Episode 061 Raw Data

	A	B	C
1	Thanks for your interest in The Question! Sign up to participate in future questions here. This file is the raw data from Episode 061, deep dive hosted on September 25, 2025 with Ben Callahan and Natalya Shelburne. There are 46 answers. The question this week was: ----- Hello, system thinkers! It’s Monday and we’re back at it this week, with a new co host—Natalya Shelburne. In preparation for this episode, I was doing a little research on Natalya. Not only is she an author, speaker, and leader, she has two bachelor degrees (one in Studio Art and one in Psychology) and a master’s degree in Creativity and Talent Development. It was clear from our conversation that she is wildly passionate about helping the humans in her orbit develop new skills and use those skills in meaningful ways. During this chat, we kept coming back to the idea that for a while, we got good at quickly learning a rapidly changing landscape of tooling: new frameworks, new tooling, new patterns. But now, things are evolving even more. Emerging technologies are changing the way we create. The tools we can reach for are so transformative, they are measurably reshaping our craft, our ways of thinking, and our creativity. In design systems, this is felt deeply: not only are we using tools, we’re building tools for our consumers to use! This week, Natalya and I want to understand how you are navigating the changes to your workflows and how you help your consumers do the same. Welcome to Episode 061 of The Question! What’s one new skill you’ve had to learn because of a change in tooling? What’s one new skill you expect your consumers to learn in order to use your design system well? What’s one skill you hope not to need in the future because of the disruption of AI?
2	Question 1: What’s one new skill you’ve had to learn because of a change in tooling?	Question 2: What’s one new skill you expect your consumers to learn in order to use your design system well?	Question 3: What’s one skill you hope not to need in the future because of the disruption of AI?
3	How to create and manage tokens in figma when Figma launched variables, bridging the gap with unsupported tokens (ehem, percentages in Figma), and MCPs. In general, I do feel like the amount of things to learn gets bigger by the day and the hours in a day remain the same.	In some cases, they still need to improve their Figma skills so the basics of design software is the minimum requirement ideally. I would love it if they were able to connect the design system to vibe code using existing components and patterns.	Connecting components across Figma and different code bases so an update made in one place trickles down to the other automatically. Automatic accessibility checks: I click a button and checks are made and improvements suggested in a way I only need to click "accept".
4	Vibe coding in cursor and claude	Vibe coding in cursor and Claude. Ability to connect our our design system into MCP servers to support vibe coding tools to function well with the system	Documentating guidelines manually
5	Tokens Studio	Using Autolayout to maximum effect: Grid, and using min/max dimensions to control filling and wrapping behaviour.	Auditing a set of wire frames for adoption of the system, or of a component build for efficiency and adoption of naming or other system conventions.
6	Obviously the sketch to figma switch auto layout, variables and all that. That's the last big one.	System thinking. I think with ai systems and the ability to change them is going to get really easy but people will need to see how everything joins up. How does changing thing a impact b, c and d.	Writing Jira tickets or documentation. I don't think ai will solve it all but as a dyslexic person having something that can help formulate my written thoughts into something logical is / will be be a game changer. Even if I just finish it off. Writing makes me so anxious as a dyslexic person.
7	Working with tokens and testing my branches in GitLab so the devs can receive the JSON code of the components I've made in Figma/Tokens Studio.	Tokenisation	Documentating and prototyping components (I don't think AI can spare us all the work, but it can definitely make it easier)
8	Prompting and generally understand how to interact with ai tools or tools powered by AI.	I expect them to learn basic user experience design principles (still too many people who only know how to draw in Figma) as well as learning basics of software development (the way an architect studies material science and physics even though they will not assemble the building themselves). I think there is also a need for system thinking in general from a conceptual lense and or course for designers learning Figma well with all the features required to use a design system as well as building interfaces in a design tool.	Maybe quality assurance testing (including accessibility testing)
9	I had to get a lot more comfortable with design tokens when we moved into Figma and started integrating them with code. Instead of just thinking about components, I had to learn how things like color, spacing, and typography get named, versioned, and mapped directly to engineering. It shifted me from focusing on pixel-perfect screens to building a system that scales, and it’s made me collaborate with engineers in a new way.	Variables and tokens as well. However not just integration into our components in Figma but when developing outside the system such as building layouts.	I hope I won’t need to spend as much time on manual documentation work. I've actually already began doing this work with AI. Things like redlines, spec sheets, or writing endless variant descriptions, but they take me away from the strategic and creative side of design. If AI can handle those repetitive tasks, it frees me up to focus on problem-solving, storytelling, and crafting better experiences.
