Ep. 25 Found in AI: If No One Clicks Anymore, What Is a Website For?
In today’s episode of Found in AI, I sit down with Karl-Gustav Kallasmaa, founder of Attensira, to unpack a question more marketers are quietly running into this year:
If AI is now the primary interface for search, what role does the website actually play?
We explore what Karl calls the death of the webpage — not as a dramatic end, but as a structural shift. As AI systems become the layer that interprets, summarizes, and recommends information, the way brands communicate clarity, trust, and relevance is changing fast.
This conversation digs into how AI engines read websites, what information they prioritize, and why many sites are doing far more work than machines actually need.
We cover:
- Why websites are becoming a reference layer — not the primary destination
- What information AI systems actually need from brands (and what they ignore)
- Why clarity is starting to matter more than clever branding
- How writing for machines and humans requires different tradeoffs
- What “machine-readable” really means for B2B and e-commerce brands
- Why content volume still matters — but only when it answers real questions
- How outdated or vague information forces AI systems to guess (and why that’s risky)
Summarized Transcript: If No One Clicks, What is a Website For?
Cassie Clark
If AI is now the first place people go for answers, what happens to the website?
That’s the question that kicked off this week’s conversation.
Hey, I’m Cassie Clark. I’m a fractional content strategist and the host of the Found in AI podcast, where I help marketers and founders learn AI search and GEO strategies so we don’t get lost in this new wave of user search behavior.
Recently, I sat down with Karl, founder of Attensira, to talk about what he calls the death of the webpage. Not in a dramatic, “burn it all down” kind of way — but in a very practical one.
As AI agents become the default interface for search, planning, and shopping, the way information is structured — and who it’s really written for — is starting to change.
In this episode, we dig into whether websites are still the primary source of truth or more of a reference layer, what information AI engines actually need from brands, and why clarity matters more than clever branding.
If you’re thinking about AI search, GEO, or how your brand shows up when fewer people are clicking, this conversation will likely spark a few useful questions and maybe challenge how you’re thinking about your website today.
Okay, let’s dig into it.
Karl’s Background: What Attensira Does
Karl:
Hi everyone. My name is Karl, and I’m the founder of Attensira. We’re a software company that helps brands stay visible in AI search by giving you all the monitoring, optimization, and content writing tools you will ever need.
Cassie:
Yeah, I think right now that is definitely a tool everyone should be looking into for sure.
Karl:
Yes. And if you didn’t catch it in 2025, you’ll have to do it in 2026 because it’s going to get even crazier.
What Does “The Death of the Webpage” Mean?
Cassie:
When you sent me a message, you said you wanted to talk about the death of the webpage. What does that mean in Karl’s definition?
Karl:
In my definition, it means… I don’t think for most businesses you necessarily need a webpage in the same way. The death of a webpage means the whole structure of a webpage should change.
When I looked at your website, in my opinion it has only two features: who am I, what kind of services do I have, and please contact me — here’s an email form.
That same information can be communicated in hundreds of words for the agents. But because humans need pictures and backstory and everything else, you’re providing a lot more context.
If we assume we’re going to have a lot of agents running around on the web, then the death of a webpage means you have to make sure you can communicate the information in as few characters as possible to the agents.
So going back to web 1.0, where webpages were simple HTML — I’ve seen a few sites where they have two different versions of the page: one for humans and one for machines.
For machines, it’s very clear what the page does if you just look at the machine code. For me to trust you, I might want images. But for machines to trust you, they don’t have to see images. That’s what I mean.
If We’re Writing for Bots, What About Humans?
Cassie:
That was going to be my next question — if we’re only writing for the bots, what about the people that show up?
Karl:
For most people, I think the default interface they’ll be using will be chat — Gemini or whatever LLM provider — or they might be using a different app altogether.
Let’s say you’re using a travel planning software, and that software needs access to a guide. Your webpage will “live there.”
People will still show up, and you should still have a website for them. But that won’t be the default way people interact with it.
