Intelligent User Flow Design

🌀 User journey maps often capture “perfect” journeys users never take. We need to stop designing paths, and start designing loops, especially in AI products ↓ We use journey maps to capture, understand and refine user's experience. However, these maps are merely an idealistic view of what users SHOULD be doing, rather than what they actually ARE doing. Linear paths don't consider detours, circling back and forth, abandonments and returns and shortcuts. In fact, our interactions with reality rarely follow a well-defined, structured script; they’re a series of adjustments and feedback loops — depending on environment, disturbances, decision-making and actions. Workflows shouldn’t be perceived as a rigid cage, but as an orchestrated loop. Matt Fick and Max Peterschmidt suggest to rethink the idea of designing paths and design loops instead, especially with AI products in play. We start with a goal, make decisions, sense what’s going on, study environment, take action and then keep checking again, and again, and again. It follows a simple structure: 🎯 1. Setting a goal First, we establish a goal: what is the user trying to achieve? Desired outcome is the foundation on which the product will ground all its actions and adjustments. We must help people articulate their goal — with slow prompting and better calibration (knobs, pre-prompts, buttons, sliders). 🌡️ 2. Studying the current state (Sensors, Environment) To improve something, we must understand its current state. We find the right sources and collect the right inputs to get a snapshot of the current state. Often there are many meaningful inputs, and often they are very difficult to predict ahead of time. 🧠 3. Making decisions (Controller) Next, we evaluate the data and compare it against the goal. We come up with meaningful actions and get recommendations, grounded in trusted sources. Mapping the reasons for recommendations is critical for building trust and confidence — with AI, but not necessarily with LLMs. 🚀 4. Taking actions (Actuator) Once we decide that an adjustment is necessary, we take an action, or we ask agents to take an action — directly manipulating the environment closer to the desired outcome. The actions are typically initiated or approved by humans, and that’s what we mean with “human in the loop”. 🧲 5. Studying and refining the new state We gather data about the changed environment, and then use these inputs to suggest the next batch of changes as output. With nested loops, when many people or AI agents are involved, output in one loop becomes an input in another and informs next decisions and actions there. An interesting and realistic model in AI world, matching the complexities of the real world better than journey maps often do. Indeed, workflows aren’t rigid cages — they are non-linear, cyclic and must be highly adaptive to be meaningful. They must sense, respond and learn — and loops do just that.

Intelligent interfaces make real-time design choices. For designers, sharing design decisions with robots can be… uncomfortable. But delegation ≠ abdication. The new work for designers is to give context and guidance to help the system make good choices. I made a guide, demo, and video for designers (link in the comments) about how to do this and keep the results on the rails. Done right, the result is a radically adaptive experience that responds to user context and intent. Layouts that rearrange themselves. Forms that choose smart defaults. Chat that “speaks” with well chosen GUI elements instead of text. It’s easier and more reliable than you might expect. The guide includes a simple, directional pattern library for giving the LLM its marching orders. For designers, sketching in simple plain-language system prompts becomes part of the design process, at least as important as drawing interfaces in your design tool. Instead of designing every interaction, you’re designing the *physics* of your application’s tiny universe. You define the behavior and constraints for making design decisions in the interface. It’s design system work for real-time decision-making. The basic recipe for wiring interface to intent: 1. Provide a constrained set of UI outputs. 2. Map those outputs to intent (“use this pattern to address that intent”). 3. Ask the LLM to understand intent and choose the right UI or action. It used to be really, really hard for systems to determine user intent from natural language or other cues. Now LLMs just get it. They grasp underlying semantics, they get slang, they can infer from context. LLMs may hallucinate facts, but they’re brilliant at interpreting intent and the shape of the expected response. This makes them a powerful and reliable partner for interpreting user meaning and delivering an appropriate interface. Check out the demo and give it a try yourself. Start writing; the interface is listening. Link in the comments (because you know, LinkedIn).

