What is an agentic design system?

An agentic design system is an AI-powered design system that uses AI agents to automatically observe, detect, suggest, fix, and learn from issues like accessibility regressions, token misuse, and documentation drift. This reduces maintenance time from 30-40% to automated workflows.

What are the five agentic flows for agentic design systems?

The five agentic flows are: Observe (monitor design system health), Detect (identify issues like accessibility regressions and token misuse), Suggest (propose fixes), Fix (automatically resolve issues), and Learn (improve over time based on patterns).

How do you encode intent in a design system?

Design systems need to encode intent, not just values. This means validating meaning across tokens, components, and documentation to ensure the system understands the purpose behind design decisions, not just the visual values.

Who should attend this session on agentic design systems?

This session is for design system maintainers, product designers, and product (design) leaders who want to reduce maintenance overhead and implement automated workflows for design system health.

How much time can agentic design systems save?

Most design system teams spend 30-40% of their time on maintenance tasks like fixing accessibility regressions, detecting token misuse, and updating documentation. Agentic design systems can reduce this to near zero through automated workflows.

What is the difference between agentic design systems and traditional design systems?

Traditional design systems require manual maintenance and constant human oversight. Agentic design systems use AI agents that automatically observe system health, detect issues, suggest fixes, implement solutions, and learn from patterns to continuously improve without constant human intervention.

Can agentic design systems work for any company size?

Yes, agentic design systems can be implemented at any company size. The session covers realistic implementation paths that scale from small teams to large enterprises, with practical approaches for different organizational contexts.

Conference Agenda March 2026

Complete schedule for Into Design Systems Conference 2026. March 19-20 online. 14+ sessions on AI Design Systems, Vibe Coding, MCP integrations. Speakers from WhatsApp, Miro, Indeed, Adobe, GitHub, Figma.

Day 1 - March 19, 2026

Prototyping for the unknown by Nate Baldwin (5:00 PM - 6:00 PM)
WhatsApp Web: Reclaiming UI Excellence through Vibe Coding by Sebastian Rousseau (6:00 PM - 7:00 PM)
Product Primitives: The New Material of Design System by Yesenia Perez-Cruz (7:00 PM - 8:00 PM)
The path to an AI enabled design system: Our approach and lessons learned by Andressa Lombardo & Eddie Machado (8:00 PM - 9:00 PM)
Agentic Design Systems by Romina Kavcic (9:30 PM - 10:30 PM)
I'm not an engineer but I ship code: How designers can ship production code and work like an engineer by Freya Stockman (10:30 PM - 11:30 PM)
Context > Probability: Design systems as AI infrastructure by Jesse Gardner (11:30 PM - 12:30 AM)
Encoding governance on agentic design systems by Cristian Morales Achiardi (12:30 AM - 1:30 AM)

Day 2 - March 20, 2026

Building real design systems with agents by Jan Six (5:00 PM - 6:00 PM)
Vibe coding with zero drift: from Figma to Storybook to Production by Shuaiqi (Mr.Biscuit) Sun (6:00 PM - 7:00 PM)
Machine-Readable Design Systems for MCP and LLMs by Diana Wolosin (7:00 PM - 8:00 PM)
From falling for markdown to solving real problems with scripts by Laura Fehre (8:00 PM - 9:00 PM)
Ship It! Vibe Coding Your First Figma Plugin by Davy Fung (9:30 PM - 10:30 PM)
Designers Who Ship: Building a Real Plugin in 48 Hours with AI by Casandra Sandu, Christoph Hellmuth & Raquel Pereira (10:30 PM - 11:30 PM)
AI Without the Chaos: Context-Based Design Systems to the Rescue by Brad Frost, Ian Frost & TJ Pitre (11:30 PM - 12:30 AM)

Agentic Design Systems

Design systems demand constant maintenance. Accessibility regressions, token misuse, documentation drift... Most teams spend 30-40% of their time just keeping things working. But a new approach is emerging: agentic design systems. Systems built for AI collaboration, where agents observe, detect, suggest, fix, and learn. I will walk you through agentic flows, and be honest about what works today vs. what's hype.

Key Learnings

Why your design system needs to encode intent, not just values and how to validate meaning across tokens, components, and documentation
The five agentic flows that enable automated design systems (Observe → Detect → Suggest → Fix → Learn)
A realistic implementation path for any company size

Who is this for

Design system maintainers, product designers, product (design) leaders

Speaker

Romina Kavcic, Design System Lead at The Design System Guide. Romina Kavcic is a design leader with 17+ years at the intersection of product strategy, design systems, and AI. She builds scalable foundations, coaches teams, and shares what's next at thedesignsystem.guide and aidesign.guide.

Tags: Design Systems, AI

Agentic Design Systems

About this Session

From my session, you will learn:

Who is this for?

Presented by:

Romina Kavcic

Secure your ticket now

Solo Pass

Team Pass

AI Bonus content included