I know AI is already doing a lot of SEO work, and honestly probably much more than most of us expected a few years ago. But I'm curious about where the limits are. Are there any SEO processes where you still feel that AI can't really do the job properly without human involvement?

I'm especially interested in things that require real experience, judgment, relationships, or understanding the business behind a website.

What have you personally found AI is still not good at?

What SEO tasks still can't really be done well with AI?

I've been thinking about this lately. AI can do a lot of SEO work now — keyword research, content, audits, competitor research, internal linking, etc. But I'm curious about the things where you still need an actual person with experience. For those of you who work in SEO, what tasks do you think AI still struggles with or can't replace properly?

Would love to hear some real examples from your own experience, not just the usual "AI can't replace humans" answer.

Mine: a free calculator tool built in a week that now brings 400 visits a month and converts at 6 percent, roughly triple our Google Ads landing pages at a fraction of the cost. One win like that funds a lot of failed tests. What worked for you in 2026, and what did it cost to find out? Trying to build a better backlog than 'post more on LinkedIn'.

I can see LLM referrals when the source header survives, but a big chunk lands in direct, especially from in-app browsers. I've tried regex on session source and a lookup table of known referrers, still feels leaky. Anyone built a setup they trust? Rough share too: We see 2 to 4 percent of sessions from AI sources on two b2b sites, wondering what others measure.

Volume is easy now and sameness is the new problem. The stuff that still earns links and gets quoted has original numbers, a named author with a stake, or a take you can disagree with. That gets hard at 8 a month. What's your process for keeping opinion and first-hand detail in every piece when the calendar is relentless? Interviews? Data pulls? Cutting volume?

Mine: a free calculator tool built in a week that now brings 400 visits a month and converts at 6 percent, roughly triple our Google Ads landing pages at a fraction of the cost. One win like that funds a lot of failed tests. What worked for you in 2026, and what did it cost to find out? Trying to build a better backlog than 'post more on LinkedIn'.

Added it to two sites in June. Server logs show barely any bot pulls and citation tracking moved zero. I keep it because it costs nothing, but the loud advice around it feels way ahead of the evidence. Someone tell me I'm wrong with numbers. Did anyone measure a real change they can tie to the file?

Hi everyone,

We're NuroSparX, an AI-first digital growth agency working with clients across multiple industries, and we're expanding our team.

We're currently hiring for the following fully remote (India) roles:

• SEO Executive (3+ years)
• SEO Lead (5+ years)
• Performance Marketing Executive (3+ years)
• Performance Marketing Lead (5+ years)

What we're looking for

• Strong fundamentals in your domain, should be able to explain WHY behind concepts
• Analytical and data-driven thinking, easy grasp on reporting, dashboards with AI
• Comfortable using AI tools like ChatGPT, Claude, Gemini, etc.
• Excellent communication and ownership
• Ability to work independently in a remote environment

We're looking for people who enjoy solving business problems, with the right guidance

If that sounds like you, we'd love to hear from you. Apply from the link on profile