Experiment
After extensive research and hands-on experimentation, I’ve yet to find an AI agent or mainstream SaaS platform capable of clearly and consistently visualizing complex, condition-heavy task flows while adhering to established flow-diagramming best practices. While many tools can sketch a flow, very few can reliably express logic. When conditional branching, exceptions, and edge cases stack up, today’s tools—and AI in particular—break down.
Other experiments (all types)
Experiment goals
Do flow diagramming best practices exist? If so, where are they, who defined them and which tool can I use to apply them to my work?
- Seek out flow diagramming best practices
- Challenge emerging technologies and existing SaaS platforms for their ability to execute against those best practices
- Evaluate the generated artifacts for their value-add potential
Historical context
Standing on the shoulders of giants
Long before UX diagrams, wireframes, or SaaS tools existed, humans were already grappling with how to represent logic. Philosophers and mathematicians from Aristotle to Leibniz to George Boole sought ways to formalize reasoning—reducing complex ideas into structures that could be examined, debated, and improved.
The foundations of modern interaction design are deeply rooted in these same logical constructs. At its core, a task flow diagram is an expression of AND, OR, and NOT. Every branch, exception, and edge case is a direct descendant of these principles. When flow diagrams fail, it is rarely because the problem is new. They fail because the logic has been abstracted, oversimplified, or visually misrepresented.
Understanding this lineage matters. Flow diagrams are not arbitrary collections of boxes and arrows; they are symbolic representations of reasoning itself. The more complex the system, the more essential it becomes to rely on a clear, consistent logic model rather than improvisation or visual flair.
- This AND That
- These NOT Those
- This AND That OR Those
- IF This OR That THEN NOT Those
Principles-first
Foundations
Before subjecting and evaluating popular SaaS platforms and emerging AI agents for their value-add potential, we should first agree on the principles and best practices to use for those evaluations.
The principles I use to determine the standards I choose for my flow artifacts are described below:
| Topic | Principle / Description |
|---|---|
| Artifact Types | The symbology should be capable of producing: Sitemaps: How information is organized Flows: How a user might move from position A to position B (often involving a conversion) |
| Mindset: Design vs. Alignment | Artifacts produced with the symbology should be considered an alignment artifact absent of erroneous visual stimuli that might distract from the relational or directional observations being made in the artifact. |
| Repeatable / Reliable | The symbology used in the artifact should have clear definitions and each artifact that leverages the symbology should do so the same way every time a new artifact is created. If the author of the artifact is creating new symbology each time a new artifact is created, the symbology fails and falls back to being merely arbitrary geometric shapes. |
| Interaction Design | When describing interaction design, the artifact must successfully communicate how the user flows through a defined task(s), and what the discrete steps are within these tasks (wayfinding). |
| Fidelity | As an alignment artifact, the level of detail needed should be as minimal as possible. Adding multiple and vibrant color or UI controls within the artifact can lead to unnecessary distractions while trying to align cross-functionally. The idea of the artifact is to arrest and visualize conditional logic (or relational hierarchies) and show user wayfinding and experience flow. Color is not required to accomplish this and only adds more time to build the artifact by introducing visual clutter. |
| Embraces Boolean Concepts | Conditional logic is rooted in binary choices that are chained together to create an end-to-end flow. Concepts of binary decision-making patterns will allow the following terms and cognitive patterns to be visualized: AND, OR and NOT |
| Swim Lanes | The ability to represent swim lanes (different environments, platforms, etc.) involved to complete a task must be possible to include in the artifact when it is helpful to do so. |
Best practice sources
👎
Source 1: International Standard ISO 5807
Information processing – Documentation symbols and conventions for data, program and system flowcharts, program network charts and system resources charts.
These standards, established by the International Organization for Standardization (ISO) and previously by the American National Standards Institute (ANSI), are detailed in ISO 5807:1985 and provide conventions for documenting data, programs, and systems. These standards inspire the symbology contained in FigJam, LucidCharts and other popular flow diagramming tools. In this designer’s opinion, these ISO standards fall short of what is needed in order to design out proper sitemaps and, more importantly, for arresting and visualizing complex task flows from a user experience point of view.
👍
Source 2: JJG Visual Vocabulary
In 2000, Jesse James Garrett introduced a set of flow-diagramming best practices that reimagined electrical schematic conventions for interaction and information design. By assigning logic-based definitions to simple geometric shapes, he transformed them into a symbolic language specific for expressing the user experience layer of relational hierarchies and transactional funnels. These symbols, when combined correctly and consistently, allow information architects and interaction designers to articulate complex, condition-heavy funnels (from the user’s point of view) with clarity and precision.
Findings
👎
SaaS: The pitfalls of today’s traditional SaaS options
After auditing multiple popular diagramming platforms, a consistent pattern emerged: most tools provide shapes, but very few provide meaning. Geometric symbols are offered without clear definitions, logical intent, or guidance for correct use. This encourages designers to treat flow diagrams as aesthetic artifacts rather than alignment artifacts, resulting in diagrams that look expressive but communicate logic poorly.
- SaaS platforms rarely define what each shape means in logical terms
- Visual freedom is prioritized over logical consistency
- ISO 5807 is treated as sufficient, despite being poorly suited for user-experience flows
👎
Artificial Intelligence: The pitfalls of AI agents and wizards
Current AI models struggle with structured reasoning and symbolic consistency. While they can generate basic flows, they fail when asked to apply a predefined visual vocabulary or maintain logical integrity across repeated attempts. Given the same scenario twice, AI tools routinely produce materially different diagrams which is an unacceptable outcome when precision and alignment matter.
- AI lacks consistency when modeling conditional logic
- Predefined symbology is not reliably respected
- Humans remain more dependable at producing logical structures
My preference
👍
I plan to keep using traditional, manual methods
Flow diagrams are alignment artifacts, not design artifacts. Their value lies not in visual expression, but in their ability to align teams around logic, intent, and edge cases. When diagrams prioritize aesthetics over meaning, they fail at their primary job.
Creating flow diagrams by hand forces deep engagement with the problem space. The real value isn’t the diagram itself—it’s the understanding gained by working through every condition, exception, and dependency. Manual diagramming builds empathy for users and clarity for teams in a way automated outputs currently cannot.
“Best practices don’t matter…”
- You’re going to need to define some sort of system to model the flow after, so why not established best practices?
- If you’re planning to produce more than 1 flow in your lifetime, redefining flow symbology and just “winging it” each time doesn’t make sense… the hard work of defining a symbology and visual vocabulary has already been done for you.
“Learning best practices takes too much time…”
- So does the churn involved with not clearly articulating important scenarios, way-finding and the logic involved that will drive the overall experience branching and interaction design.
- If you haven’t taken the time to learn and practice them a few times, you can’t really argue against them.
Designer resources
When stakes are high and complexity is real, guessing your way through logic is not a strategy. Clear flow diagrams require more than tools—they require a shared visual language and an understanding of why each symbol exists and how it should be used.
A consistent visual vocabulary allows teams to stop debating diagrams and start debating decisions. Once you internalize a system such as the one introduced by Jesse James Garrett, you stop improvising logic and begin communicating it with intent and precision.
Used together, these resources help reinforce best practices while keeping the focus on logic instead of decoration. The resources below are provided as starting points. They are most effective when used deliberately and consistently across projects and teams.
Related Content
Other Experiments
SaaS Review – FigJam
FigJam’s native flow diagramming library uses the International Standard ISO 5807
⭣
Generic shape labels
Missing connectors
Connectors
Context unaware
