Decisions behind our Elly chatbot
As you probably already know, September 5th was a big day for us as we launched the MVP version of Elly, our AI-based shopping assistant. We described most of her features in this article, but wanted to share our thoughts on why we launched Elly the way we did with the entire community.
Here are the answers to some of the questions that we asked ourselves frequently in the last couple of months:
Why an MVP?
MVP stands for Minimum Viable Product. When you are in the start-up phase like Eligma currently is, you have limited funds, resources and especially time. The issue is that you cannot be sure whether your product indeed solves a customer issue and if your way of dealing with that issue is a correct one. Launching an MVP is a way of testing some of these hypotheses to see whether a continuation of your product development makes sense. In our case, we wanted to see how our users would respond to AI-based recommendations. While not every one of those recommendations was made to all the users’ liking, the feedback was generally positive, which fills us with confidence about moving forward.
Why a chatbot?
Oracle’s survey states that 80 per cent of businesses are planning to use chatbots for customer interaction by 2020. While that is a telling and eye-opening number, it was not the main reason why we decided to present Elly as a chatbot. We chose a chatbot because we saw it as a perfect way to showcase Elly’s capabilities. While we could have opted for a standard filtering system, we felt it was not expressive enough. Not only did we want our users to find the most suitable products, but also to do so in a way that feels innovative, fast, convenient and just a little bit quirky. We looked at the chatbot solutions available on the market, but felt they were too limited for what we were trying to do. In the end, we decided to create a custom hybrid solution which we were confident would be up to the task.
Why specification-based recommendations?
While we realize that a lot of people are very passionate about the brands they support and buy, we also know that there are a lot of consumers who want to get the most for what they spend. The beauty of our system is that it allows for both ways of shopping. Elly finds products with the best quality/price ratio; however, users can still choose their favourite brands. Since we are not a store or supported by a particular brand, we can put the consumer’s needs first. This also means that we are not in competition with any stores, brands or manufacturers, but rather give them a new and innovative channel where they can offer their products.
Why mobile phones?
This was a very important decision because we were aware this category needed to be something that users would be familiar with but would still include enough specifications for Elly to use her algorithms to select the most suitable products. We diligently did our research and, in the end, it was the nearly 1.54 billion smartphones sold in 2017 (Statista, 2018) that persuaded us. We knew how often people use their smartphones and how passionate they are about them. There were two issues with this category that we could think of. One is that people usually buy their phones through their carriers and the other is that they are very passionate about the brand they want. We realized that our MVP’s main focus was to showcase the technologies we are developing and not to actually sell mobile phones, so the first issue was not really an issue anymore. We tackled the second issue by formulating Elly’s chat flow in a way that gives users control over which brand they prefer. While this means we do not recommend phones by specifications alone, we feel that it’s a good step in the direction of AI-based shopping choices.
Going forward
As Elly herself states, she is very young and still learning, but we are working hard to make her as smart and reliable as possible. In the future, we will add more and more product categories and stores to enable an awesome shopping experience. She will also learn from the people using her, so we encourage you to visit https://helloelly.com/ to talk to her and help shape the future of commerce.
Join our vision and our community!
