Elie Khoury

You can now try Gemini 3 Pro in Figma Make! What stands out to me from early testing is how useful Gemini is at exploring visual directions + ideas that designers can then shape with their craft. As AI models get better, Figma gets better and designers can go further. I share more thoughts in the attached video, but a few quick highlights below... Gemini 3 is excellent at 0->1 generation I've never seen a model do one shots like this so consistently! It's honestly very addictive... using it in Figma Make reminds me of play Civ 5 😅 Just one more prompt... ! A wide variety of styles and aesthetics With detailed prompting, you can consider different visual directions very quickly. This creates more understanding of the option space you're working in, giving you more time to deeply explore the path you want to take. It is a highly steerable model Of course we also tested Gemini 3's ability to work with Figma design files in Make. The model translates designs with precision + offers impressive interactions. Working with Design Systems I also wanted to see how Gemini 3 could work with Design Systems in Figma, in particular with Make kits and npm packages. Again, the results were impressive. It's great for exploring options and refining details while staying on brand. New possibilities I'm especially excited to see the creative possibilities skilled, high-craft designers will explore with models like Gemini 3 + Figma Make. Additionally, these types of tools help non-designers express ideas at higher fidelity — bringing new perspectives to the design process. If you want to try Gemini 3 in Figma Make, go to "Settings" in the upper right and click "Experimental models." Then toggle it on. Have fun and please let me know if you create something cool!

2,828

83 Comments

The future of design isn't UX. It's AX. John Maeda's 2025 report made it clear. AX = Agent Experience. Design is shifting from humans to AI agents. Not interfaces for people. Interfaces for bots. Here’s why it matters: AI agents don't click. They read. AI agents don't navigate. They interpret. AI agents don't guess. They need clear paths. Most UX today is built for human confusion. AX demands machine precision. If you want to stay ahead, start designing for: APIs over GUIs Instructions over interfaces Clarity over creativity The best products tomorrow won't just be user-friendly. They'll be agent-optimized. Clear inteface for the machine, invisible interface for the humans.

1,068

110 Comments

User interface modes reduce #usability. Here's the story of #UX pioneer Larry Tesler's battle against #modes, first at Xerox PARC and then at Apple. We enjoy the fruits of his labor to this day. My full article on UI modes 👉 https://lnkd.in/g7U7W98V

47

6 Comments

Cool stuff we've seen folks do with Mastra's new dynamic agent functionality 1. Customize an agent's system prompt by running user metadata through an LLM call 2. Build several dozen templated agents for a multi-location hospitality chain, using location details to customize the system prompt and tool selection 3. Use a smaller model if a user is on a free plan and a larger one if they are on a paid plan 4. Customize the agent's toolset based on the user's role Blog post in comments!

27

6 Comments

Document review is a task that is going to get 99%+ automated by AI agents - they can do it faster and more accurately than humans. This is a great case study by CondoScan showing both the pains of buyers reviewing a ton of docs during the homebuying process, and also how they were able to build an automated document workflow architecture with LlamaCloud and LlamaIndex workflows. Blog: https://lnkd.in/gc-uscEb LlamaCloud: https://lnkd.in/grgiv-vB Workflow: https://lnkd.in/gRsUHuV8

