cognitive engagement)

Actually, bots are the perfect example of blended AI. A bot is ready to answer 24/7 and does not mind taking care of even the most tedious task. Moreover, when using chatbots it is possible to measure and evaluate the quality of the ongoing conversation. This means that if a member, customer or citizen talking to the bot is not receiving the help that he or she wants, it is possible to transfer the dialogue to a real person and that way offer a more deep and rich dialogue. It is a highly customer-centric solution – that does not need to be complex at all.

My point is this: You need to consider AI, but It makes sense to start small and think big – learning as you go. Start using artificial intelligence in a small area of your organization and test it, measure the effects and take that knowledge with you to build a proof of concept. The important thing is to get started – find a process where AI makes sense and choose the right solution for your organization.

Leave a comment

You must be logged in to post a comment.