AI operators think in workflows, not words. Stop tweaking prompts. Start designing systems. Months of optimization led me to the breakthrough strategy of focusing on workflow design. Here's how it reshaped the game: Marginal gains. Infinite loops. Shaky outcomes. Then it hit me: the flaw wasn’t in my prompts... it was in the architecture. The Prompt Engineering Dilemma: Solitary prompts are delicate. They shine in trials but crumble quietly in real-world use. Context changes. Edge cases appear. The model confidently produces garbage. You can't solve architectural problems through prompt engineering. The Prompt Architecture Shift Stop asking: "How do I improve this prompt?" Start by asking: "How do I design a system for graceful prompt failures?" This involves: Decomposition Break a complex task into simpler ones. Each prompt handles one function, and together they accomplish what a single prompt can't. Validation Layers The output of every prompt gets checked before the next step: - Does this look right? - Does it match the expected format? - Does it contain the required elements? Failures get caught. Reruns happen automatically. Bad outputs never propagate downstream. Context Management What information does this prompt actually need? - Not everything. - Not the whole document. - Just the relevant context for this specific task. Smaller context windows. More focused instructions. Better outputs. Fallback Paths: What's the plan for model errors? Human help. New prompts. Quick recovery. Our design handles setbacks. Failures are expected. Prompt engineering refines; prompt architecture shapes the vision. The other makes entire workflows reliably excellent. The operators who win at AI aren't better at writing prompts; they're better at designing systems where prompts are just one component. What's one workflow you've built that chains multiple AI calls together?

You've heard me say that UX should be invisible, that the user should use the design seamlessly, without drawing attention to itself. It should enable users to interact with the system naturally, without unnecessary interruptions or confusion. Here's how UX could be invisible: - Align with User Mental Models: The design should match how users think and expect things to work. This means understanding users deeply—how they approach tasks, their mental shortcuts, and their expectations. When the design aligns with these mental models, users don’t have to pause and learn; they just act, and the interface works as anticipated. - Streamline Tasks and Remove Clutter: An invisible UX simplifies tasks by removing unnecessary steps and presenting only what is essential at each stage. Every element on the interface has a purpose directly tied to the user's goal. By stripping away anything extraneous, users can complete their tasks without distraction. - Guide Users Subtly, Not Forcefully: Instead of overt instructions or heavy-handed guidance, the interface should provide subtle cues that guide users gently. This could be through visual hierarchy, natural language, or affordances that hint at what actions are possible. Users should feel in control and empowered rather than managed or restricted by the design. - Error Prevention and Recovery: The design should anticipate potential user errors and prevent them before they occur. If errors do happen, the system should offer simple, immediate ways to correct them without penalty or frustration. - Consistency in Interaction Patterns: Consistent design patterns help users build a reliable mental map of how to interact with the system. Use familiar conventions so users feel comfortable and confident. Consistency reduces the learning curve and makes the interaction feel second nature, contributing to the sense of an invisible UX. - Proactive Support Without Interference: Interfaces could offer proactive help—like suggestions, auto-completions, or predictive inputs—exactly when needed, but without overwhelming the user. The support should feel like an enhancement rather than an interruption. - Design for Flow: Design for flow, where users are fully engaged and can move through tasks without disruption. Remove points of friction and create smooth transitions between different parts of the task, allowing users to maintain their momentum and focus. - Functional Simplicity: Invisible UX focuses on the core functions that directly contribute to user goals, avoiding unnecessary features or complexities that might confuse or slow down the user. Good UX is not about showcasing every possible feature but about prioritizing what’s truly necessary for the user’s success. In summary, create an experience that is so aligned with the user's needs, expectations, and behaviors that it becomes an almost subconscious interaction. The user should achieve what they set out to do with minimal thought about the interface.

