Deep integrations - Our approach to integrations
6 Feb 2025
3.5 min read
For any AI agent to be useful, it needs to be able to access all the systems needed to do whatever job you’ve asked it to do. Think about it like a human agent: if they are going to get to the bottom of every ticket, they need to be able to pull up an order, check shipping statuses, issue returns, request refunds, and so on.
For an AI Agent to be able to resolve the most common kinds of questions that retailers receive, it has to be able to integrate with these systems to access all the relevant information.
That’s why we look to build deep integrations with every system our customers need. If we don’t have the integration already, we’ll add it – free of charge.
What do we mean by deep integrations?
Deep integrations is a term we use here at DigitalGenius to describe the way that we look to connect with the different systems a retailer uses. This means enabling our platform to connect to your business processes, allowing AI agents to take actions such as cancelling orders, changing addresses, or authorising refunds based on your business rules.
This is necessary because most chatbots can’t really do very much. They may be able to look up a tracking link, or find an answer in a knowledge base, but they’re unable to do anything that has real impact for the customer.
Think about a traditional, button-based chatbot. These chatbots are built to try and create shortcuts for customers to be able to find information. Now with generative AI, these buttons aren’t necessary any more, but the chatbot is still trying to guide customers to information.
But what if a customer already has all the information because they’ve looked it up, and now they need the bot to actually do something. Without deep integrations, most chatbots are stuck, and just pass you along to an agent, which actually ends up doing more harm to the overall customer experience. It’s frustrating for consumers who just want to get through to someone because the bot can’t help.
All types of integration
The truth is that not all integrations were created equal. Anyone who has ever used a system which is “integrated” with other platforms will know that some technologies work extremely well together, while others can just do some basic things.
Take integrations with carriers such as Fedex, DHL or DPD. A basic integration will be able to pull a tracking link from those carriers for a particular order. But what if the tracking status shows that a package has been stuck for 5 days? Or what if it’s been marked as delivered but the customer says they haven’t got it? That’s where a deeper integration is required.
This deeper integration means that an AI Agent can pull all the information and then apply business rules to it. This could be something like: Has the package been static for more than 5 days? Then consider it lost, and issue a new one.
What are the benefits of deep integrations?
Like in the example above, deeper integrations mean that you arm your AI Agent with more information to be able to take the right action.
But more than just arming an AI with information, deeper integrations allow for your AI Agent to actually take action. If business rules dictate that a replacement order should be shipped, then the agent has to go from the carrier to the order management system or ecommerce platform and issue a new order.
In old-school computer language it means being able to both read information and write it.
Without deep integrations, any AI-powered chatbot will be passing almost every question to an agent to pick up.
Whereas with deep integrations chatbots can actually resolve customer tickets within the conversation, meaning that a human agent never has to step in, and the customer gets what they want. And because it’s AI, it happens immediately, round the clock, and in a matter of minutes.
To see our ready-built integrations, take a look at our integrations page.
