Artificial intelligence is redefining the entire user journey and forcing teams to rethink the role of SEO in growth decisions, between post-click insights and context-driven content
SEO is changing in structure, not just in form. Zero-click searches, AI Overviews, AI Mode and generative answers are rewriting the way people look for information, weigh alternatives and make decisions.
Value is no longer concentrated solely in rankings or clicks, but in the ability to be relevant before the user even reaches your website.
We are living through a phase shift: more complex, harder to read with traditional metrics, yet decisive for those who know how to watch the right signals, question old certainties and rethink the role of SEO within product, marketing and growth decisions.
In SEO Confidential we meet the people who work every day along these precarious boundaries, where data, product and strategy intertwine. In this episode our guest is Bengü Sarıca Dinçer, SaaS SEO Manager at Designmodo and freelance consultant for international SaaS companies.
Bengü has been working for over twelve years across startups and established brands, and brings a clear-eyed, strongly data-driven vision of modern SEO, always keeping a practical focus on the business context.
In the interview we talk about what really happens after the click, about e-mail marketing, about artificial intelligence used as an ally, all the way to the big themes of visibility in generative systems and brand reputation in the era of LLM answers.
A conversation designed for those who are not looking for shortcuts, but want to understand how to read the change, govern it and turn it into more informed strategic choices, before it becomes a problem to chase.
“AI is confident, but it is not accountable. And that makes it dangerous in the hands of irresponsible people”
A well-crafted newsletter is one of the most effective tools for building trust, staying in touch with your audience and delivering real value to the brand over time. Yet many companies send newsletters and e-mail campaigns without truly understanding what happens after the click: why is it so important today to use tools like Google Analytics to understand whether e-mails are actually working?
Because e-mail platforms measure e-mail performance, not business performance. I don’t care whether 40% opened the e-mail or 39% didn’t click the button inside it. The truth lies in what people do once they land on my website, not in what my e-mail tool tells me.
GA4 reveals what everyone loves to ignore: what people actually do when they reach the site. Do they read? Do they explore? Do they convert? Do they leave after one second? Understanding this is a huge opportunity for any kind of business. Once you understand what your users do after seeing your e-mail campaign, you can optimize your processes and eventually convert them.
However, most teams skip this part. And I think it’s because they don’t know how to leverage GA4, or because they assume e-mail metrics are enough. But they aren’t. Post-click data is the reality check. That’s why tracking e-mail campaigns properly matters.
Which metrics really matter when evaluating an e-mail campaign in a serious, business-relevant way?
I believe opens and clicks are only a small part of the story. They are not outcomes. For me, everything meaningful happens after the user visits the website, as I said before. That’s when intent becomes visible. So I look at:
Engagement metrics: session duration, pages viewed;
Micro-conversions: sign-ups, interactions with tools;
Macro-conversions: purchases, demo bookings;
Retention signals: did this e-mail bring the visitor back later?;
Revenue attribution: did this campaign contribute in a measurable way?
At the end of the day, I try to understand whether the e-mail supports the business goals. If the answer is no, the campaign didn’t “work”, regardless of what the e-mail platform’s metrics say.
When it comes to tracking, the most common mistakes often come from the famous UTMs (Urchin Tracking Module, parameters added to links to track precisely where traffic comes from and which campaign it belongs to), used in a confused or inconsistent way. How badly can a simple inaccuracy in link building distort your analysis? Is this point still underestimated?
A single UTM mistake can completely misrepresent your traffic sources. And yes, it is still highly underestimated. Even small inconsistencies like “Email_Campaign_X” versus “e-mail-campaign-x” split your data, making attribution unreliable and leading to decisions based on a distorted version of reality.
Most teams still think UTMs are “just labels”, but wrong UTMs can make a successful campaign look like a failure, or vice versa.
Bengü, we know that looking at what happens after the click completely changes how performance is read. Let me ask you: what is the most uncomfortable discovery that emerges when you analyze real user behavior on the site instead of stopping at open or click data?
I think I can say it is often shocking to see that clicks do not equal engagement or interest. Someone can accidentally click a button in the e-mail, land on the site and leave after two seconds. Or they can scroll the page a little, get confused and abandon it before taking any action.
Many campaigns that look like winners from the e-mail tool’s point of view turn out to generate minimal value when you check on-site behavior. There is often a big gap between what the e-mail suggests will happen and what users actually do. And that discrepancy is uncomfortable, because it forces teams to face problems with messaging, landing pages or the value proposition, not just e-mail design.
Many people talk about using AI in SEO, but without understanding what it really means. On this note, I’d love to ask you: in which practical activities do you find that artificial intelligence truly saves time and effort?
For me, AI is like a co-pilot for thinking, or a research tool for exploring new knowledge. I generally use AI to speed up experimentation and validate ideas before acting. It saves me a lot of time in tasks such as:
- Summarizing large data sets;
- Spotting patterns in log files;
- Testing hypotheses;
- Running quick comparisons or competitive scans.
In essence, it helps me find an answer faster, but the interpretation still remains mine.
More and more content is losing rankings for no apparent reason: how can AI help understand what has really changed in search habits?
