When McDonald’s introduced self-service ordering kiosks to its restaurants, it elevated the discussion of human-machine collaboration in the workplace overnight.
While one of McDonald’s USA’s former chief executives worried automation would break linear career progression opportunities at the company — that is, eliminating junior jobs that provided a starting point to work one’s way up through the ranks, just as he had — his fears did not ultimately play out. Instead, revenue increased as customers spent more at a screen compared to having someone ask them to upsize.
Kiosks also freed up workers who could then be redeployed behind-the-scenes, particularly to the kitchen. In the words of one McDonald’s HR executive, “We have fewer people behind counters but jobs were not lost. They were redesigned from just taking orders to now more customer-facing and hospitality-driven roles.”
All of this happened so seamlessly that self-service is now just a part of the McDonald’s experience.
Clearly, there are lessons here for all organizations that are embracing human-machine collaboration in the digital workplace through the use of technologies such as artificial intelligence (AI), robotic process automation (RPA) and natural language processing (NLP).
The volume of use cases for these technologies, and the urgency in which they are being deployed, has significantly accelerated over the past two years. This is happening across a number of disciplines, including customer support operations and increasingly in finance departments.
AI, for example, is already in use at 58% of organizations, according to PwC; a further 35% have active proof-of-concepts in motion, and most see promise in them. In a separate survey by Deloitte, 78% of respondents said they have found use cases for RPA.
The McDonald’s approach offers clear guidance for getting human-machine collaboration right, and for making the implementation so seamless that it is barely noticeable.
Introducing 'Smart' Technology? Keep it Human
One key learning is that keeping people in an automated process is a necessary ingredient to success. The clue is in the title of human-machine collaboration: both people and machines are important. It doesn’t work if a person’s role is minimized or excluded. The "human" in this context includes both employees and customers.
AI, RPA or NLP that is deployed to augment people — rather than replace them — is going to be a more acceptable proposition for employees to accept. If machines are brought in with the express intention of replacing people, the organization may see increased short-term profitability but also a corresponding reduction in employee satisfaction. That may ultimately have a long-term negative impact on profitability, since it is hard to resurrect employee satisfaction scores after a major misstep.
Completely cutting employees out of a process is short-sighted. Machines don’t get good at executing processes out-of-the-box. They need to be trained on historical data, and they also benefit from the supervision of real humans with domain expertise.
A chatbot, for example, needs to draw on a knowledge base written by humans. People are also good at dealing with the 1% of cases that a chatbot can’t handle.
If a customer has exercised all the options in the bot and their problem remains unresolved, they need a path out. Speaking to a person should always be a fallback option.
In processes where the role of employees is diminished, expect to see corresponding flow-on impacts to the customer experience.
Customers are smart. They’ll quickly figure out if they’re talking to a machine. As businesses adopt human-machine collaboration, we’ve got to be careful how much of this we’re pushing out to customers because it’s got to be valuable to them and intuitive. That’s where McDonald’s got it right, but where other adopters continue to go wrong.
Customer interactions are about empathy and sentiment. Organizations that do not prioritize the customer experience quickly clock up low sentiment — net promoter — scores.
Expect More Lightbulb Moments
As organizations become more aware of the full spectrum of human-machine collaboration use cases, adoption will increase.
Many of the teams and divisions now implementing human-machine collaboration tools are in the business, such as finance or HR. In our experience, they are often unaware of the full extent of what’s now possible.
For example, finance teams can use machines for invoice scanning and exception handling. Automating this was previously considered too hard for many finance teams. When they find out they can use technology out-of-the-box to open every email and attachment and extract the key details of the invoice, it leads to a lightbulb moment — and deeper investigation of the art of the possible.
The challenge now is to build awareness for what kinds of workloads can be made more efficient and cost-effective with human-machine collaboration.