I went to an AI UX workshop last night expecting recycled LinkedIn advice about "building AI trust through transparency." Instead, Isabella Yamin tore down LinkedIn's job posting flow using her CarbonCopies AI framework in real-time, while founders shared raw implementation struggles. It completely changed how I'm rethinking Maibel's onboarding flow. Here's what I stole from B2B SaaS principles to redesign emotional AI for B2C: 1️⃣ Progressive disclosure with purpose LinkedIn's fatal flaw? Optimizing for completion ease > Outcome quality. Recruiters are drowning in irrelevant applications because AI never learns what "qualified" means. The personalization paradox: How do we give users enough control without overwhelming them? Users don't want "frictionless". They want INFORMED control. 📌 At Maibel: I was falling into the same trap, making emotional coaching setup so simple that the AI couldn't understand user context. Now? Progressive complexity with clear trade-offs. Show users how their choices impact outcomes. → Want deeper insights? Add more context. → Want faster setup? Here's what the AI can't personalize. 2️⃣ Closed-loop data intelligence: What Platfio gets right They've built a platform for software agencies where where every data point feeds back into the entire system. User preferences in marketing flows shape proposals. Campaign performance shapes future recommendations. Every interaction becomes intelligence for future recommendations. 📌 At Maibel: Most wellness apps store emotional check-ins like digital journals. I'm turning them into predictive feedback loops. Emotional intelligence isn’t static but COMPOUNDS. Today's reflections shift tomorrow's suggestions. Patterns fuel prevention. Users' inputs on Monday could predict AND prevent Friday's breakdown. 3️⃣  Multi-modal creativity: Wubble's transparency approach Translating images and files into music - who'd have thought? They've cracked multi-modal creativity where users become co-creators, not passive consumers. The breakthrough moment for me: What if users could see how their visual environment contributes to emotional context? 📌 At Maibel: Users upload images of their day and see how AI analyzes emotional cues: cluttered workspace = overwhelm, junk food = stress eating. Multi-modal understanding users can contribute to and influence. 💡 The bottom line? B2B Saas gets one thing right: Every interaction has to earn trust. In B2B, failed AI means churn. In emotional AI, failed trust breaks belief in tech entirely. 📌 Here's what we're doing differently at Maibel: → Progressive complexity → Context-aware feedback → Multi-modal participation → Intelligence that compounds with every input. It's not just about building WITH AI. I'm designing systems that learn understand YOU before you even need to explain yourself. Kudos to Isabella, Shivang Gupta The Generative Beings, Shaad Sufi Hayden Cassar and everyone who shared deep product insights.

We just rewrote our Business Case Builder from the ground up. Here's why. The old flow tried to do too much at once - context, use cases, and inputs all mixed together. It worked, but it created cognitive load. Users had to hold too much in their heads. The new flow separates three distinct mental models: 1. Context gathering Instead of asking users to structure their knowledge upfront, we let them dump everything - transcripts, notes, files. The system does the structuring. Humans are better at providing; AI is better at parsing. 2. Intelligent use case matching This is the sophisticated part wrapped in a simple interface. Under the hood, we're analyzing all available use cases against the provided context - looking at industry signals, pain points, stakeholder patterns, competitive mentions, deal characteristics. We surface the ones most likely to drive value. The shopping cart UI? That's just the interface. The intelligence is in the recommendation engine. But we never force the AI's judgment—users can browse the full library, compare alternatives, and make the final call. 3. Input review This is where the magic happens: we surface every assumption in one flat list. Tab navigation makes it fast. Transparency makes it trustworthy. The result: users complete business cases 3x faster, and they understand what they're creating. Good product design isn't about adding features. It's about removing friction between intention and execution. Huge shoutout to Asa and Miron for shipping this release! 🔥

