Hackbot 2019: Highlights from This Year’s Chatbot Hackathon
Hi there, I am your Chatbot for learnings. How can I help you?
I am looking for a good way to learn what SAP does with Chatbots and implement my own use case.
Great, we have just introduced our Hackbot event. This is a great way to get to know the solution.
What is a Hackbot?
Wait and read… 😉
There are many ways to learn and get familiar with a software solution. There are books, webinars, official learnings, blogs, websites, hands-on tutorials or just basically somebody showing the tool. But in the recent years, a new concept has proven to be quite successful for getting familiar with a solution. We have done it already with several solutions, like SAP Predictive Analytics, SAP Analytics Cloud or SAP Cloud Platform. I am talking about the Hackathon.
While some people might associate this term with some geeky security hackers that sit in their garage trying to overcome the barriers of the government security system, this is not entirely what happens. Nevertheless the original thought of this is actually quite fitting because Wikipedia defines a hacker as “…any skilled computer expert that uses their technical knowledge to overcome a problem”.
And this is exactly what we do: we take a real life business scenario, any sort of scenario, and we use the solution to solve it over one to five days, depending on the complexity of the use case. There are presentations, hands-on exercises and expert trainers who will help you out with it. Either we provide the customers a real-life scenario or the customers can use one of their own. And, while we might not technically go to a garage, a creative working environment helps to get the brain cells working.
The SAP Conversational AI Hackathon—Hackbot
Just recently we have done a Hackathon for SAP Conversational AI, our chatbot solution. We called it Hackbot. We invited five customers from different industries such as insurance, retail, automotive and public sector in groups of two to three participants from different backgrounds, IT and business. This duality is key in order to succeed with the project. The customers were required to bring their own real-life use case that they could work on for the next three days. Beforehand we gave them some guidance on what they should think about when they pick a use case.
In order to make the best out of the experience, you have to think about the following questions beforehand:
- Does this use case have a Return on Investment? Does this chatbot reduce the workload of employees? Does this chatbot increase customer satisfaction?
- What is the business scenario?
- What is target group or end user of my chatbot? What are the goals, pains and tasks of that user?
- What would be typical expressions this persona uses in regard to their intent?
- What kind of backend systems would need to be connected in order to get the required information?
Those are just a few of the questions that you should be answering during an evaluation of a useful chatbot use case.
All participants put a lot of thought in their use case and were able to bring a great business scenario for the Hackbot. So, the fun could begin.
What You Can Learn from the Process
On the first day we had some introductions of the Hackbot and the participants, followed by some exciting sessions about the product and possible use cases, so the participants will get an idea of what’s possible. Afterwards we scheduled some time for defining the use case in detail. That means, really specifying the business scenario and persona and create the use case hierarchy. By defining a use case hierarchy, you define the overall use case and its subthemes.
For all the subthemes there are different intents that need to be defined. An intent basically represents an idea of what the chatbot is supposed to understand. This helps to bring some structure into the scenario. Part of that brainstorming was also to think about the conversational flow.
A conversational flow represents the actual conversation between the chatbot and the end user with all its expressions, questions and answers. And this was, when some participants really started to struggle. Understandably! Because up until the conversational flow you think, you really know your use case. But then you actually have to write down how the conversation between the chatbot and the user goes. This is when you realize the complexity. After the first expression of the chatbot user you really need to think about the “what happens if…”.
If I think back to my university times for example, I wished I had a chatbot that could have helped me through the requirements of an application. So, let’s say you have a chatbot that helps students to get information about the requirements of a bachelor’s or master’s degree. There are certain requirements attached to that.
- How will the conversation go on if the end user chooses master’s?
- How will the conversation go on if the end-user chooses bachelor’s?
- You may need to know what nationality is that person, where did he or she do the bachelor’s degree.
- How would the conversation continue if that person is German but studied in another country?
- How would the conversation continue if that person is not German but studies in Germany?
- And so on and so on.
- Those conversation flows can get complex, that’s why it is very important that you have done all the prework of really specifying your use case.
Each customer mastered the first day with really great progress. The second day the hacking time finally started. Now it was all about realizing the use case in the tool. We had a long day ahead of us. The morning started out quite relaxed. We played a fun game, had a little recap of the day before and a hands-on session, in order to get familiar with the tool itself. Then it got serious.
It was really a rollercoaster of emotions. And yes, I am still talking about software. You could hear whoops of joy from one corner of the room and grunts of despair from the other side. Luckily, we always had some great experts around who have experienced all this in other customer situations and who could help out our participants. And at the end of the day everyone had implemented parts of their use case. There were even some that didn’t want to leave. Like a real hacking party, we finished the day off with some nice beer and pizza.
The last day was presentation day. In the morning the customers had some time to finish the final touches of their use case and to build up a nice business presentation as a summary of their business scenario and how they have solved it. And then it was presentation time. We actually made quite a competition out of this. We invited some impartial guests who haven’t been involved with the use cases at all, so they could judge the value of the use case and how well this is presented. The other judges were the other customers and us, as trainers.
The Results and the Verdict
The chatbot use cases were amazing. It was absolutely impressive to see what the participants have managed to develop in such a short amount of time. This is actually the beauty of a tool like SAP Conversational AI. You can sign up for free in order to try out the technology and then you can start building on the platform without being a programming expert.
With the help of the trainers, they were able to create great use cases focused on external users for customer service, for example, and internal users to help employees daily with IT tickets, procurement, or orders. I have to be honest it was extremely hard to pick a winner. In the end, the tiniest details made the difference.
All in all, it was an extremely productive three days. We have received great feedback from the customers who said they would recommend this event to other participant, they’ve felt like they really learned a lot about chatbots in that small amount of time, and they’ve really loved the atmosphere.
We are all looking forward to our next exciting Hackbot.
By the way, as your personal chatbot I am usually not that chatty, but this format always gets me so excited.
Special thanks for an amazing organization and preparation from Isabelle Thurau, Business Development and Tomasz Janasz, Solution Management.