We’re all making an attempt to determine the way forward for content material advertising.
Because of generative AI, each a part of our self-discipline is in flux: the ways and techniques that yield nice outcomes, the instruments and processes we use to realize them, even the unit economics that justify it.
One suggestion put ahead by Josh Spilker at AirOps and Loreal Lynch at Jasper is that firms ought to rent a brand new function, the content material engineer. Right here’s my understanding of the function:
The content material engineer
A content material engineer is an AI-native content material function with a give attention to techniques pondering, utilizing AI to scale content material output and improve high quality.
As a substitute of making content material, they construct automated techniques to create content material at scale. It’s a hybrid function that includes parts of content material technique, workflow automation, search engine optimisation, and immediate engineering.
The first objectives of the content material engineer embody:
- Bridging the hole between artistic content material advertising and technical AI implementation
- Scaling written content material manufacturing and automating content material repurposing, distribution, and personalization
- Safeguarding model tips, authorized and regulatory compliance, and editorial requirements
- Iterating to enhance content material high quality over time
Sidenote.
I believe this can be a honest illustration, however please learn the unique articles and determine for your self.
I’m extraordinarily bullish on generative AI (you possibly can examine my AI content material course of right here). It’s reworking our trade in beforehand unthinkable methods, and we now have to adapt.
However as I take into consideration hiring for this new period, the content material engineer is sort of diametrically against the abilities and talents I consider we’d like. I believe the content material engineer is the fallacious ability set for the way forward for advertising.
Right here’s why.
The content material engineer’s remit is automation, taking content material methods that work nicely and scaling them to the Nth diploma utilizing AI. However the golden period of scaled content material is already behind us.
Ahrefs is a good case research in scaled content material. We’ve got 2,161 articles revealed throughout half a dozen languages. We’ve got a number of programmatic content material methods, peaking at round 34,000 automatically-generated pages and hundreds of thousands of pageviews.
However on the time of writing, the Ahrefs weblog receives a fraction of the clicks from Google Search that it did at its peak. And only a few weeks in the past, we selected to maneuver our most profitable programmatic content material technique away from our predominant web site and to a separate area.
Scaled content material creation is essentially the most accessible it has ever been. It’s additionally the least efficient it has ever been. Indexing new pages is troublesome. Google has repeatedly penalized and demoted web sites for “scaled content material abuse”. Many profitable programmatic methods are just a few clicks away from being copied wholesale, in their very own model of the search engine optimisation heist. And because of AI Overviews and AI Mode, the unit economics of large-scale content material manufacturing simply don’t work like they used to.
This isn’t a coincidence. The flexibility to automate one thing and the diminishing of its returns are essentially linked. Any a part of content material advertising that may be totally automated by AI is destined to develop into nearly nugatory.
For those who can publish 1,000 programmatic touchdown pages per week, your rivals are incentivised to do the identical. Because of the democratisation of generative AI, it’s low cost and straightforward to take action. Quickly, each competitor is compelled into the identical plan of action, and the tactic is rendered ineffective.
(What number of automated pitches do you get from AI SDRs every week? And the way rapidly did their novelty put on off?)
The helpful half-life of “scaling” content material with generative AI is at the moment measured in months, and can quickly be measured in weeks or days. Hiring a content material engineer to supervise content material automation could make sense within the brief time period, however it dangers diverting restricted vitality and assets to dying methods, moderately than new content material methods with the potential to supply outsized returns.
There are numerous content material methods nonetheless yielding progress, however they require effort and experimentation, and never automation and scale. I really feel headwinds for programmatic and top-of-funnel content material, and tailwinds for authentic analysis and information content material, platform-native thought management content material, short-form and long-form video content material.
Generative AI has essential roles to play in every of those areas, however the content material engineer just isn’t the skillset for the job.
In a world drowning in AI slop, publishing high quality content material issues greater than ever. However there’s a huge distinction between gradual, incremental enhancements to consistency and output, and really creating content material that anybody cares about.
AI has undeniably raised the ground of content material advertising. For just a few {dollars} and some minutes of time, any firm can generate satisfactory, practical written content material. The baseline high quality of content material advertising is the very best it has ever been.
