AI Tip of the Week
Gerry, a colleague of Dona’s, wanted her to write a foreword to his book. (This happens regularly.) But this time, he took the initiative and did the thing.
Gerry gathered a bunch of Dona’s past talks, interviews, and writing—anything with her voice in it (there’s a lot out there). Then he used AI to analyze her tone, favorite phrases, and style. With AI, he drafted a foreword as if Dona were writing it, bringing together themes in the book with her perspective on the subject. He edited a bit, then sent it over to her — not with a “can you write this?” but with “here’s a draft—feel free to tweak it, but we think it sounds like you.” And it did sound like her. The voice, the rhythm, even the attitude—it was close enough that Dona just laughed and said, “Yeah, that’s pretty much what I would’ve said.” It’s Dona, so of course she re-wrote it completely, but this time she had a starting point and an overall structure – win win for everyone!
4/7/25
We’ve covered several types of AI Power Users so far, but one of our favorites is an AI Power User in Operations (Ops). The ops role is pretty much the purest form of Doing The Thing we can find. Ops is taking a vision — whether written down in a 100 page document or scrawled across a napkin — and making it reality. Let’s have a look at how we use AI to help us do ops for one of our businesses, Towards Arcadia.
In addition to Prima Dona Studios, we also have a men’s shirt line, Towards Arcadia. We haven’t focused much on it for a bit, but we have several beautiful collared shirts, hoodies, and t-shirts on towardsarcadia.com. When Jeremiah wears one of our TA creations, people keep coming up to him and yelling WHERE DO I GET A SHIRT? It’s getting to be a thing.
We’ve decided to give this Towards Arcadia thing a go. We’ll continue to work with our tailors to do bespoke shirts, but to be able to deliver beautiful TA shirts to our customers on a reasonable timeline, we’ll need to work with a shop that can do small batch. We’ll be able to use AI to help us with some of the tasks, but others we’re going to have to do by hand.
To get started, we’ll list all the things that go into our ops process. Of course, this varies for each business – but it might help if we’re specific about the sorts of things that go into our own biz. Then, we’ll distinguish which of the tasks we can augment with AI, and which are the ones we’ll need to do ourselves. Finally, we’ll prioritize all of these to get to our work.
Prioritization
To prioritize tasks, we’ll have a look at
(1) Importance / Urgency – How important is it? We’ll want to use AI for things that will prove real ROI
(2) “Gen AI-able” – Is it the sort of task for which creativity would be good? Would it be fine to have occasional randomness / hallucination? We’d want to use AI for creative, flexible tasks like drafting an email to a prospective client, not for something like regulatory filings with the SEC.
(3) Do we have our data in order? We’ll only want to use GenAI for tasks where we HAVE the data, and can USE the data.
Grounding Principles
Throughout this exercise, we’ll seek to remember our grounding principles for using AI:
1) AI doesn’t touch customer (all engagement)
2) Anonymize customer info when using AI
3) We’re transparent when using AI
With that aside, let’s get to our workflow to scale beautiful Towards Arcadia shirts!
1. Talk to people (research)
First up, the most fun part of our job, talking with current customers (of other TA shirts & hoodies) as well as prospective customers.
- Us: Talk to customers to understand their needs
- AI: Organize feedback
- AI: Highlight potential differentiators vs competitors
- Us: Make final assessments of where we’ll focus:
- Most men’s shirts are – you know, boring. So colors & interesting patterns
- If a men’s shirt has Flippy cuff ALL THE WAY UP TO THE ELBOW
- Comfortable
- Pas de pockets (because we’re not THAT type of nerd)
2. Create samples with tailors (design)
After we have a good handle on what we’re looking to do, it’s time to move out of the talking phase into the sewing phase! We’ll work with our bespoke tailors on these, making lots of one-in-the-world size M shirts!
- Us: Hand-drawn sketches
- AI: Help find bespoke tailors & rate offerings
- Us: Go in with fabric and work with the tailors until we have designs we don’t hate
3. Wear-test samples (test)
This is Jeremiah’s favorite part of the shindig. It’s time to get him & our friends into some stunning shirts!
