AI Is Agreeable. That Doesn't Mean It's Right.
How to use AI without compromising your judgment, your product, or your credibility.
I love working with AI. I use it almost every day, and one of the things I appreciate most is how beautifully it can organize my sometimes very scattered thinking. But I've discovered something about my extraordinarily helpful digital assistant that I suspect many of us have encountered as we've begun figuring out how to work with these tools.
It can be a bit of a yes-(wo)man.
I can toss out a half-formed thought, make an offhand comment or suggest something I haven't fully considered, and suddenly I have a beautifully organized argument supporting an idea that probably deserved a few more questions. Just because AI agrees with me doesn't mean either of us is right.
That's what I want to explore in Part 2 of this series. Not whether AI is good or bad for a product brand, but how we use it to become more knowledgeable, more productive, and better at what we do, without handing over the judgment that makes our work valuable.
My Very Helpful, Occasionally Misguided Assistant
I'm not an AI expert. I'm a designer and product development consultant who has spent decades bringing apparel to market. I've simply become an enthusiastic user of a tool that has found its way into almost every corner of my life.
One of its greatest benefits is helping me organize information and ideas. Anyone who knows me knows that one thought leads to another, which opens up three more possibilities. They're usually connected, but putting everything into a useful structure takes work.
Recently, I used AI to plan a trip to Paris. I supplied my hotel, dinner reservations, shopping interests, and recommendations I'd collected, then asked it to organize a practical itinerary. The first version was helpful, but we went back and forth. I wanted more time for shopping, smaller independent stores, and recommendations that reflected my taste rather than a generic list of popular boutiques.
There were suggestions I loved. Others? Let's just say we had some differences of opinion about what constitutes interesting fashion.
My job was to be clear about what I liked, what I considered tacky versus genuinely interesting, and why certain recommendations missed the mark. In many ways, it reminded me of working with a design assistant or another team member. You don't simply tell someone that an idea isn't right. You explain your thinking so they can better understand what you're trying to accomplish and apply that understanding to the next task.
With AI, that meant adding context, challenging recommendations, and refining the itinerary until it reflected the trip I actually wanted. I could also set preferences that might make future planning more useful, though I would still need to check whether it carried those preferences forward accurately.
AI organized the plan. It didn't decide what kind of plan I wanted.
I use it similarly in my teaching at FIT, where I might ask it to test the clarity of an exercise I created, make it more engaging, or challenge students to apply what they've learned rather than simply repeat information. In my consultancy, it's become a useful sounding board for organizing research, reviewing documents, structuring proposals, and turning a stream of consciousness into something coherent.
In each case, AI helps me work through information and possibilities. The objectives, priorities, and decisions remain mine. And that distinction becomes especially important when the subject moves into territory where I may not have the same depth of experience.
When Agreement Starts Looking Like Expertise
I've learned to ask it to challenge my assumptions, question what I may have overlooked, and explain why it's recommending a particular direction. Otherwise, it can take one offhand comment and build an entire recommendation around something I never intended to establish as fact.
To be fair, AI can challenge us. But it doesn't always recognize when a seemingly minor detail should trigger a much bigger conversation. That's especially important for founders using it in areas where they don't yet have much experience.
Imagine you're developing a supportive bra and ask AI to recommend a fabric. You might receive a perfectly plausible answer, complete with suggested fiber content, stretch characteristics, and potential suppliers. It might even pull together information from hundreds of existing products and research on emerging textile technologies.
But what size range are you developing? What level of support is required? How will the material behave under tension or after repeated washing? What construction will you use, and how will all those elements work together?
AI can synthesize an enormous amount of information about bras. What it cannot bring to that recommendation is firsthand experience of how those fabrics feel, stretch, recover, and perform on different bodies.
Someone who has spent time developing, manufacturing, fitting, or wearing these products may recognize immediately that there isn't enough information to make a responsible recommendation. And that's the important distinction: recognizing what's missing is part of expertise.
I've been thinking about this in terms of learning tennis. You can spend hours hitting a ball against a wall and improve your coordination and consistency. But if your grip or swing is wrong, the wall won't tell you. You may simply become very good at practicing the wrong technique. In other words, practice makes permanent, not perfect.
