Every marketing team can now produce ten times the content at a fraction of the cost. So can every competitor, which is exactly why content volume stopped being a strategy.
When a capability becomes universally available, it stops being an advantage and becomes a cost of entry. That is the story of AI in marketing this year.
What Got Commoditised
- Blog volume. Publishing frequency is no longer evidence of investment or expertise.
- Ad copy variations. Testing fifty headlines costs nothing, so everyone does.
- Basic personalisation. Name in the subject line has been table stakes for years and is now genuinely trivial.
- Competitor monitoring. Everyone has the same visibility into everyone else.
- Generic thought leadership. The most saturated format in B2B, and the easiest to generate badly.
What Became More Valuable
The inverse of that list. Original research, because it cannot be generated. Genuine customer relationships, because they cannot be automated into existence. Distribution you own, such as a list, a community or a podcast audience, because it does not depend on an algorithm. And point of view, because a model averages opinions and a strong position stands out precisely by not doing that.
The pattern is consistent. Anything requiring access to something real, whether people, data or judgement, went up in value. Anything requiring only production capacity went down.
Where Budget Is Moving
| Away from | Toward |
|---|---|
| Volume content production | Original research and data |
| Broad top-of-funnel SEO | Citation in AI answers |
| Generic email nurture | Owned community and events |
| Agency content retainers | In-house specialists with domain depth |
| Vanity traffic reporting | Pipeline attribution that survives scrutiny |
The Measurement Problem Nobody Solved
Attribution was already imperfect and AI answers made it worse. A prospect can research your category thoroughly, form an opinion about you, and arrive as direct traffic with no trace of the journey.
Reported data suggests a majority of B2B websites lost meaningful organic traffic across 2024 and 2025, with average declines around a third. If your reporting treats sessions as the goal, that looks like failure even when pipeline is flat or growing.
Move your primary metrics to branded search volume, direct traffic, and self-reported attribution on forms. The last one is unfashionable and remarkably effective: simply ask people how they heard about you.
What Still Works
Being genuinely useful to a specific audience, consistently, for longer than your competitors have patience for. That was true before AI and the economics now favour it more, because the alternative strategy of outproducing everyone became available to everyone simultaneously.
Conclusion
Stop competing on content volume, because that race has no finish line and no prize. Invest in original research, owned distribution and a clear point of view, all of which resist generation. Shift measurement to branded search, direct traffic and self-reported attribution before falling session counts get mistaken for falling performance. Use AI to remove production drudgery, not to replace the parts that were ever the differentiator.
Frequently Asked Questions
Should we publish AI-assisted content?
Yes, with genuine editorial judgement and something original in it. Assisted production is fine. Unreviewed generation at volume is how a domain acquires a reputation problem.
How much should we spend on original research?
More than feels comfortable. One solid annual data study typically earns more citations, links and sales conversations than a year of ordinary blog posts.
Is paid search still worth it?
Yes, and watch the mix carefully. As organic clicks fall, competition for paid positions rises, and your cost per click will tell you that story before any report does.