I heard this phrase in a McKinsey Podcast replaying a fascinating discussion on “Moving B2B companies into the digital world“. At some point in that discussion the participants talked about how data-driven suggestions meet human interaction and how this mix can make a difference. I thought, yes, that’s it and then some examples came into my mind. Indeed, it makes a difference to a sales manager when the system scores deals to identify the likelihood to close. He has then a much better basis to decide which opportunities shall get highest attention.
Taking decisions is mainly based on the fact how reliable and sufficient the information is that one has – and actually in most cases I never have all necessary information, be it in my business or private live. How often do I think “..if I would have known that, I would have decided differently..” OK, sometimes I deliberately do not want to look for all available information and take it into consideration, I just follow my gut feeling, as it happened when I bought my last car 😉
But in business I regularly put my gut feeling aside, I want to have all information that is possible to obtain. So that I am able to make a well based decision. And here the “machine” can help me, as it can take much more information into consideration than I can. And it can help me with experience, with much more experience than any single person can have.
Based on experience a marketer decides which customers should be addressed with a specific message, as he or she knows the offered products and the customers. Combining the experience of many and use this as basis for decisions, this when the ‘human-machine-mix” takes-off. Imagine our marketer segments the customers based on their buying propensity score. And this scoring does the “machine”: based on past experience it anticipates future behavior. It also can suggest what it is the best channel to address a customer to capture his full attention or what products should be suggested to which customer.
And here two other examples how past experience of many can be used to support the single human:
Think of an accountant focusing of accounts receivables. One of his tasks is to match incoming electronic bank payments against open customer invoices to clear them. Ideally the bank statements match automatically to open receivables, but sometimes this is not possible, as e.g. the customer has not given sufficient information with the payment. Now the accountant must do this matching manually based on his experience. Wouldn’t it be great to use the experience of all accountants for this task? And this is what SAP Cash Application does, the machine learning model is trained by analyzing historical clearing documents and can then match payments and invoices. The results will be returned for automatic clearing or review by the accountant.
The machine learning models of SAP Service Ticket Intelligence use historical data to assign an incoming service ticket to the right category and even suggest solution proposals for this ticket to the service rep. Before, a person had to manually examine and categorize the ticket based on his or her knowledge and experience. Now the system can do this and taking much more information into consideration than the single human. And the application improves accuracy as it continuously captures data – with every new ticket that is categorized.