AI 2027 Conference: Practical AI for Design Teams
Explore the Dec. 1 online conference on AI-ready design systems, Figma, agentic workflows and practical lessons for UX and engineering teams planning for 2027.

AI 2027 Conference: What Design and Engineering Teams Need to Know
AI tools are showing up in design and engineering workflows faster than most teams can decide where they actually fit. A convincing demo is easy to make. Turning AI into something that works inside a real product - along with clear ownership and documentation that people can use - is a different story. What teams need are practical examples: where it works, where it breaks, and what should stay firmly in human judgment.
On December 1, 2026, SmashingConf will host the AI 2027 Conference, a live online half-day event for designers and engineers. It's geared toward hands-on topics like AI-ready design systems, AI and Figma, agentic coding, AI experience patterns, and day-to-day workflows. The goal isn't to speculate about what AI could do in theory - it's to show how practitioners work.
This guide covers who the conference is for, what the schedule includes, who's speaking, and how to judge whether it's relevant to your team. It also points out where the program leaves room for questions, including the operational and ethical issues that always come up when you're using AI at work.
What is the AI 2027 Conference?
The AI 2027 Conference is a live online event from SmashingConf on December 1, 2026. It's aimed at people working across UX, product design, front-end development, and design systems. The stated aim is to help attendees improve team workflows with AI - without treating every new tool as if it's automatically a breakthrough.
That distinction matters. In a real product team, adopting AI isn't only about whether a model can generate a screen or write code. You also have to decide what context the tool gets, how people review its output, where that work fits into existing processes, and who's accountable when the output is wrong. A good session should help you answer those "how does this actually operate?" questions - not just show a smooth prompt and a neat result.
Who should consider attending?
The program is a good fit if you're:
- A UX or product designer exploring AI-assisted prototyping or AI features.
- A design system engineer or owner working on documentation, components, or design-to-code workflows.
- A front-end developer evaluating coding agents and structured design context.
- A team lead setting practical boundaries for AI use across design and engineering.
- A practitioner looking for workflow examples, not broad predictions about the market.
You don't need to be an AI specialist. The event focuses on using AI in established design and development work. It'll likely be most useful if you already have a workflow you want to improve - because you can compare what you do today with what the speakers propose.
What the event is - and is not
The conference is framed as a working session for people figuring out how AI fits into their day-to-day roles. The organizers lean into practical sessions, tools, setups, and details that people tend to skip. There are moderated breakout rooms at the end where attendees can bring questions from their own workflows.
That said, it's not positioned as a full AI policy or ethics summit. The published topics are about design practice, Figma, documentation, agents, and design systems. If you care about privacy, copyright, accessibility, or human accountability, bring those questions - just don't assume they're guaranteed to have dedicated slots in the published agenda.
It's worth understanding that scope before you register. This event is meant to help teams make better day-to-day decisions about work - not to solve every organizational question that comes with adopting AI.
What will the agenda cover?
The agenda runs from designing AI experiences into prototyping, documentation, and design system workflows. That ordering keeps the focus practical: start with the product experience, then explore how to build or prototype it, and finally look at the systems and context you'll need to support work across a team.
The published schedule is:
- 7:45 a.m. - Doors open
- 8:00 a.m. - Welcome
- 8:10 a.m. - Designing AI Experiences in 2027
- 8:45 a.m. - Prototyping AI Functionality
- 9:20 a.m. - Nobody Will Read Your Documentation
- 9:40 a.m. - Design Systems and Figma MCPs
- 10:15 a.m. - Alternative Workflows for AI and Design Systems
- 10:50 a.m. - Real-World AI and Design Systems Workflow
- After the talks - Two moderated breakout sessions, then closing remarks
For the 8:00 a.m. start, the published time-zone equivalents are 8:00 a.m. in San Francisco, 11:00 a.m. in Toronto, 5:00 p.m. in Paris, 8:00 p.m. in Abu Dhabi, and midnight on December 2 in Beijing. Double-check the event page against your calendar - the doors open 15 minutes before the welcome.
From AI product patterns to prototypes
The first two talks focus on AI as part of the user experience. Vitaly Friedman's Designing AI Experiences in 2027 covers design patterns for AI experiences. Anna Arteeva's Prototyping AI Functionality focuses on prototyping and agentic coding in professional design practice.
They connect, but they're not the same problem. A prototype helps a team test an interaction before committing to a build. Product design patterns help a team make decisions about how AI should behave when users hit uncertainty, incomplete answers, or the need to correct a result. Both are more useful when they're treated as parts of a product workflow - not just isolated demonstrations of what a tool can produce.
Documentation, Figma and design system workflows
The second half shifts toward the shared infrastructure that supports design and engineering. Luke Murphy's talk, Nobody Will Read Your Documentation, looks at what coding agents need from design system documentation. Christine Vallaure covers design systems and Figma MCPs. TJ Pitre presents alternative workflows for AI and design systems. Brad Frost rounds out the program with a real-world AI and design systems workflow.
Underneath all of it is a practical constraint: AI output depends on the context it's given, and on how that context is structured. If your documentation is hard for people to use, it's unlikely to become suddenly useful to an agent just because it's connected to another interface. Teams still need to decide what information is authoritative, how it's kept up to date, and how design intent maps to implementation.