- Us: Self-host Designs
- Us: Have potential customers self-host
- Us: Write down comments, complements, and any thoughts
4. Think Marketing (marketing)
After Jeremiah and friends have worn shirts on planes & stages, and we’ve gotten a sense of what designs resonate, it’s time to start getting the word out.
- Us: Finalize design
- AI: Suggest marketing/sales ideas based on design, customer feedback and market trends
- AI: Draft launch copy (email, site, IG)
- Us: Review AI drafts, rewrite in our voice, and schedule rollout
5. Find small-batch (research)
Once we’ve locked in our designs and tested the fits, it’s time to scale beyond one-offs. That means finding a production partner we can trust.
- Us: Define criteria (e.g., turnaround time, ethics, sustainability)
- AI: Find small batch-factories near Seattle
- Us: Meet and talk to them, and select one
6. Pattern-making size grading (design)
With our production partner in place, we need to translate our designs into scalable patterns and size options.
- Us: Deciding what sizes and cuts we want
- AI: Create a sample size chart based on audience
- Us: Test the samples with real people
7. Source fabric & notions (research)
Even the best designs fall flat without the right fabric and details—so we go hunting.
- Us: Define qualities we’re looking for (e.g., handfeel, weight, materials, sustainability)
- AI: Find high-quality shirt fabric & notions
- Us: Go get fabric from local sources (including Joann’s, which is going out of business!)
- Us: Order small quantities and wear test fabric
- AI: Organize wear-test feedback into a short report
8. Production run & QA (design)
With sizing, fabrics, and partners ready, we launch our first real batch. This is where ops gets real.
- Us: Finalize production plan and timelines
- AI: Based on previous orders and goals, predict sizing for prod run
- Us: Place factory order
- Us: Handle any adjustments with bespoke tailor
- AI: Generate QA checklist for finished goods inspection
9. Handle customer orders (marketing)
Our shirts are out in the world—now we keep the experience great for our early supporters.
- Us: Add to our e-commerce site (e.g. take pictures, write website description [with AI draft] & publish)
- Us: Pick inventory, ship to customers & handle any refunds/returns requested
- Us: Collect social commentary & feedback for use in future product development
- AI: Analyze feedback to flag common trends/issues/opportunities
Where next?
If you’re running ops—whether for a fashion line, a startup, or a side hustle—you’re juggling about 87 different things at once. That’s where AI comes in. Think of it as your tireless, chaos-tolerant operations assistant. From organizing customer feedback to suggesting marketing copy, helping you find vendors, or summarizing wear-test notes, AI can help you move faster without dropping the ball. The trick is knowing where to bring it in: use AI to accelerate research, draft creatively, and find patterns in messy data. Then, step in as the human with taste and judgment to make the final call.
As an AI Power User in Ops, your job is to prioritize ruthlessly and match the right tools to the right tasks. Ask yourself: is this task important and generative? Do I have good data for AI to chew on? Will it help me move faster or make a better decision? If yes, delegate it to your AI assistant. If not—if it’s high-stakes, relationship-heavy, or mission-critical—you keep it. Great ops is about momentum, clarity, and vibe. And using AI right means you can focus more on the parts of the work that only you can do.
AI helps us speed through research tasks, draft marketing copy without the blank-page panic, and organize feedback faster than you can say “flippy cuff.” We’re not asking it to sew a button or read Jeremiah’s mind about a shirt vibe—that’s still on us, and Dona will attest that it’s pretty challenging—but we are asking AI to surface insights, suggest options, and grease the wheels so we can focus on the magic bits: human judgment, aesthetic choices, and actually talking to people.
While it’s tempting to throw AI at everything, we have to be thoughtful. Sometimes it’s simply not possible to use AI. Others, it’s possible but still not a good idea. We’ll prioritize AI use based on ROI, creativity fit, and data readiness—because AI is only as good as the stuff we feed it. If it’s a task that needs a little flair and some pattern-matching? AI’s in. If it requires precision, personal touch, or accountability? That’s on us. This isn’t about automating the soul out of our biz—it’s about scaling the stuff that can scale so we have more time to put love into the rest.
Hope this helps!
<3 D & J
Leave a Reply