On 29 May 2020 at 16:30, a virtual KINTalk took place—via Zoom where around 50 people participated—this time, also from beyond the Netherlands!
ING’s YODA: an onboarding chatbot experiment
Mihaela Ionescu (innovation methodology coach) and Sander de Bruijn (head of global employee experience) from ING kicked off our KINTalk with their presentation on how to improve Employee Experience with chatbots. Employee experience matters a great deal because a differentiating employee experience can engage employees and result in high performing teams, which in turn can generate a differentiating customer experience and create value for the company (check out our post on how chatbots can improve employee onboarding).
To harness the potential benefits of chatbots and improve employee experiences, ING focused on improving the onboarding process, which refers to the period from when an employee accepts an offer to when (s)he feels embedded in the environment, about +100 days into the organization.
Mihaela explained that ING takes an approach blending design thinking, lean startup, and agile into a methodology that they developed themselves to accelerate innovation.
The steps are to (a) to identify and validate customer (employee) pains, (b) design experiments in order to draw (c) evidence-based conclusions and inform next steps. Following this approach, they first identified pain points in the onboarding process. They learned that pre-boarding (when the offer is accepted to before the employee comes physically to the office) was cumbersome, mainly due to the lack of communication. They concluded that clear communication, a personal touch, and a warm welcome were needed.
After defining the problem, they set up a chatbot experiment focusing on enhancing the emotional experience of their employees—making them feel welcome, seen, and able to interact with someone to get help. The simple chatbot, YODA, guides new hires throughout the onboarding process. The purpose of the experiment was to see if ING could establish connection in a digital environment rather than to test the technological possibilities of chatbots.
The results showed that even though YODA didn’t really add significant functional value (and that wasn’t the goal), new employees reported having a greater than 10% improved experience during the onboarding journey with YODA. Sander highlighted that the experiment showed that even a very basic chatbot like YODA can provide a personal touch to the onboarding process. Mihaela added that employees were quite excited, happy to share their experience with friends, and felt that they had received the attention necessary from the beginning.
Question: “What if at the end, people turn more to chatbots than to their colleagues?”
Mihaela: “How we see it is that for some questions, people will turn to chatbots and for some others, to humans. Chatbots have evolved a lot to show empathy, but still, they cannot mean empathy. It’s not one or the other, but both and for different purposes.”
Question: “What is the best strategy to encourage people engage more with chatbots? How can we create value from and nurture chatbots?”
Mihaela: “We learn to trust a system as we engage with it. What we’re trying to do is to find customer value. We try to deliver value, focusing on small steps instead of a technological overhaul.”
Insights from ChatLab
Next, Mahmood S. Zargar from KIN Center for Digital Innovation took the stage to share the findings of Master thesis lab, ChatLab, that he’s been running together with Ella Hafermalz (read more here).
As with ING, the purpose of ChatLab was not to prove anything on a technological level, but to explore how chatbots can bring practical value for new applications. Leading questions for the ChatLab were: Can we use chatbots for collecting qualitative data (for research or HR purposes)? If it works, how well does it work and how do people feel about the process?
The result? As Mahmood put it, “it kind of worked!”
Some of the key findings included expected ones such as that chatbots elicit less information than human interviewees and that chatbots often yield shorter interactions. One counterintuitive finding was that people showed higher tendencies to respond in a socially desirable way when interacting with chatbots than with human interviewees.
Regarding how people feel about the process, results showed that breakdowns were fine, but letdowns were not. People do not expect perfection from chatbots. If chatbots share their shortcomings upfront, breakdowns are perceived as ok. On the other hand, investing time and effort in solving a problem only to see it unresolved in the end, was not perceived to be cool. In short, the key factor was people’s expectations towards chatbots and how they are managed by chatbots and not so much in the performance of chatbots.
Although we could not enjoy the usual ‘borrel’ with beers and snacks afterwards, the virtual event allowed for broader participation from people across borders! We were delighted to see the event run so smoothly and the enthusiastic responses from participants.
Interested to know what would be discussed at our next KINTalk? Sign up to our newsletter and stay tuned on Eventbrite to hear about cutting-edge digital technologies and innovation stories at the intersection of science and practice! We hope to see you (maybe in person) at our next KINTalk!
Author: Bomi Kim