AI can be useful here for analyzing user intent or the content velocity of competitors. User behavior is changing these days, as are the interfaces people use, and on top of that, traditional SERP features have changed a lot, and sometimes analyzing everything manually takes an enormous amount of time.
AI can help us understand subtle shifts in intent, compare and identify winning content formats by query, and so on. However, we need to be aware that AI does not magically hand us the answer, but it can drastically cut research time.
AI can help, but it cannot replace human expertise: what risks do you see when AI is used without expert oversight, especially in activities that can affect results and revenue?
AI is confident, but it is not accountable. And that makes it dangerous in the hands of irresponsible people.
If you trust outputs that shouldn’t be trusted, optimize based on hallucinations, create content with the wrong intent, make decisions disconnected from context without specific expertise, then you risk spreading misinformation, inconsistency and even damaging the brand.
So yes, AI is powerful for scaling your work, but it needs us as its north star. Without someone who understands the big picture, AI becomes a risky tool that knows how to be very convincing.
In the SaaS world (Software as a Service: software accessed via the internet, with no local installation, through a subscription) AI is becoming an integral part of products, but often without a clear strategy: what is the most common mistake you see in companies that integrate AI “just because they have to” instead of building features that are genuinely useful for users?
The most common mistake is building AI features just to be able to say “we have AI”, without considering what users actually need. And you know what? Users notice it immediately.
But at that point, you can’t talk about strategy if software companies are just trying to chase trends and shipping features nobody asked for.
Instead, AI can be valuable if the idea is built around a clear value such as “What specific problem does this solve for our user?”
If that question goes unanswered, the feature usually ends up being useless. Strategic AI integration should focus on solving user problems faster and better than before.
AI platforms promise automation and speed, but in SaaS quality still depends on the ability to interpret data well: which processes should remain under human control to avoid wrong decisions that can compromise growth, retention or revenue?
I can only speak as a user and from a product SEO perspective. So, in my opinion, everything that involves judgment, interpretation or long-term consequences should remain under human control.
And that includes pricing and revenue strategy, retention and churn analysis, customer success decisions, the interpretation of ambiguous data, product prioritization and understanding why something is happening.
Of course, AI can provide support, but it cannot understand nuance, timing or emotional signals. These responsibilities still belong to humans.
The tools that measure visibility in LLM-generated answers promise to “photograph” a context that is actually opaque and unstable. How much can brands really trust this data, given the unpredictable nature of generative models?
Not completely. LLM outputs change constantly, and results are generated mainly based on prompt variations, model versions, internal updates, randomness, context windows, and so on.
So these tools can help you spot patterns or opportunities, but they are not “analytics tools”. You can’t use them as KPIs, and you certainly can’t rely on them for forecasting.
Today Search Console offers no specific metric for AI Overviews or AI Mode, leaving a huge blind spot for those working on organic visibility. What impact does this gap have on brands’ ability to assess real traffic and make informed investment decisions?
Yes, this blind spot creates a huge gap in data quality. AI Overviews, AI Mode and LLM answers are already influencing user behavior, but Google Search Console provides no specific or dedicated data on these interactions, making it impossible to measure their real impact on impressions, clicks and traffic.
This means we cannot attribute traffic correctly, quantify shifts in intent or confidently forecast organic performance. We are analyzing a channel whose main new component is invisible. That is a serious problem for strategic decision-making.
When generative systems report inaccurate data or attribute information that was never published, who takes responsibility for the consequences? Will brands have to prepare to manage reputational damage caused by AI-created content that talks about them with no oversight? How can they defend themselves?
We cannot control AI. So the area of responsibility is blurred, but the consequences hit the brand first. Therefore, the only thing brands can do is measure how quickly and effectively they can react when AI systems misrepresent or misinterpret them, without being able to directly prevent or monitor these errors.
I think they should treat generative results as a new layer of reputation management. To do that, they should monitor AI-generated mentions to catch inaccuracies early and create a process for requesting corrections where possible, as well as developing authoritative content to build trustworthiness.
Below you’ll find the video where I comment on the interview with Bengü — let me know what you think in the comments:
The new SEO doesn’t reward those who run, but those who understand
Today the winner is not the one who produces the most output, but the one who interprets the context best.
The click is no longer a certainty, the data is no longer complete, AI accelerates everything but takes no responsibility. And this is exactly where the human role becomes central again, made of judgment, method and the ability to read what’s happening.
From my point of view, this is the most interesting direction SEO could have taken.
More uncomfortable, yes. More demanding, certainly.
But also more mature.
A discipline that stops chasing easy metrics and goes back to engaging with product, marketing and real decisions, accepting opacity as part of the game instead of denying it.
The disintermediation imposed by artificial intelligence makes visibility an essential condition.
A brand must exist as a recognizable, authoritative and cited entity, capable of generating trust within the information context. Being present doesn’t mean ranking, but becoming a relevant source that AIs recognize, use and respect.
If you work in SEO, marketing or SaaS, this phase is not something to endure. It must be understood, observed and governed.
This is exactly what we want to do with SEO Confidential, week after week, giving a voice to those facing these changes in the field.
Thanks to Bengü for her valuable insights, and thank you for following this episode too.
See you next week with a new must-read guest.
