# Most media publishers are solving the wrong AI problem
Source: https://simonbeauloye.com/writing/future-media/ai-restructures-publishing/
Published: 2026-04-17
Pillar: future-media
Author: Simon Beauloye (https://simonbeauloye.com)
License: CC-BY-4.0 (attribution required)
Cite as: Simon Beauloye, "Most media publishers are solving the wrong AI problem", https://simonbeauloye.com/writing/future-media/ai-restructures-publishing/
AI-use policy: https://simonbeauloye.com/ai-policy.txt
> A recent dinner in Dubai with senior publishers. Most ask how to use AI to produce more content — I'm increasingly convinced the bigger lever is rebuilding the organisation around AI, not integrating it onto separate functions.
## Takeaways
- Content production is the fastest-commoditised result of AI in publishing. The real business opportunity sits one level up, in rebuilding the organisation itself.
- When production compresses from days to minutes, every upstream and downstream workflow must change, or you end up with drafts no one can publish responsibly.
- Three mOOnshot digital examples (newsletter automation, affiliate data pipelines, and internal dashboards) show how we built platforms that would have been too costly before.
- Organisational-level AI adoption is harder because budgets and leaders are aligned with functions.
- A 5-person AI-native publisher can match a 50-person legacy operation at a fraction of the cost.
- Treating AI as an operating system rather than a content tool is the architectural shift that rewrites the economics underneath business growth.
## Signals
- Claim: Luxe Digital Privileges sends newsletters to 30,000 subscribers three times a week via a single n8n workflow with a ten-minute human review gate.
Year: 2026
Source: mOOnshot digital operating data
- Claim: mOOnshot digital's affiliate platform ingests daily data from 150+ affiliate partners via an n8n pipeline with normalised outputs feeding the CMS.
Year: 2026
Source: mOOnshot digital operating data
- Claim: mOOnshot digital moved from a fifteen-person writing team to a small team of operators managing an AI infrastructure covering development, finance, legal, market research, writing, editorial review, SEO, and GEO.
Year: 2026
Source: mOOnshot digital operating data
- Claim: mOOnshot digital operates without a dedicated internal development team; operators build their own internal tools using Claude Code.
Year: 2026
Source: mOOnshot digital operating data
- Claim: In the Reuters Institute's January 2026 survey of 280 news leaders across 51 countries, 67% reported that AI efficiencies had not saved any jobs, 16% had slightly reduced headcount, and only 9% had added new roles — consistent with publishers bolting AI onto existing org structures rather than restructuring around it.
Year: 2026
Source: Reuters Institute, Journalism, Media and Technology Trends and Predictions 2026 — https://reutersinstitute.politics.ox.ac.uk/journalism-media-and-technology-trends-and-predictions-2026
## Article
I was recently at a leadership dinner in Dubai with editors and executives from leading regional publications. Smart people, serious businesses, all wrestling with the same question most of the industry is asking right now: *how do we use AI to produce more and better content?*
It's a fair question. But the more I run [mOOnshot digital](https://moonshotdigital.com) with AI at the centre of daily operations, the more I'm convinced the biggest results won't come from using AI to produce more content. They'll come from rebuilding organisations around AI from the ground up, rather than integrating AI agents into separate functions one at a time. I run mOOnshot digital from inside that transition. We have streamlined our operations — from a fifteen-person writing team to a small team of operators managing an AI infrastructure that covers development, finance, legal, market research, writing, editorial review, SEO, and [GEO](/glossary/geo/) (the full story is in [I replaced a 15-person writing team with AI agents](/writing/ai-publishing/ai-rebuild-retrospective/)).
Revenue is still below its peak for our publishing businesses as online advertising and affiliate revenue have changed beyond recognition. But I've learned that **the real story is not about writing.** It is about everything else.
## Why producing more content with AI is the smallest prize
The fastest-commoditising part of AI in publishing is content generation. All the major LLMs can produce a decent 1,500-word article. Most publishers have access to the same tools. Whatever edge exists is now limited to whoever is using the latest model for a short period of time.
I have watched this play out inside our own business. Two years ago, writing throughput was genuinely hard. Today it is the easy part. **The [bottleneck](/glossary/bottleneck-shift/) has shifted to the human workflow around the AI:** brief quality, editorial review, visual production, fact-checking, distribution. When the production step compresses from days to minutes, everything upstream and downstream has to change, or you just end up with a large pile of drafts no one can publish responsibly. I've gone into the mechanics of this in [The bottleneck shift: why AI in publishing is now a human-time problem](/writing/ai-publishing/ai-publishing-bottleneck-shift/).
From the conversations I've had with other publishers over the past two years, most of the AI investment has gone into back office functions: summarisation, translation, transcription, SEO. These save time inside the existing org chart but they barely affect total headcount. The 2026 [Reuters Institute's survey of news leaders](https://reutersinstitute.politics.ox.ac.uk/journalism-media-and-technology-trends-and-predictions-2026) supports that impression: 67% of publishers said AI had not saved any jobs, and only 9% had added new roles. The tools are in. The org chart underneath has barely moved. That is exactly why these investments only deliver small gains compared to rebuilding the org chart itself.
This is why the question I keep coming back to has shifted. Instead of thinking about AI as a content production machine, I now focus on redefining our entire organisation with AI as part of the foundation. **What would that new organisation look like?**
In my own experience, that second question is harder to sit with, because the honest answer is usually a smaller, flatter, more operator-heavy company than most businesses operate today.
