Does AI content rank? Yes. But that is no longer the point

Last week, MarTech covered Semrush’s new study on whether AI content ranks well in search, and the headline was about as surprising as rain in Manchester: yes, it can. Google is not automatically punishing AI-written content, and content quality still determines outcomes. Useful, clear, relevant pages can perform whether a human drafted every line or not.

That should calm one debate and intensify another.

Because if AI content can rank, then ‘can it get on the page?’ is no longer the interesting question. The more uncomfortable one is what happens when everyone can produce search-competent material at scale, with decent grammar, clean structure and just enough surface-level usefulness to pass as good.

The answer is not hard to see. More output. Less distinction. More polish. Less real conviction. Search fills up with content that reads perfectly well and leaves almost no mark. It ranks, it nods politely at intent, then it vanishes into the wallpaper.

SEMrush analysed 42,000 blog posts and found that AI content is not inherently blocked from ranking. MarTech’s summary of the study landed on the right conclusion: search engines are evaluating AI-assisted pages the same way they evaluate any other page, by usefulness, relevance and clarity.

Yes, AI content can rank on Google if it is useful, relevant and clear. The method of production is not the deciding factor. The stronger question is whether the content adds enough original value to compete once many other brands can now publish similarly competent material at speed.

That distinction matters. Ranking has always been a means, not an outcome. Yet AI has made it temptingly easy to confuse technical eligibility with commercial effectiveness. A page that lands on page one but says what fifty other pages already say has achieved something, certainly. It just may not have achieved anything you can take to a revenue meeting with a straight face.

Google’s own guidance has been consistent on this point. Generative AI can help with research and structure, but content created primarily to manipulate rankings or mass-produce low-value pages risks falling into scaled content abuse. Google’s ranking systems prioritise helpful, reliable, people-first content, not content that exists merely because a workflow made it cheap to generate.

No, Google does not automatically penalise content just because AI helped create it. What it does warn against is scaled content abuse, where content is mass-produced mainly to manipulate rankings rather than help users. Quality, originality and value still do the heavy lifting.

That is the policy answer. It is also the easy answer.

The harder truth is that search quality and market quality are not always the same thing. A page can be good enough for Google’s systems and still be strategically forgettable. It can satisfy the machine’s threshold for usefulness while doing very little to make a buyer trust you, remember you or choose you.

This is where the current AI content conversation remains oddly timid. Much of the trade coverage still circles the compliance question, as though the main issue were whether AI content is allowed into the building. It is. The more pressing issue is what it looks like once everybody gets inside.

AI is very good at improving grammar, smoothing structure and producing broadly acceptable answers. It is much less reliable at generating sharp judgement, first-hand experience or the sort of commercial tension that makes a reader stop and think, ‘Fine, these people actually have a point.’ Left alone, it tends to average things out.

Sensible. Balanced. Safe. Magnolia messaging, to apply a term coined by The Rubicon Agency. Safe enough to offend no one, and persuasive enough to move almost no one.

AI content often fails after ranking because visibility is not the same as differentiation. Many AI-assisted pages are readable and technically relevant, but too generic to persuade, be remembered or shape preference. They meet the brief for search while missing the brief for actual market impact.

That is not a small problem. In B2B technology especially, where buyers face complicated choices and long sales cycles, content must do more than answer the query in front of it. It needs to signal judgement. It needs to show that someone behind the brand understands the category, the stakes and the trade-offs. Otherwise, you are just another competent voice in a queue of competent voices.

The Rubicon Agency is already on the record arguing against vague, vacuous content and in favour of more distinctive, proposition-led thinking. We’re not inventing a new belief here – it’s extending an existing one into the AI era.

There is a mild irony here. AI lowers the cost of producing decent content, which means decency itself becomes less valuable. The commodity becomes the baseline. What gets expensive again is not production, but perspective.

That does not mean every blog post needs to be a manifesto. Some queries deserve straightforward answers. Some pages should simply help. But even practical content benefits from specifics, original framing and evidence that a human mind has actually interrogated the material rather than merely rearranged it. Real examples. Clear trade-offs. A sentence or two that sounds like it could only have come from this company, not from any company that subscribed to the same model last Tuesday.

The Rubicon Agency already has a useful framing device for this in The Content Spectrum, which positions content according to buyer need, product maturity and sales stage rather than pretending every asset has the same job. That thinking becomes even more relevant now. AI may be good at generating a competent middle. It is much less dependable at deciding when a piece should provoke, reassure, reframe or sell.

Brands should use AI for acceleration, not authorship by default. Let it help with research, structure and draft momentum, then add what models usually flatten out: clear judgement, first-hand insight, sharper examples, stronger voice and a point of view that reflects the brand rather than the average of the internet.

The commercial point is simple. Search performance still matters. So does efficiency. But if AI makes it easier for everyone to publish acceptable content, acceptable becomes a weak ambition. The brands that win will not be the ones producing the most polished neutrality. They will be the ones that decide what they actually want to say, then say it clearly enough that a buyer remembers who said it.

AI content can rank. That debate is settling. Good.

Now for the more useful one.

If production gets faster, where does the saved effort go? Into more volume, more templates and more faintly competent pages that all smell the same? Or into better judgement, tougher editing and stronger ideas that are actually worth surfacing in search? Google’s guidance gives you the minimum standard. The market will demand more than that.

That is where the opportunity sits. Use AI to remove drudgery. Then spend the reclaimed time on the bits that still resist automation: deciding what matters, what is true, what is commercially at stake and what your brand is prepared to stand for in public. If that sounds less scalable than pressing ‘generate’, that is because it is. It is also where the advantage still lives.