You’re doing the work that should’ve been handed off

I don’t think most designers are slow. I think we’ve just gotten weirdly good at staying busy with work that doesn’t need us. Not obviously wrong either. The kind that feels responsible and thorough, like you’re doing your job properly. That’s what makes it hard to spot.

This hit me after a client call, somewhere between work, travel, and waiting for a flight home from a Toronto offsite. Laptop open at the airport, half a drink beside me, brain already on the next thing.

It was a good call. Clear patterns. Same objections I’ve heard before. You know that feeling when you already know what matters before you even open your notes.

Then I did what I always do. Opened the call recording, scrubbed through it, started writing insights, cleaning them up, turning them into something product can use, something engineering won’t ignore, something sales can repeat.

Halfway through I stopped. Why am I doing this part manually? The thinking was already done in the call. What I was doing now was just packaging it.

So I tried something else. I had Claude structure the insights while I moved on to the next task. Not perfect, but good enough to react to and shape, good enough that I didn’t have to start from zero.

That’s when it clicked. I didn’t get faster. I just stopped doing that part. That sounds small. It isn’t. It kind of fucks with how you think about your job.

This isn’t about prompts. It’s about who still does the work

This shift isn’t really technical. It’s how you work. Most people are still using AI inside the task, helping them write, think, iterate, so they stay in the same loop and own every step. It feels like progress, so you don’t question it. That’s the trap. It’s convincing enough that you just keep doing the same shit, thinking you’ve leveled up.

Delegation breaks that loop. Once you feel it, you start noticing how much of the job is repetition: rewriting insights, reformatting deliverables, rebuilding the same thing in slightly different ways. That work feels important because it’s always been manual. But manual doesn’t mean valuable. It just means nobody questioned it.

There’s another shift that creeps in. When you stop doing the packaging, you’re forced to be sharper about the thinking. What actually matters here? What’s noise? You can’t hide behind volume anymore. That’s where judgment gets exposed, and that’s uncomfortable. And if that’s uncomfortable, the next part is worse.

Hand off the wrong part and your work gets worse

I don’t delegate everything. That’s the trap. Swing too far and your work gets hollow fast. I wrote about this in the LinkedIn post that sparked this whole thought. If I let AI generate insights from scratch, it sounds right and feels clean, but it misses the nuance you only get from being in the room.

So I keep the front half: discovery, pattern recognition, framing. That’s mine. What I hand off is everything after the decision: structuring, formatting, first-pass synthesis, anything I’ve already explained more than once. I stay close enough to correct it quickly, not babysit it, just steer it. That balance actually works, and it forced me to question something I didn’t want to question.

More effort is quietly making your work worse

This changed how I look at my role. I used to equate effort with quality: more time, more passes, more control. Teams reward that behavior, which makes it even harder to let go. Now it just looks like friction.

It forces a harder question. If this part of the work disappeared tomorrow, would the outcome actually be worse? Sometimes the honest answer is no. That’s where it gets uncomfortable. If that’s true, the next question is obvious.

The small tasks stealing your week (that you don’t question)

Next week, notice one thing. What’s the last task you repeated twice? Not a big workflow, a small thing you explained again, rebuilt again, or reformatted again. That’s usually the entry point. Not everything needs to be delegated, but something probably does.

The math gets stupid fast.

Say you spend 45 minutes packaging research after a call. Two calls a week is 90 minutes. That’s 6 hours a month. At a conservative blended cost of $75/hour, that’s $450/month of designer time spent mostly turning thinking into a shareable format.

Now multiply that across a year. That’s $5,400 in time for one tiny repeated task. Not strategy. Not judgment. Not the part where your taste or experience creates value. Just the packaging layer.

And that’s the gentle version.

If that same packaged insight helps product avoid one bad sprint, or helps sales repeat the customer pain more clearly, the value jumps way beyond the time saved. That’s the part people miss. Delegation isn’t only about saving hours. It’s about moving your judgment faster through the company.

Once you see it once, you’ll start seeing it everywhere.

Ask this before you touch the next task

Where in your workflow are you still doing work you’ve already mentally solved, you’ve explained before, and someone else could execute with direction? What would break if you stopped doing it yourself? Most people don’t like the answer.

Steal this Claude Skill for ongoing research

This is the kind of thing I’d turn into a skill instead of rewriting the same prompt every time. I was also reading this breakdown on using Claude Code for interview synthesis, and it hit the same point from a more technical angle: stop manually repackaging research every time you need to share it. Not because I want AI doing research for me, but because I want it aggregating the work I already did and packaging it for people who need to act on it.

I’m a product designer working in a
B2B SaaS environment.
I regularly run client calls, Loom reviews, user interviews, and stakeholder conversations.
I am usually in the call, so I already have context on what mattered.
Your job is not to invent insights.
Your job is to help me package and aggregate the signal
from research inputs I provide over time.
When I provide transcripts, notes, Loom summaries, call recordings, or raw observations,
help me create:
• a clean summary of what happened
• the repeated patterns across conversations
• the strongest supporting moments or quotes
• what this likely means for product, design, engineering, sales, and GTM
• what is still unclear, risky, or unvalidated
Please:
• separate observed evidence from interpretation
• call out where the signal is strong vs weak
• flag anything that sounds like assumption, not evidence
• avoid generic UX language
• write in plain English a cross-functional team can use
• frame findings in terms of business impact, adoption, retention, revenue, risk, or operational cost
Do not make the work sound more certain than it is.
Do not invent themes just to make the output feel complete.
Do not replace my judgment.
Help me distribute it.
        

That’s the useful version of AI for me. Not a replacement brain. A way to stop manually repackaging the same thinking every week.

Why some designers will get outpaced (and won’t see it coming)

I don’t think designers get replaced by AI. I think some get outpaced by people who stop doing work they shouldn’t be doing anymore.

That gap won’t show up all at once. It’ll show up quietly, one delegated task at a time.

If you want to see the full AI delegation conversation, that’s where we get into delegation, the risks of moving too fast, and what actually changes when you stop treating AI like a tool and start treating it like support.

Till next week,

Tyler White Helping designers prove the ROI of their decisions