A few months ago I landed on a company's website while doing research for a client, the kind of site with a lot of information and very little of it useful. Product pages, certifications, a history section, all of it there, none of it telling me what the company actually did or who it did it for. So I did what I do now without thinking twice. I asked an AI chatbot to summarize the site for me. Tell me what this company does.

It worked. But it also struck me as a strange thing to need. A visitor shouldn't have to hand a company's own website to a machine to get a straight answer.

The New Front Door

My search behavior changed sometime around when AI summaries started appearing at the top of search results, and I began trusting them enough to act on them directly. Now I use AI tools daily, for due diligence on new leads, competitor research, understanding our own ICPs (ideal customer profiles) more deeply. I go through results the way I used to go through a Google search, starting at the top and working down, building a shortlist from whatever surfaces first.

What I've noticed is a gap between how AI describes a company and how the company describes itself, particularly on the pages that matter most, the homepage, the landing pages built for a specific audience. Those sections are often filled with language that could belong to almost any business. AI-driven innovative solutions to power your business forward, or some version of it. It doesn't tell a visitor, or an AI, anything real.

AI Exposes Strategic Confusion

AI is becoming a new front door for business, shaping how buyers understand a company before they ever visit its website. (Photo: Matheus Bertelli/Pexels)

We ran this test on our own site at Level™ and didn't love what came back. Some earlier SEO work had pulled our language sideways, away from plain description and toward something more generic. We're now rewriting our FAQ section with the actual questions our clients ask, and pulling Brand Strategy out from under Brand Identity, where it had been buried as a subset instead of stated clearly as its own service.

The confusion wasn't something AI introduced. It found something that was already there. This shows up constantly with the OEM and supply chain manufacturers we work with, companies that talk about their products at a level too high to pin down what they actually build, for whom, and why it matters. Ask AI to compare a few of them and the summaries start to sound interchangeable, full of the same handful of words. Innovative. Quality-driven. Words that don't tell anyone anything, and that you couldn't use to explain the company to your parents.

Structure Is Part of the Brand

Some of the companies we work with are barely visible to AI at all, not because they're hiding anything, but because there simply isn't enough written down. Long-established suppliers with a website nobody has touched in years. Newer companies that haven't seeded much information yet. And for companies with more content on their Chinese site than their English one, AI seems to draw more heavily from whichever version has more to work with.

There's also a risk in how AI fills the gaps when a company hasn't given it enough to go on directly. It may reach for a partner's site, a media mention, an old event page, and build its answer from that instead. It's a specific feeling, hearing a summary of yourself assembled by someone who doesn't know you very well. You sit there thinking, that's not quite it.

Trust Requires More Than Claims

A company’s website was once its digital front door. Today, AI may introduce the business before a buyer ever arrives. (Photo: Timothy Huliselan/Pexels)

We've started building confidence scoring into our own research process, asking how deep the sourcing goes, whether it's just a website or something broader, and treating the input as light, medium, or heavy depending on what's actually out there. It matters because AI will confidently hand you an answer built on thin information without telling you it's thin. We've caught it inventing or blurring details more than once. We now use multiple tools to check each other, and a person still reviews the result before we trust it.

I used to think Taiwan's discipline in engineering and manufacturing would translate naturally to this, that the same precision applied to communication would be an advantage. I'm less sure now.

I keep seeing technical teams who are precise about the wrong things, or precise about only half the thing. A battery gets described by capacity in milliampere hours and chemistry type, lithium ion, full stop. What's missing is the translation, how many hours that actually runs before someone needs to think about it again. A solar panel gets described by efficiency percentage and wattage output, accurate numbers that mean little until someone says how many appliances that powers, or what it actually saves over a year. The specs aren't the problem. Stopping at the specs is. Communication tends to sit low on the list internally, well behind engineering and operations, and it often falls to whoever is available rather than someone thinking about how the work sounds to a buyer instead of an engineer.

Communication has always been the harder skill here, regardless of how AI enters the picture. If anything, AI just makes that gap easier to see, and harder to explain away.

The simplest thing a business leader can do this week is put themselves in their customer's seat and search for the problem they solve, then read what comes back. Then ask AI directly what it thinks the company does. The answer won't be comfortable. It will be accurate.The answer may not be comfortable. But it will be revealing.

The thoughts above come from my own time and experience working in Taiwan, not a claim to speak for the place as a whole. Just one perspective, offered in case it's useful.

(This piece reflects the author's opinion, and does not represent the opinion of CommonWealth Magazine.)

CommonWealth Magazine welcomes op-ed submissions. Please send your article proposals to [email protected]

About the Author:

Joshua Roberts is the Founder of Level Interactive™, a Taipei-based brand strategy and design consultancy helping B2B and technology companies get clear on their story and communicate it well. A former Global Creative Director at Acer, he has worked across agency and in-house roles over more than 20 years in Taiwan. Outside of client work, he spends his time hiking, traveling, and practicing close-up magic, a habit that goes back further than his design career does. Contact: [email protected] or connect on LinkedIn.

Have you read?

• How Taiwan Can Build Global Brands Without Losing Itself
• Line Logic: Why Trust in Taiwan is Earned Not Announced
• What Hiring Reveals About a Company’s Brand

Uploaded by Ian Huang