I remember when web development was hard — adjusting for web and mobile, keeping things fast, all of that. It’s a hassle to maintain.
If you use a platform, it’s way easier.
What Should Businesses With Traditional Websites Do?
Cassie:
For businesses that have traditional webpages right now, what would you suggest?
Karl:
If you have a WordPress site, ideally there should be a plugin — because it doesn’t make sense that an individual business owner figures it out themselves.
But if you are a developer, there are plenty of things you can do. In the instructions of your website, you can say: “Hey, we have this platform site that’s only meant for agents so they would have access.”
There’s also a marketing angle here. I’m not building anything in this space, but there’s a huge marketing need.
If you think about the era of SEO in the early 2000s, it was keyword stuffing — it kind of worked. Then everyone did that, and the next stage was: fastest website wins.
Right now, agents are doing a lot of work trying to mimic being human. That wastes resources.
A good way to compete — and honestly save the environment — is to reduce the size of a webpage.
So as a small business owner, I wouldn’t do much besides researching platforms out there and seeing if they offer anything.
And the second thing is: always think about machines. The joke is they’re human too — but think about machines when you’re writing content. Do they understand what you mean? Sometimes you’re saying too much.
What Information Is Non-Negotiable for AI Systems?
Cassie:
When you’re thinking about it like “give as much information as possible and nothing more,” what are the non-negotiable pieces of information a brand should communicate so these engines understand what the brand is about?
Karl:
It depends on the brand. Let’s take e-commerce.
You should communicate materials, price, volumes — like the term sheet you’d send — but maybe not more.
For B2B, you have to communicate trust signals, safety, and so on.
How I approach it is: LLMs are some of the smartest things on the internet, but they don’t necessarily want to visit your website if they can avoid it.
There will be an increased push toward efficiency. They’ve already saturated most of their markets.
So if they can use information from 2022, they will — unless you provide something novel that’s useful.
For example: when is your business open? If you have a calendar, it should be up to date.
What kind of information would a machine need to know about my business? Otherwise it will guess — and that’s the last thing you want.
Let’s say you offer refunds as an e-commerce site, but there’s a minimum order fee of 20 that’s hidden. Then the LLM will follow industry practices and guess.
What I’ve seen with customers is: if you provide better text to machines, you will rank higher.
And machines and humans think alike — as weird as it sounds.
If I were describing you, “who am I,” I can tell you in two sentences. I wouldn’t mention what I studied or my favorite color. It’s irrelevant. You just got the answer as soon as you could.
Clarity vs Branding: What Matters More Now?
Cassie:
That makes a lot of sense, because when AI engines started becoming a thing, people kept saying your prompt has to be specific and clear — otherwise you get a mess as an answer.
So what works here? Are we shifting to clarity over branding? Do we still need branding?
Karl:
Clarity. I’m team clarity, because branding is harder to quantify.
Values matter — but if you say you’re environmentally friendly on your website and you don’t communicate it elsewhere (product labels, materials used), then you’re not really environmentally friendly.
At the end of the day, you want to be somewhere in the answer.
Then when a human visits, they’ll read about the brand.
I’m more interested in utility-first. But you shouldn’t neglect brand either.
Sometimes people don’t know all the solutions out there. Like if I mention a design tool, you instantly say Figma or Canva — you don’t research the top 10. You just search for Figma.
So you still need branding — but I’d focus on clarity.
It’s hard to do effectively because models know so much… and yet they still can’t count certain things correctly.
The Hard Part: Positioning and Focus
Cassie:
I think the thing that’s going to be difficult is being super clear on your positioning and who your product is for — because that requires everyone to focus on what they’re actually doing.
Karl:
Yes. And there’s also something you can do if you have a little bit of code.
On traditional websites, you can tell one set of information to the LLM and another set to humans. You can front-load context for machines.
For example, you can say: this product works best for these types of people. But if you put that on a product listing page for humans, it might feel irrelevant.
Do Marketers Need to Learn to Code?