193

12 Comments

Most AI products are designed like tools. The best ones are designed like teammates. At LCA, we call this AX (Agentic Experience). It’s a new UX paradigm for AI agents: part product thinking, part org design. The best agents, like the best employees, don’t just execute - they clarify, anticipate, push back, personalize, and grow with you over time. Here are 8 principles we're using when designing AX - inspired by what makes a world-class teammate: 1. Transparency, Tapered Over Time Early on, agents should show their work. We want to see wtf is going on. As trust builds over time, you don't need to see 'how they got there' and the answers will do just fine. 2. AI Prompting You Great agents don’t just wait to be told what's next, they suggest next steps and flag issues after a request is done. There's nothing worse than checking in with an employee and nothing was done because 'they were waiting for instruction'. 3. Clarify Before You Commit Before diving in, a smart agent pauses to ask the right questions. Let's be honest, no one nails the perfect prompt, and a little clarification can go a long way. Less turns, more delight. 4. Pushback Is Professional Sometimes what you ask for isn’t what you need. Good agents offer better paths. Great agents will offer the better path and already have it done and demoable (Demo > Memo. h/t Aparna Chennapragada). 5. Multimodal for the Moment Text, video, voice ... great agents pick the right format for the job and can bend to what you need, when you need it. 6. Loop in Other Experts When things get complex, your agent should call in backup - other specialized agents. Like a great teammate, sometimes knowing 'who' can solve the problem better vs 'how' is 10x more effective. 7. Context is King Great teammates don't ask the same question twice. The more you work together, the better it should understand your goals, tone, and style. 8. Personalization Is the Moat Over time, a great agent knows how you think. They know what 'polished' means or what a 'report' means (to you). They know when you want things fast vs. perfect. It even senses when you’re tired, frustrated, or in flow. This level of personalization is not fluff, it’s a competitive advantage. -- We’ve spent 20 years designing for clicks, taps, and swipes but AX is different. It’s not about using the product, it’s about building a relationship with a thinking partner that gets better every time you interact. In an age with a higher quality bar for "good", an explosion of copycat products, and a major shift in consumer behaviour, products need to move from “this is a helpful tool” to “I can’t live without this.”

128

20 Comments

🚨 New UX Pilot AI Feature: generate consistent, on-brand designs with ‘Customized Themes’ Yup the days of typing hex codes are officially over No more writing #00FFJ0 or “radius 40px” into the prompt box Just set everything once inside your custom theme:  🎨 Color Settings > Primary colors > Secondary colors > Accent colors > Neutral colors > Card colors > Border and input colors > Destructive colors 📐 Border Radius and Effects > Radius > Shadows > Effects ✏️ Font Types > Sans > Serif > Mono Set it up one time Save your theme Reuse it whenever you want across new projects You already could upload your Brand Guidelines and PRDs You already could import your Design Systems from Figma Now you can also create your own Custom Themes And it just takes a few clicks/seconds to do UX Pilot just keeps getting easier to play with 🙂 Try it for free: https://lnkd.in/dYNtPzFS

81

24 Comments

B2C moves, B2B adapts. Nearly every major shift in B2B started with a consumer expectation—self-serve checkout, clean UX, mobile-first experiences. The latest bellwether? Shopify x OpenAI. AI-assisted shopping is no longer experimental—it’s here. Which means B2B is about to try and follow. And on paper, B2B looks surprisingly ready for AI-native buying: -Discovery? LLMs can already parse G2, product docs, reviews, and specs. -Selection? Agents can compare features, fill out RFPs, and suggest better-fit products. -Payment? Ramp, Brex, Stripe—no shortage of rails. But one thing breaks the loop: price. Buying B2B software still feels like haggling for a rug in Marrakech. No posted price. No logic. Just a negotiation based on how hard you push and how desperate they are. If we want AI to truly transact in B2B, pricing needs to be structured, verified, and accessible. Until then, we’ll keep pretending that “let’s hop on a quick call” is a normal way to shop.

39

1 Comment

The AI agent market is racing toward $32.5B by 2034. I’m seeing the shift firsthand at Copilot.live. Here’s what the numbers don’t tell you: $5.2 billion today. $32.5 billion in 10 years. That's a 20.1% CAGR. But here's what the numbers don't tell you: Customer service holds 45% of the market share. Makes sense. That's where the ROI is immediate and measurable. 350 million AI agent units are already deployed in 2024. The growth trajectory suggests exponential increases ahead. The real story isn't the growth rate though. It's that we're witnessing a fundamental shift. Virtual personal assistants command 30% of the market. Autonomous agents take another 25%. North America dominates, but Asia Pacific is rising fast with China leveraging government incentives and big data infrastructure. The pricing ranges from $100 to $1,000 per agent depending on complexity. Companies are prioritizing advanced NLP and machine learning capabilities. That's when the real transformation begins.

136

16 Comments