AI has changed the process of software development. Most discussions seems to focus on engineering (e.g., code generation). As an innovation and UX veteran, I'm more interested in its impact on Product Conceptualization and Definition.
In this article, I want to discuss why User Flows remain critical for software development and how they can be modernized as part of your AI tool kit.
What is a User Flow
A user flow is a visual or written representation of the steps a user takes to complete a task within a product or service. It maps out the journey from entry to completion, illustrating how users interact with different touch points, screens, or system components.
Why User Flows in the Conventional process
Conventionally, user flows is a key part of a product requirement document. It clearly outlines to the team what a user does, at each step, when interacting with the product. Written in a user-centric way, it complements technical documentation, and is very useful in providing a common language for stakeholders to describe the product.
Similar to wireframing, user flows are designed to maximize certainty and align team understanding. A clear and agreed user experience upfront hugely reduces risks and costs associated with later engineering fixes.
Why User Flows in the AI-enabled Process
Things have changed. Incorporating AI in the workflow has reduced time and costs, resulting in a less risk-averse mindset - after all, we can build, launch, refine, and repeat at little costs now.
Well, at least in theory.
The point is: fixing oversights - no matter how minor - always come with a cost!
As a North Star
My take: the ability to iterate quickly doesn’t and shouldn’t eliminate the need for intentionality. Without a guiding structure, teams risk building reactively rather than strategically.
User Flows acts as a North Star. They prevent misaligned expectations, unnecessary rework, and lost momentum.
Imagine iterating to improve rather than fix, that’s good use of your resources.
For Prompt Engineering
Working with AI code editors such as Cursor, I have learned that step-by-step user interactions translate directly into step-by-step prompts. GenerativeAI responds exceptionally well to this type of instructions.
For QA testing
And just like in a conventional approach, User Flows remain useful for creating test scenarios, edge cases, and expected user behaviours for QA testing. This helps whether you are conducting automated or manual QA tests.
Modern User Flows
A living, simple guide
I think modern User Flows can be more lightweight and adaptive. While still defining details upfront, it is more like a dynamic, living guide that grows by incorporating iterative user feedback.
Visual or text, it really depends on your communication needs. The key is not to overdo it for the sake of documentation, but also don’t skip this important step.
Aided by GenerativeAI
Many AI tools can generate User Flows automatically.
For me, that’s a beautiful thing - saving me lots of time drawing out, from scratch, what’s on my mind.
Using the same prompt (description of a product concept), I have tried both ChatGPT and Miro’s built in AI assistant. The former produces a simple, accurate image that I can incorporate into a project brief. The latter creates a baseline User Flow on which I can build and expand - excellent!
Final Thoughts
In the AI era, it’s easy to be tempted to under-plan and over-ship.
In my opinion, it should be about moving fast with clarity - and that’s exactly what modern User Flows help achieve.
What do you think, do you use User Flows? And do you find that there’s a need for something different these days? Let’s discuss!