Nowadays, companies see chatbots as a promising solution, but one that demands a lot of resources. Specifically, time to get them running and tech talent to build them. Therefore, making them fast, yet right and ready for the future, is key for modern businesses.
That’s precisely why we launched the new Chatbot Configurator, a highly automated way to get a working chatbot in minutes. But before diving deeper into our tool, let’s see the challenges we’re trying to address here.
The Challenge of Building a Chatbot
Chatbots offer an additional communication channel with your users, using process automation and infused artificial intelligence (AI) to keep costs low. There’s undeniable hype around chatbots, but their value may not be easy to grasp at first glance.
Similar to what happens with many new technological innovations, chatbot solutions follow Gartner’s Hype Cycle:
At the peak of inflated expectations, one expects a frictionless experience for the end-user and that the chatbot magically self-improves and learns over time. These expectations lead both users and developers to get more and more frustrated when chatbots fail to meet their expectations.
With this in mind, I’ll share a few guidelines that we found to be critical when creating the best chatbot solution for your company, using OutSystems:
- Start small and deliver value right from the beginning. You can then evolve your chatbot based on real user feedback, instead of providing a highly complex solution that may not work for your user base.
- Map the chatbot scope and conversation flow beforehand. It’ll help reduce the pain of changing or adding something in the flow after implementation.
- Bear in mind the hype cycle and try to manage inflated expectations right from the start.
Getting a first working chatbot in front of your users is one of the main challenges in implementing these solutions. The best way to understand the value of chatbots is to allow users to try them, collecting feedback as soon as possible. Then, adjust and evolve accordingly.
Introducing Chatbot Configurator
The ideal scenario for any business would be to integrate a chatbot solution in apps and get it to end-users quickly. Not only timely but also without needing massive tech knowledge to get it running, right? Well, that’s exactly what Chatbot Configurator allows you to do.
Working together with the OutSystems.AI Chatbot (more about it here), this tool will automate most of what you need to get your chatbot. You start by connecting Chatbot Configurator to an AI provider in a guided path, and then you’ll be able to create fully working chatbots in minutes. In the end, your chatbot will be ready to integrate into your apps. And better yet, you can start from real use cases.
Quickly Building a QnA Chatbot
The main entry-level use case for automated communication channels is the QnA Chatbot, which replies to frequently asked questions (FAQs). It automates a process that usually consumes a considerable amount of time and is highly repetitive.
In Chatbot Configurator, you can select the QnA Chatbot option to use knowledge bases (KB) and easily incorporate them into the chatbot. All you need to do is give your chatbot a name, select your KB, and in a few minutes, it’s ready. Yep, it’s that fast!
If you’re just exploring or don’t have a KB yet, the tool also comes with a pre-built one. Just select it, try the chatbot, and then evolve it according to your use case. It’s all powered by the Azure QnA Maker, a service that incorporates AI using natural language processing (NLP) to handle the input questions and find the best corresponding answer.
Fully Customizable Solutions
QnA chatbots allow you to quickly scale your operations processes, such as for internal support. But what if you want the chatbot to answer with dynamically tailored messages for status checking and tracking?
To customize your chatbot responses with personalized answers (such as: “Hi, Daniel! Your request #1234 is currently in transit. [Click here] to follow your order.”), you’ll need to use the information and data you have on your OutSystems apps.
The Simple Chatbot option in Chatbot Configurator is the best pick for this scenario. It creates a generic chatbot resource connected to your OutSystems webhook application and is ready to integrate with other components. It’ll look something like this:
Further Expanding Your Chatbot With AI
After catching user messages, you must define the best way to deal with them. And that would be to avoid using a basic logic flow with rules and let AI deal with those decisions. You can do that with the Azure LUIS Connector. The main difference between the two options is that LUIS supports natural language. It doesn’t just match specific phrases, but actually allows for more natural and informal conversations.
You can then integrate any other service according to your own specific scenario. Consider other AI features to complement your solution, such as OutSystems.AI Language Analysis, that’ll allow your chatbot to identify trends and keywords, and detect users’ language and translate it. The following sequence diagram is a simple extension of the previous one:
To top it off, you can add other patterns to your chatbots, such as multi-step conversations that allow you to jump back and forth between multiple topics, and a human-in-the-loop system to blend human interactions in the conversation flow. With a centralized brain, you can also expose the same chatbot in multiple web and mobile channels.
Create Your First Chatbot Now
Chatbot Configurator is the best way to make the most out of our chatbot component, and also a quick and easy way to add an extra automated communication channel to your apps. You can either start from a simple chatbot and adapt it to your needs or create a highly efficient QnA chatbot with custom or sample knowledge bases.
More than just marketing hype, chatbots bring real value. And it’s now far easier to create and evolve one than it was before. Give Chatbot Configurator a try and have a fully operational chatbot up and running in just minutes!