Top Five API Automation Patterns

Introduction

Don’t take APIs for granted. Yes, APIs will empower your automations by spanning across any number of systems your process workflows require. But they also provide the foundational glue for any number of RPA use cases.

Our recent white paper, “How API integration supports process automation”, discusses the procedural importance APIs provide for scalable, end-to-end automation. APIs are pervasive, “touching nearly every modern application, service, and user”.  

API patterns

In the modern robotic process automation (RPA) landscape, we’ve come to depend on a series of API patterns. These primary patterns provide the ‘glue’ that binds and integrates use cases for automated workflow design, user interface (UI) automation, and document understanding. The more complex the workflow, the more likely you’ll find multiple API patterns existing within a single workflow.

By examining the following top five API automation patterns, you’ll notice the degree to which APIs provide foundational support across RPA.

1. Application to application: App-to-app (or point-to-point) API integration provides a standard way of creating process-driven sequences across multiple third-party apps. Event-driven triggers support these app-to-app, or step-by-step sequences. Triggers generate next steps in processes such as order-to-cash, contact-to-lead, and applicant-to-hire. Each process may include multiple triggers that activate point-to-point sequences within a single workflow.

Note: App-to-app integration has an interesting and shared point of intersection between integration –platform –as –a service (iPaaS) and RPA.

2. UI robot to application: one of the hallmarks of RPA is UI emulation. Here, robots can be configured to support rule-based processes within each next step of an automated workflow. Robots can be trigger-activated as well, by changes within an application, such as a new employee being added to a human capital system. Robots can also support enhanced use cases when artificial intelligence (AI) and machine learning (ML) are enabled. 

3. Document-to-application: robots need to be empowered to read/decipher documents from any application, such as invoices or purchases orders, before taking the next programed step in a workflow design. Physical documents and PDFs can be processed and transformed into a JSON (Javascript Object Notation) format, enabling APIs to move these documents into finance and enterprise resource planning (ERP) systems. 

4. Human in the loop: interestingly, robots can assign process tasks to human team members for approval, escalations, and exceptions. Centralized monitoring gives visibility to API and UI-based robots working alongside people to gather data from applications. Here, for example, mundane copy-and-paste operations can be handed off to robots, increasing human productivity. 

5. Attended/unattended robots: unattended robots can support an entire workflow without human intervention. Attended robots work in tandem with humans, sharing tasks, but can be combined with attended robots, using a hybrid deployment model. APIs are incorporated into robots--including UI and document understanding tasks—all working together to automate the entire scope of a workflow. 

Conclusion

It’s accurate to say that software robots and API patterns based on best practices—are working together to alleviate worker tedium and reduce process inefficiencies.

In fact, APIs have become a must-have for the expansion of the modern, end-to-end automation platform. Again, this is further discussed in detail in the new UiPath white paper,  “How API Integration Supports Process Automation”.