10	New Shortcuts	How to prompt to find with ease the information that they're looking for	Anything repetitive that I've learnt to do myself before
11	Prompt writing. Some others down stream from this, but prompt writing competency seem to be a foundation for some new technology partnering/tooling workflows.	Where to look for the answer to the questions they might have and actually ask it (before its a crisis).	Not sure i'm thinking of skill shedding so much as task shedding — manual maintenance or production based tasks that are necessary without automation eat time and budget. This also just makes me think of bloated processes that may be reliant on these types of production modes, maybe there are ways to expedite production that can help streamline review or approval processes.
12	How to effectively interact with AI to get better outcomes from both research insights and code generation	Honestly they still need to learn tokens better than they do. Also they need to learn strategies for when to or when not to use new tech with our system	Drawing pictures of UI, and spec’ing these drawings for people to build it.
13	working with code in storybook	semantic naming of styles and tokens	prompt creation - AI is annoying and created extra work
14	AI Prompting!	MCP Server connections, and prompting as well.	I'd rather not lose any of my beloved skills! But, to your question, I'd be benefited from not having to remember backwards and cross-browser compatibility issues and considerations, and expect those to either be inexistent or get resolved by themselves, through AI assistance.
15	Learning how to integrate the design system into vibe coding tools, and how designers need the system packages loaded by default.	Learn about different methods to search and discover content.	Ability to differentiate changes more easily.
16	I'm trying to find the time to learn some of the tools I've discovered here, like Storybook.	Reading. We'd really like to not have our consumers learn anything new.	How to convert a Word document to an HTML file.
17	Change management communications across a broad range of skill level and... Figma Variables	Understanding the foundations and the boundaries of what can / can't change on a whim	Deeply interactive prototyping
18	Presenting and teaching are wildly under-discussed skills for design system leaders. Building the system has challenges, of course, but the people management is the real long term challenge.	Getting our design and development teams used to the idea of central token values is our biggest challenge at the moment.	I hope not to be expected to rely on AI rather than sound design thinking and human creativity.
19	JSON for Design Tokens Prompting for AI	Figma's advanced features (using variables and autolayout mainly)	frankly, none. It's not because AI is helping that you should stop learning how to do things and how things are working
20	Use of Variables in Prototyping	Design System critique feedback and contribution.	size variable creation (too teadious )
21	Building accessibility annotations to support design system documentation for design and developers	Use correct libraries	not sure
22	More technical tools, less design-specific ones.	Thinking in semantic terms to be able to apply "low level" design system decisions.	Requirements gathering.
23	Writing, storytelling, and typing :) I'll try to sneak in three for one. With a move toward chat and prompt-based design/engineering, my need to be able to summarize and describe my intent has grown exponentially. At best these skills align with what I should have already been doing as part of the process of writing design briefs, and user stories. I've have had to describe and detail goals, need, intent, user type, use case and outcomes for my design, product and engineering partners and now that same template becomes more relevant when working with generative technology.	Accessibility advocacy. It's very easy for a component that has been documented to meet accessibility needs to be used to create an inaccessible experience. It takes partnership, meeting real users, and shared principles to get to a place where individual consumers and contributors know enough and appreciate their role and sharing the responsibility.	Repetition. There are so many smaller utilitarian tasks that I think the robots are so much better at. I don't need to create color and variable swatches or clean and prep details of a feature for handoff if there are know conventions, templates and examples.
24	Well, Figma a couple years ago. I’ve recently been digging more into the prototype feature.	We’re rebuilding our main application, and not all the developers are used to React yet.	I would love for design tools to get closer to the code. Figma is great, but it’s still just documentation.
25	getting better at writing/using APIs	How a structured theme/token object works. On the design side, the concept is often new. On the eng side, they often don't realize it's typesafe or how to use it or why it matters	keeping stories/pattern examples up to date
26	AI is a big one here. Learning how to use it and write good context to get what i need has been huge. Alongside this, the other skill i needed to get better at in new contexts is overall troubleshooting. When using AI, things can...happen. Its a great starting point for helping to execute on things like coded prototypes if you dont know a language, but to be really successful with it, ive found i need a level of language comprehension in order to troubleshoot it. When using it for overall help in writing and storytelling, ive found that I need to leverage a different part of my brain to critique and refine its outputs. No one wants to read AI written grossness, so its really important to learn how to give better directions and take from it only what you need. Ive just found this requires a different type of skills than doing all the tasks from scratch and refining your own work.	Troubleshooting (ive been big on this lately). I dont expect consumers to know everything about a system, but if they have a general understanding of a few key foundations, they can troubleshoot much faster themselves than if they were to always rely on us to help. This requires a foundational understanding of token structure, and a high level knowledge of concepts like "slots". At best, I would love for them to simply learn how to use inspect to look at component code where they can see pretty clearly whats possible.	Honestly, I hope I need all my skills still but I hope that AI can help me execute at scale better. I dont want to do the same small tasks over and over so maybe i hope i just dont need all my time saving shortcuts!