I’ve been designing + building products for 20 years. One AI project changed everything I thought I knew. It was 5 years ago. The brief: an AI assistant for financial advisors. "Easy" I thought. I brought the playbook - understand users, map needs, prototype, iterate. Within weeks, every method had failed. User-centred design has given us incredible tools: journeys, personas, usability testing. It created a shared language for innovation and put users at the centre of product development. But it also gave us something dangerous: the illusion that good process guarantees good outcomes. Where design methods break: 🔴 They treat all problems as design problems. Not every challenge needs a workshop. Some need engineering breakthroughs. Some need business model innovation. Some need regulatory change. When your only tool is empathy, everything looks like a user experience problem. 🔴 They assume user needs reveal future possibilities. Advisors thought they wanted better dashboards. Not "AI that predicts my clients needs and anxiety levels". Revolutionary products create needs people didn't know they had. 🔴 Confuses good process with good results. Following the method perfectly doesn't guarantee you're solving the right problem. Great design comes from insight, not adherence to frameworks. What building AI systems has taught me: 🤔 The old tools need rethinking. User research couldn't predict interactions with something that evolves. Journey maps couldn't map AI that creates new paths. Prototypes couldn't capture systems that learn and change. 🤔 The real design challenge isn't the interface - it's the intelligence architecture. Should the system interrupt or wait? Learn from the user or protect their privacy? Optimise for efficiency or explainability? These aren't UX decisions. They're ethical and technical decisions that determine trust, dependency, and agency. 🤔 And critically: AI systems create feedback loops that change user behaviour over time. Traditional design assumes static user needs. AI design requires predicting how your solution will reshape the problem space. We're designing systems that could shape human behaviour for generations. User research and workshops aren't enough anymore. We need a new playbook. What I've learnt: 🟢 Ask "should we?" before "how might we". Consider consequences, not just possibilities. What data does this use? How does it learn? What could break? 🟢 Develop systems thinking. Your decisions ripple through complex networks of technology, behaviour, and culture. 🟢 Design for responsibility, not just iteration. Every design choice becomes a values statement when scaled through AI. 🟢 Question the AI narrative. Not every problem needs an AI solution. Some need better human processes. 🟢 Partner deeply with engineers and data scientists. The best AI experiences emerge from true collaboration, not handoffs. The craft evolves. The responsibility remains the same. Let’s write new rules. Who’s in?

"AI will make traditional interfaces invisible!" I keep hearing this, but my deep dive into over 100 AI workflows in SaaS products and reading about Microsoft's commitment to human agency in AI patterns have convinced me otherwise. Here are the UI patterns I'm seeing: Traditional interfaces serve a crucial role in the consumption of AI products. Advanced prompt engineering can be packaged into user-friendly point-and-click UI, short-hand sentences, or step-by-step wizards for less tech-savvy users. I refer to these as Prompt Triggers. They come in 3 flavors: 1. User Prompting 2. Prompt Templates 3. Prompt Builders Let's unpack these patterns with real-life examples. 1. User Prompting - The free-form approach that allows users to type and send any text-based response to the system. Best when users need fewer limitations. Example: Adobe Photoshop uses a floating action bar for the user’s text prompt. 2. Prompt Templates - Pre-constructed text prompts triggered by a button or action in the UI. Best for enabling specific tasks, like text summarization. Example: Scribe uses an inline editing UI with a list of pre-constructed text prompts. 3. Prompt Builders - Guided wizards that help users build detailed prompts without writing them. Example: Gamma uses a guided wizard for building a detailed prompt for presentation generation. For more weekly UI design insights and inspiration for AI patterns, check out my library at Teardowns.ai.