Karl:
My advice is: learn how to code. It’s not that hard.
I studied economics. But I realized these days it’s quite easy to learn basic coding.
Basic stuff can unlock a lot. It’s like learning to write. Yes, you can get by without it — but you probably shouldn’t.
Cassie:
Oh no. I literally put in a blog post today that I have no plans to learn how to code. And now you’re telling me I probably should.
Karl:
Maybe not. But coding allows you to estimate how long it takes to customize things.
As a brand, you need some sort of advantage over others.
I had this issue — I had ideas, and “vibe coding” wasn’t a thing, so I had to learn.
If you don’t know how to code, you have to pay other people for things like lead gen or email writing — or you spend time learning.
Even “vibe coding” is brilliant — it’s the future — but there’s still a place for traditional coding.
Practical Tips: What Brands Should Do for AI Search
Cassie:
Final question. Do you have any tips for e-commerce brands or B2B SaaS who want to get ahead of AI search? What should they absolutely do to be as machine-readable as possible?
Karl:
First: the game you’re playing is the content game — and volume matters.
Obviously don’t produce slop. My definition of slop is content that’s low value.
If there are 10 million e-commerce sites, we shouldn’t have 10 million pages describing the same average stuff. Like: “the rise of free shipping.” Probably not.
But there are so many questions that haven’t been answered — or the content out there is out of date. And by out of date, I mean it hasn’t been written in the last three months.
So my first advice: write as much as you can. Use AI — but still read it. Make sure it’s valuable.
And don’t worry if the answer is only 100 words. That can still help a lot.
For e-commerce, people have questions about fabrics or your brand. Answering those helps AI understand what your brand is.
For B2B SaaS, content also communicates trust. People do more research.
It’s easier than ever to write content.
Final tip — I’m sad that it works — create lists where you say you’re the best at X, put yourself at the top, and publish many of those lists.
The more lists you’re in, the more likely you’ll appear.
These systems are not always that smart. A lot of the time they’re predicting the next word.
I remember reading a study that said: the more lists you’re in, you have a higher chance of appearing.
For B2B SaaS: don’t neglect content.
You can do funny videos on TikTok, but it’s hard to rank there. TikTok and Instagram are media-first, and they don’t necessarily transcribe video reliably.
If your video is you laughing in the toilet — that’s funny, it gets engagement — but it doesn’t communicate trust and safety.
So keep writing blog posts. Use AI.
A lot of brands have a thin content layer because they think humans won’t read it — but that’s okay.
Agents read it. They will see the answer there. You’re writing for them too.
Don’t worry — it’s not like a PhD thesis that no one reads. (Okay… someone will read it.)
Practical Takeaway: What You Can Do This Week
Cassie:
Okay, before you head out, here’s one practical thing you can do this week to make your website more ready for AI search.
Pick one core page on your website — your homepage, your main product page, or your services page.
Then read it like a machine would and ask yourself:
- Is it immediately clear who this is for and what problem it solves?
- Are the most important facts easy to find without scrolling, clicking, or guessing?
- If an AI engine had to summarize this page in two sentences, would it have enough information to do that accurately — or would it have to guess?
If anything feels vague, buried, or overly polished at the expense of clarity, that’s your sign.
You don’t need more content. You need clearer signals.
And if you’re not sure where those gaps are — or you want a second set of eyes on how your brand shows up inside AI answers — that’s exactly what I help with.
I offer AI search visibility audits that look at how systems like ChatGPT, Gemini, and Perplexity understand, summarize, and recommend your brand — and where that understanding breaks down.
Thanks for listening. I’ll see you in the next episode.
If we haven’t met yet….
Hi, I’m Cassie, a fractional content strategist for early-stage startups who focuses on AI search optimization.
If AI-generated answers are already influencing your buyers, your content strategy needs to account for that. I build content programs that connect strategy with execution: clear positioning, systems, and content that actually drive revenue.
If you’re ready to stop guessing and start growing, here’s how we can work together.