However on the identical time, AI has utterly failed to boost the ceiling of content material advertising. We aren’t swimming in new and thrilling types of content material advertising on the again of AI innovation. Most firms adopting generative AI in content material are cost-motivated: effectivity, productiveness, doing extra with much less, these are the recurring themes that characterise AI adoption in content material advertising.
Month-to-month content material advertising spend stays the identical no matter AI use. AI is getting used to stretch the identical finances farther, not unlock new and thrilling types of buyer acquisition.
This isn’t a restrict of AI. It’s a incredible artistic sparring associate, a instrument for studying new abilities and difficult preconceived concepts. High quality is nicely inside generative AI’s remit, however as an alternative, most firms select to focus it on automation and scale, like paying Picasso to color your fence.
Hiring content material engineers would perpetuate this downside. The content material engineer is essentially a top quality assurance function, designed to cut back errors and errors, to extend uniformity and model consistency. It isn’t a job involved with precise high quality: creating outstanding issues that ship outsized outcomes.
We don’t want extra consistency in content material advertising: we’d like weirdness and irregularity. We don’t want model voice as a lot as we’d like actual, distinctive, particular person voices. We don’t want engineers to productize content material like so many widgets on an meeting line; we have to use AI to amplify the most effective traits of craftspeople and obsessives, and embrace a definition of high quality that’s messy and surprising.
Content material engineers act as a bridge between artistic content material advertising and technical AI implementation. However I believe this bridge is already pointless.
Generative AI fashions are barrelling headfirst in the direction of large client adoption, with the intention of democratizing entry to writing, coding, and design. Just about day-after-day, mannequin suppliers discover new methods to make their AI simpler to make use of: creating new and opinionated workflows, releasing new guardrails, and studying from the successes of their present customers so as to nudge future customers in the direction of higher outcomes.
I‘ve spent this week utilizing Google’s newest picture mannequin, Nano Banana, to generate photos for my speak at Ahrefs Evolve. The mannequin creates the very best high quality photos I’ve ever generated, and it does it with a single line immediate.
Zero immediate engineering ability required.
We see the identical trajectory in different generative fashions. Deep analysis has compressed dozens of back-and-forth immediate exchanges into two fast messages. Most AI instruments now robotically choose the absolute best mannequin to execute your immediate. MCP servers make accessing APIs a matter of writing plain textual content requests.
The barrier to utilizing AI is changing into decrease and decrease. With every passing day, nice outputs develop into simpler to realize, requiring much less finessing, much less tweaking, and fewer technical abilities. Generative AI is simple to be taught; it must be, or your entire trade crumbles. Nice advertising, and nice writing, are far tougher abilities to grasp.
Very similar to immediate engineers, the content material engineer looks like a stopgap function, a short-term repair for issues which can be being solved by the instruments themselves. (And what number of firms have you learnt with full-time immediate engineers?).
AI can already write novels and construct software program—within the grand scheme of issues, getting your product’s technical particulars appropriate and utilizing en dashes in applicable locations are trivial issues to resolve, and never value hiring towards.
In some methods, hiring a content material engineer is to understate the significance of generative AI. AI fluency is one thing to be anticipated of us all, in the identical method that web fluency is. It isn’t one thing to be constrained to a devoted function.
Ultimate ideas
To reiterate: we’re all making an attempt to determine the way forward for content material advertising. Content material advertising has to evolve to accommodate generative AI, and it’s much better to posit an answer than to hold on as if nothing has modified.
However from my perspective, the function of the content material engineer carries the insidious threat of hiring for abilities we gained’t want for lengthy, on the expense of these we actually will.
On the excessive, this resolution to index hiring on AI abilities dangers making a staff of n8n jockeys who spend their time scaling unremarkable content material methods that don’t assist the enterprise. Worse, it will increase the chance of lacking out on the brand new and rising advertising alternatives that can outline the success tales of tomorrow.
At Ahrefs, I would like my content material staff to index closely on advertising and writing abilities. I belief that nice entrepreneurs can be taught AI, however I don’t belief that AI consultants can develop into nice entrepreneurs. Give me a alternative between hiring a fantastic author with no understanding of generative AI, and a middling author with fluency in n8n and MCP servers, and I’ll take the good author each day.