During meetings with stakeholders, we often hear about 𝒓𝒆𝒅𝒖𝒄𝒊𝒏𝒈 𝒃𝒐𝒖𝒏𝒄𝒆 𝒓𝒂𝒕𝒆𝒔, 𝒊𝒏𝒄𝒓𝒆𝒂𝒔𝒊𝒏𝒈 𝒓𝒆𝒕𝒆𝒏𝒕𝒊𝒐𝒏, 𝒂𝒏𝒅 𝒐𝒑𝒕𝒊𝒎𝒊𝒛𝒊𝒏𝒈 𝒄𝒐𝒏𝒗𝒆𝒓𝒔𝒊𝒐𝒏 𝒇𝒖𝒏𝒏𝒆𝒍𝒔. If you're feeling confused and overwhelmed about how to do all of this, you're not alone. Here's something for those new to the world of metric-driven design. Trust me, your designs can make a real difference :) 𝗙𝗶𝗿𝘀𝘁 𝘁𝗵𝗶𝗻𝗴𝘀 𝗳𝗶𝗿𝘀𝘁, 𝗴𝗲𝘁 𝘁𝗼 𝗸𝗻𝗼𝘄 𝘆𝗼𝘂𝗿 𝘂𝘀𝗲𝗿𝘀 𝗔𝗡𝗗 𝘁𝗵𝗲 𝗯𝘂𝘀𝗶𝗻𝗲𝘀𝘀 → Talk to real users. Understand their pain points. But also, grab coffee with the marketing team. Learn what those metrics mean. You'd be surprised how often a simple chat can clarify things. 𝗠𝗮𝗽 𝗼𝘂𝘁 𝘁𝗵𝗲 𝘂𝘀𝗲𝗿 𝗳𝗹𝗼𝘄 → Sketch it out, literally. Where are users dropping off? Where are they getting stuck? This visual approach can reveal problems you might miss otherwise and which screens you need to tackle. 𝗞𝗲𝗲𝗽 𝗶𝘁 𝘀𝗶𝗺𝗽𝗹𝗲, 𝘀𝘁𝘂𝗽𝗶𝗱 (𝗞𝗜𝗦𝗦)→ We've all heard this before, but it's true. A clean, intuitive interface can work wonders for conversion rates. If a user can't figure out what to do in 5 seconds, you might need to simplify. 𝗕𝘂𝗶𝗹𝗱 𝘁𝗿𝘂𝘀𝘁 𝘁𝗵𝗿𝗼𝘂𝗴𝗵 𝗱𝗲𝘀𝗶𝗴𝗻 → Trust isn't built by security badges alone. It's about creating an overall feeling of reliability. Clear communication, consistent branding, and transparency go a long way. 𝗠𝗮𝗸𝗲 𝗶𝘁 𝗲𝗻𝗴𝗮𝗴𝗶𝗻𝗴 → Transform mundane tasks into engaging experiences. Progress bars, thoughtful micro-animations, or even well-placed humor can keep users moving forward instead of bouncing off. Remember, engaged users are more likely to convert and return, directly impacting your key metrics. 𝗧𝗲𝘀𝘁, 𝗹𝗲𝗮𝗿𝗻, 𝗿𝗲𝗽𝗲𝗮𝘁 → Set up usability tests to validate your design decisions. Start small - even minor changes in copy or button placement can yield significant results. The key is to keep iterating based on real data, not assumptions. This approach improves your metrics and also sharpens your design intuition over time. 𝗗𝗼𝗻'𝘁 𝗿𝗲𝗶𝗻𝘃𝗲𝗻𝘁 𝘁𝗵𝗲 𝘄𝗵𝗲𝗲𝗹 → While it's tempting to create something totally new, users often prefer familiar patterns. Research industry standards and find data around successful interaction models, then adapt them to address your specific challenges. This approach combines fresh ideas with proven conventions, enhancing user comfort and adoption. Metric-driven design isn't about sacrificing creativity for numbers. It's about using data to inform and elevate your design decisions. By bridging the gap between user needs and business goals.

Design decisions are bidirectional. I see this pattern a lot in our customer work at ZURB where a team does solid design work, ties it to a business goal, and brings the concept into a review expecting momentum. Meetings get some head nods, a few questions come up… and then the meeting ends. Nothing actually moves. After a few rounds of iteration, it starts to feel like everyone is fighting the system. What’s usually underneath that tension is a misunderstanding about how design impact works. Many designers assume impact starts by creating something from an expressed goal that sounds business-relevant, then handing off the design to the business to understand, interpret, and deploy once the idea is done. The expectation is that the value will become obvious after the fact. Let’s call this the push method. Example: Design improves onboarding based on user feedback and presents a new flow. The work is framed as “better UX.” No metric is tied to the decision. Leaders debate opinions, scope, and risk. The design gets watered down. In practice, this is a hard way to build momentum with stakeholders. Most business problems are fuzzy. Goals like attracting customers or gaining market share are not concrete. They are just as creative and undefined as designing an interface. This is where a pull method tends to work better. Example: The team starts with a business moment like onboarding → activation → conversion. They ask where users get stuck in that flow, then use UX metrics like completion, comprehension, or time on task to make that moment visible. Design decisions are pulled from the flow, not pushed from isolated user feedback. When business goals and capabilities are the starting point, they become real constraints. Users are introduced as friction in the system. The work shifts from pushing solutions to shaping the business problem around user needs. That framing changes the design problem. More importantly, it creates a shared vocabulary with the business. You are no longer translating after the fact. You are defining the problem together and learning what actually works for customers. Starting from business goals does not limit design or weaken the user experience. It does the opposite. It reveals clearer paths to wins that would otherwise get chipped away as the creative process unfolds. If this tension feels familiar, join hundreds of product and design leaders in the Glare forum working through it together. https://lnkd.in/ggHXcVQZ