27	In-depth understanding of Figma’s branch, merge, and component propagation behaviors	How to ascertain that they’re using resources from the correct libraries	I’m not sure one comes to mind.
28	Figma Make	How to prioritize design system adoption	Being required to create PowerPoint presentations for meetings ... why are we even still using PowerPoint in 2025? I have no idea.
29	Controlling a product's density through design tokens and mode options	How to apply the desired density mode to their product	I hope not to lose a skill thanks to AI. In fact, I feel that my full skill set is needed for AI to train with our design system.
30	Using AI effectively in data base query tools. Nimble work arounds to safegaurd accuracy in the process to compensate for LLM fails.	If ew design "consumers" as product team users, I expect them to learn the basics of typography and visual hierarchy.	I hope not to need to learn 3-5 coding languages.
31	I can't think of anything at the moment.	Not really new but I have shared with devs/designers how to easily find things in Figma with the search and filter. Seeing their face light up is rewarding.	If glasses are provided (and they are stylish) or contacts, have accurate live closed captioning to assist with visual processing. It would be great to assist with foreign languages too.
32	Adapting and patience with new tools. Accepting that the reality of landscape is going to change and it’s best to stay ahead rather than stick to legacy ways of working because they have worked before.	Troubleshooting	Documentation and researching to help back up design decisions.
33	i was doing design systems during the move from Sketch to Figma. i've had to learn about Web Components as the web systems move from being in Angular, React, etc to one framework (w/ wrappers) to serve them all. now, gonna need to work on how AI integrates	need them to keep up with Figma; its out of our control that the tool keeps evolving, its got to be considered part of the job to keep up with that skill. it feels like consumers of the DS expect us to help them with that.	would love some of the really routine manual items to be handled by AI. but honestly i'm still learning what AI can do for us....our company is slow to adopt (b/c of risk issues in banking), so we're a little behind the curve
34	Prompt engineering	Emergence of design and engineering skillset	Internal meeting loops
35	Documentation for AI assistants’ refinement	Simplicity - as the design system grows I’m not sure they’re resisting the temptation of using or requesting more components/variations instead of rethinking the actual UX	Starting a completely new career
36	How to thoroughly prompt ai to get intended results more consistently	How to thoughtfully go beyond the system when they reach the edges	Writing documentation: agent follows along my process and documents the solution for me. first touch customer support: agent reliably and correctly routes common questions in slack to solution/docs
37	What tools we can use for our system based on company constraints, environments, and end goals. Understanding the different tools and vetting them to use.	Comprehension. I say this half jokingly. In order to use any design system well, people have to take the time to comprehend the documentation.	Answering questions that can be found in our documentation.
38	Patience... NOt sure about a specific skill but it certainly created the need to rethink workflows, governance and accountability from all parties involved in the design to build process	not sure	none
39	I think a major thing is AI, really having to understand how these tools work and some of the more nuanced terminology,.	I am starting to expect some of my consumers to have a general understanding of Design Tokens and the "system" as a whole. All to often our consumers think of the design system as just a set of components.	I think being able to bridge the gap between design and engineering, I can utilize AI to help me understand things that normally might take more time to understand.
40	Figma Variables. I used to used Token Studio.	Instance Swap for component slots	Can’t think of any. I enjoy using all of my current skills.
41	Figma plugin api	Designing in the same syntax as code	Generative ai
42	Quite a bit. But that’s why I love this space. There’s always growth and the opportunity to learn.	Maybe markdown or writing to help facilitate and collaborate with documentation.	I use AI to help streamline documentation and writing so maybe in the future ai can eloquently write all the docs.
43	MCP	Interpreting design over pixel perfection	Documenting the design system
44	Unsure	Unsure	I hope we’ll need less tedious manual documentation writing where you spend hours reformatting specs or updating references across multiple places. Ideally, AI will automate that lift, freeing us to focus on storytelling, decision-making, and ensuring our documentation has clarity and purpose.
45	Understanding how code files relate to each other (import/export) more intimately. Helps a lot with prompting precise changes.	Using AI to write their own documentation. Using AI and available internal tools to push their own code changes, or at least start a PR.	I think having the knowledge and ability to read technical languages won't ever go away. But on a day to day, when AI works, writing line by line isn't needed. But ability to debug/solve problems when tools fail separates us from machines.
46	How to read code.	Prompting.	Writing documentation.
47	Vibe coding	using Design tokens correctly and proactively	anything that involves repetitive manual labor
48	Shifting more into education (for our consumers), strategy and planning now that the burden of implementation has been relieved with AI tools in the mix.	To have a better grasp on design system best practices and how to make good decisions for the scalability and maintenance of the frontend. A lot of our developers are full-stack and don't have a strong sense of how to best use the design system.	Ready to let go of coding at the super low level of implementation. Would rather not spin my wheels on fixing bugs and setting up build configuration, etc.
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100