The agenda also doesn't promise a single universal setup, and that's a plus. If the sessions show real trade-offs between workflows, you'll be able to map the ideas to your own team size, system maturity, and delivery constraints. A design system used by a small team won't have the same constraints as one used across a large organization - and an AI-assisted workflow has to fit the system you can realistically maintain.
Who is speaking at the conference?
The six speakers bring experience across product design, UX education, design systems, front-end work, and AI tooling. That mix matters because AI workflows cut across roles: designers may prototype with code, engineers may need design context, and system owners have to keep shared patterns understandable and current.
Vitaly Friedman and Anna Arteeva
Vitaly Friedman works on UX, front-end, and performance problems, and runs workshops on AI design patterns. In Designing AI Experiences in 2027, his focus is on the product experience side of the program. The useful question isn't "how do we add a model to an interface?" It's "how do we make an AI feature understandable and usable?"
Anna Arteeva is a product design leader and agentic coding instructor with a front-end background. Her session, Prototyping AI Functionality, focuses on how designers build technical fluency and use coding agents beyond one-shot prototypes. A key distinction here: a prototype can help you test an idea or communicate expected behavior, but that doesn't automatically mean the feature is reliable enough (or ready enough) for production.
Christine Vallaure and TJ Pitre
Christine Vallaure is a designer and educator who founded Moonlearning.io. Her work sits at the intersection of design and code, and her talk is Design Systems and Figma MCPs. It's relevant if you're trying to connect design tools and structured system information to AI-assisted workflows.
TJ Pitre is Product Architect for AI and Design Systems at Figma and founded the design systems consultancy Southleft. His session, Starting Anywhere: Alternative Workflows for AI and Design Systems, is built around different ways to approach the problem. It's a useful framing for teams that can't simply swap tools and start fresh - or rebuild their design system from scratch.
Brad Frost and Luke Murphy
Brad Frost is a design system consultant, web designer and developer, and the author of Atomic Design. His talk, Real-World AI and Design Systems Workflow, is about applying AI to design system work. The practical value depends on whether he clearly explains the workflow, its inputs, and where people should review (or intervene) rather than blindly accept outputs.
Luke Murphy, Principal Design Advocate at zeroheight, presents Nobody Will Read Your Documentation. His talk focuses on design system documentation and what coding agents need from it. zeroheight is joining as a partner; its platform is used to create, manage, and publish design system documentation.
Overall, the speaker lineup covers both the design and engineering sides. If you're looking for something you can actually test, the best takeaway will be a method you can adapt - rather than a tool recommendation that ignores team size, system maturity, or delivery realities.
How can teams get practical value from the event?
A conference is easier to judge when you connect it to a specific problem you're already dealing with. Before you attend, write down one workflow that causes friction. It could be prototyping an AI feature, finding the right guidance inside a design system, keeping documentation current, or giving a coding agent enough context to produce genuinely useful work.
Then use the sessions to evaluate the workflow, not just the tool. Ask what information goes in, what the tool produces, how you check the output, and what you do when it's wrong. That helps you separate a repeatable process from a demo that only works under perfect conditions.
Bring a concrete workflow question
There are two moderated breakout sessions. To make that time count, come prepared with a short description of your situation:
- Name the task. What are you trying to design, document, or build?
- Describe the current process. Where does the work slow down or require repeated effort?
- Identify the constraint. Is it missing context, review time, tool access, or inconsistent design guidance?
- Define a useful result. What would improve speed or quality without adding a bunch of extra maintenance?
- Ask about failure modes. How do you catch inaccurate output, protect sensitive information, and keep a person responsible for decisions?
That keeps the discussion grounded. "How should we use AI?" is too broad for a breakout. "How can an agent use our component guidance without generating unsupported patterns?" gives other attendees something specific to respond to.
Evaluate the business impact, not just the novelty
If you're a team lead, the business case isn't that AI is new. It's whether a workflow change improves delivery, consistency, or the quality of user-facing work enough to justify the cost. And the cost isn't only subscriptions. You may need to restructure documentation, review AI output, train people, and maintain integrations.
Start small. For a prototype workflow, compare how long it takes to create and revise something testable. For documentation, track whether contributors can find and apply the right guidance. For generated code, look at review time and how often you end up doing substantial rework. These metrics won't capture every benefit, but they do make trade-offs visible.
Keep ethical and operational risks part of the same evaluation. Consider what data the tool receives, whether people can tell when responses are uncertain, who reviews high-impact decisions, and how errors get corrected. The conference's advertised sessions don't promise to settle all of that - but you should consider it when deciding whether an AI workflow is ready for real work.
Key takeaways
- The event is aimed at designers and engineers who want practical AI workflows, not general predictions.
- The agenda spans AI experience design, prototyping, documentation, Figma, and design systems.
- The speakers bring different perspectives - no single workflow will fit every team.
- Bring a real process problem to the breakout rooms and ask about inputs, review, and failure cases.
- Measure improvements against delivery and quality, while factoring in maintenance and risk.
Conclusion
The AI 2027 Conference is a focused chance to see how AI is being used across design, prototyping, and design system work. Its real value should come down to specifics: how speakers structure their workflows, what context they feed into tools, and where humans remain responsible for decisions.
If you work in UX or engineering, don't attend just to be inspired. Bring one real problem and a clear way to judge whether a proposed approach would help. That will give you a better chance of leaving with something testable - and a more realistic view of the trade-offs than you'd get from a tool demo alone.