A good instructor notices what you're doing, asks questions, and introduces information you didn't know you needed. Playing against someone more experienced takes that learning further, forcing you to respond to situations you hadn't anticipated.
AI can be an extraordinary practice partner. But exposing your thinking to people, experiences, and evidence that challenge what you already believe is enormously valuable. The goal isn't to replace one source of answers with another. It's to become better equipped to recognize which answers deserve further investigation.
When the Product Is Real but the Picture Isn't
A friend recently raised another question that brings us back to the conversation we started in Part 1. What happens when you've done everything right, but want to use AI for a shortcut at the very end?
,
Imagine you've designed a beautiful swimsuit, selected the materials, worked through the fit and construction, sampled, tested, and refined it. You've even invested in professional photography of the actual garment. Now you'd like to use AI to show that swimsuit styled with a coordinating sarong or a pair of jeans, perhaps in a setting you didn't have the budget to photograph.
The product is real. The design is yours. The development work has been done. Can using AI to tell that story make the product feel less authentic?
That's a great question, and it doesn't have a simple answer. I can certainly see the appeal, particularly for a small brand without the budget for multiple photoshoots. There's also a meaningful difference between using AI to show the versatility of an existing product and using it to make something look like it hasn't been developed.
But another consideration is customer trust.
If a customer knows one image is AI-generated, does she start wondering about the others? Is the fabric texture accurate? Does the garment really fit that way? Is the color representative? Is the product itself exactly what she's seeing?
I don't pretend to know precisely how customers will respond as AI imagery becomes more common. That's a conversation worth having with people who specialize in marketing, e-commerce, and consumer behavior. From a product development perspective, I do know that if you've invested in making a genuinely good product, you should be careful not to let how you present it undermine confidence in the work you've done.
For fit-critical apparel, that's especially important. Customers are making purchasing decisions based on what they believe a garment will do for their bodies. A swimsuit that appears to offer more coverage, support, or shaping than the actual garment is setting an expectation the product may not meet. And those expectations have consequences beyond the image itself.
Authenticity Is About More Than the Tools You Use
I don't believe using AI automatically makes a brand less authentic. We've always used tools to communicate ideas, from illustration and photography to retouching and digital presentation. What matters is what we're communicating and whether the customer can trust it.
But just because something is technically possible doesn't mean it's the right decision for your brand. Your imagery is part of your relationship with your customer, and I wouldn't want to put trust at risk simply because a shortcut became available.
The same principle applies to how founders use AI behind the scenes. I'm not going to offer a master class in prompting. People far better qualified can do that. But from the perspective of someone who has spent her career developing products, there are a few guardrails I'd recommend:
Use AI to expand your research, not replace it. Explore competitors, materials, technologies, and market opportunities, then verify what matters through credible sources and real-world investigation.
Be cautious when an answer sounds too easy. Product development involves trade-offs. If a recommendation doesn't account for fit, materials, construction, cost, and your intended customer, there's probably more work to do.
Protect the accuracy of your product imagery. Whether you're using photography, retouching, or AI, customers should receive a product that reasonably matches what you've shown and promised.
Seek out perspectives that challenge your assumptions. That might come from a supplier, a customer, someone on your team, an experienced specialist, or your own testing and research. The important thing is to expose your thinking to information you may not have considered.
Use AI to prepare for more productive conversations. Understanding terminology, organizing objectives, and identifying questions can help you make better use of the knowledge and experience available to you.
None of this requires you to become an AI expert. It requires thoughtful consideration of the business you're building and the promises you're making to your customer.
The Goal Isn't Just to Work Faster
One thing I've valued throughout my career is working with people who know what I don't. Sometimes they reinforce an idea, sometimes they improve it, and sometimes they explain why something that looks wonderful on paper simply won't work. Often, they do all three.
Those conversations aren't obstacles to the design process. They're part of what makes the product better. And I believe AI can make them even more valuable by helping us arrive better informed, more organized, and more prepared.
If you're building an apparel brand and want to make sure your development strategy is working for your brand, not against it, that's exactly the kind of clarity we work through together. Book a call here to talk through where you are and what your product actually needs.
Comments