Information architecture is one of those terms that everyone loves saying but few can define. Not many people call themselves information architects, and those that do often come from a library science or taxonomy background.
But IA is becoming more of a buzzword every year, in part because of the value it can bring to software.
Good IA in SaaS means:
Providing enough context that the user always knows where they are, but not too much that it becomes overwhelming
Staying consistent with labels and vocabulary
Organizing content and navigations in an intuitive way
These principles share a large overlap with design heuristics. Users should be aware of the system status, understand where they are, and make decisions based on real-world experiences.
So: pretty simple. But how do we ensure that the architecture is actually good, and not just thought about? A big chunk of the answer lies in just asking the right questions. Sometimes this means usability studies, like holding card sorting exercises or testing how easily users can create a specific task, then analyzing the routes they chose to take. Other times, just asking yourself whether the right context is available - where the user is, where they're coming from, and where they're going - will direct the design in the right direction.
I won't go into more detail now because others have done so better than I ever could. If you're interested in learning more about IA, start with these three books. They're well-organized (who woulda thought?) and provide an overview of the basics while also providing concrete steps.
How to Make Sense of Any Mess by Abby Covert
Understanding Context by Andrew Hinton
Information Anxiety II by Richard Saul Wurman