Table of contents
Best practices for integrations
All ready to get started with your own integrations? Here are some final best practices to keep in mind when supercharging you tech stack with your integrated chatbot:
Keep it simple
While we strongly encourage you to integrate your chatbot with your tech stack, we do recommend to consider your integration carefully before you start working on it. Because an integrated chatbot can have a significant impact on your business, it’s important to consider whether an integration is really necessary.
You don’t want to pollute your workflows with unnecessary or cluttered datastreams. So when you think you need to integrate your chatbot with an external application, think about what result you wish to achieve with this integration. You can think about if there are ways to achieve this same result without an external integration. By keeping things simple you can focus your ideas on integrations that will actually make a vital difference in your business.
Which data, when
When integrating your chatbot with your tech stack, it’s important to carefully consider which data you want to send between systems. So before you start building your integration, take your time mapping out the datastreams.
Thinking about which data you want to retrieve and where you want to send it can be of great help when designing a new integration. But when you’re considering which data you want to use, it’s also important to think about when you want your integration to trigger.
Do you want to trigger your datastream right at the beginning of a conversation, or perhaps later on, when your chatbot has already answered some of the questions? Especially with more complex integrations, knowing when you want to send your data could easily help you plan the build. As you can work on your integration step by step and roll it out in several iterations, you can make the build more manageable.
Because there are so many use cases for an integrated chatbot, it’s easy to lose track of which possibilities suit your situation best. When that happens, you’ll often find yourself thinking in the same patterns over and over again. But don’t let that dampen your creativity!
Sometimes it’s best to take a step back and take a good look at the tools in your tech stack and your workflows. Be creative when thinking about ways to integrate your chatbot and don’t be afraid to experiment.
Especially with custom integrations, it can seem challenging to experiment, as custom integrations can take a lot of time to realise. But if you’re not yet sure how to approach your new, integrated chatbot, try testing out simpler versions of your integrations by using middleware. This way you can drastically lower the threshold and you can easily learn and improve your plan for your final integration, which can then be custom built, if still necessary.