## What does organisational restructuring around AI actually look like?
Here are three concrete examples from inside mOOnshot digital. None of them are about content generation. All of them are about collapsing work that used to need teams.
### Newsletters that used to take hours
We run [Luxe Digital Privileges](https://privileges.luxe.digital), an affiliate platform with 30,000 newsletter subscribers. Historically, producing a three-times-a-week send involved someone pulling the best current deals from multiple partner networks, writing subject lines and intros, laying out the email, and scheduling it. Multiple hours of skilled time per send.
We now run the whole thing through an n8n workflow. It pulls the live offers from partner APIs, scores them, formats the email, writes platform-native subject lines, and queues the send. A human reviews before it goes out. The send is one person's ten-minute gate, not a two-person afternoon.
We built the platform itself with Claude Code in a couple of days. Frontend, backend, partner API fetching routines, and the email integration, all written iteratively with the tool. Two years ago, this would have been a scoped-out multi-month project with outside developers and a meaningful budget.
### Affiliate data pipelines that a team used to own
The same platform ingests daily data from 150+ affiliate partners: offers, pricing, stock, commission rates. Before, this was a spreadsheet, an email thread, and a junior ops hire. Now it is an n8n pipeline with normalised outputs feeding directly into the CMS.
The underlying engineering is not glamorous. It is integrations, retries, schema checks, per-source health monitoring. But it has removed an entire recurring data-ops job from the business.
Because the data flowing in is now fresher and cleaner, it has also opened the door to further affiliate optimisations on [Luxe Digital](https://luxe.digital) itself. Better real-time offer sorting on the editorial site, more responsive coverage of live promotions, tighter matching between partner campaigns and what the audience is actually reading. Work we could not have done reliably when the pipeline was a spreadsheet and an email thread.
### Internal tools built by operators, not engineers
The most underrated shift is what Claude Code and similar tools have done to internal tooling. Jobs that used to require a contracted developer for a couple of weeks — a link audit script, a custom dashboard, a one-off data cleanup — are now built by the person who needed the tool, in an afternoon.
This doesn't mean that anyone with a Claude Code subscription can now build their own internal tools easily. A technical mindset still matters, probably more than it seems at first. I don't have a computer-engineering degree, but I have been building websites for over twenty years across multiple tech stacks, and I'm comfortable in design tools as well. That background is what tells me when to push back on an AI recommendation, how to verify and validate what it produces, and what it takes to turn an afternoon's working prototype into something safe to deploy to production. Handing Claude Code to someone without that foundation will yield average results at best, and occasionally dangerous ones.
We do not have a dedicated internal dev team. We do not need one. But all our operators are tech-savvy and they understand what it takes to solve problems with AI.
That is the pattern underneath all three examples. AI does not just sit on top of the work. It absorbs the job that used to require a specialist role in between.
## Why is org-level change harder than tooling change?
The honest answer is organisational inertia, and I have a lot of sympathy for it. Restructuring around AI means fewer roles, different roles, and a reporting structure the current org was not designed to support. That is not a tooling problem. It is a change-management problem, which is always harder, particularly inside larger organisations with longer histories.
Three specific frictions tend to show up:
1. **Budgets are allocated by function.** Editorial, product, tech, ops, each with their own P&L owner. An AI infrastructure that cuts across all of them has no natural home and no natural sponsor.
2. **Talent overlaps with identity.** Asking whether a role is still needed when an AI system can do 80% of it is a much harder conversation when the role is a long-serving editor or a department head.
3. **The board often reads content volume as progress.** Producing more articles is a legible KPI. Restructuring the org so it needs fewer articles to hold the same margin is not.
None of this is about smart versus less smart organisations. It is about the speed at which different structures can adapt. Smaller, [AI-native](/glossary/ai-native-publishing/) operators have fewer of these frictions, which is the structural advantage they get to work with.
## What does a 5-person AI-native publisher look like in practice?
A useful way to see the shape of the opportunity is to compare two hypothetical publishers serving the same niche at roughly the same audience size.
| | 50-person legacy publisher | 5-person [AI-native](/glossary/ai-native-publishing/) publisher |
|---|---|---|
| **Editorial** | Editor-in-chief, section editors, staff writers, freelancers | One [editor-operator](/glossary/non-engineer-operator/) directing an AI pipeline with human review gates |
| **SEO and GEO** | Dedicated SEO team plus outside agency | Operator running an AI pipeline for keyword research, briefs, internal linking, [GEO](/glossary/geo/) checks |
| **Newsletter and audience** | Newsletter editor, designer, QA | Automated n8n pipeline with a ten-minute human review |
| **Data and reporting** | Analyst plus BI contractor | Operator with dashboards built by the team in Claude Code |
| **Ad ops and commercial** | Ad ops manager, direct sales team, programmatic specialist | Programmatic and affiliate stack managed by one commercial operator |
| **Internal tools** | Contracted dev team or long roadmap queue | Built by the operator who needs them |
| **Cost base** | Heavy fixed staff cost | Low fixed cost, variable AI spend |
| **Speed to ship** | Quarters | Weeks |
The 5-person version is not better at every single function. A dedicated SEO team will out-research one operator on a hard brief. A proper newsroom will out-report a pipeline on a breaking story. But across the whole business, the structural advantages add up: lower cost base, faster decisions, fewer dependencies, no committee.
For most mid-market publishing niches, that combination wins. Not because the small team produces more, but because the small team survives a market where revenue per article has collapsed and the old cost structure can no longer be financed.
## The question I keep coming back to
Start from zero. If you were launching a media business today in your niche, with AI tools as part of the foundation rather than an add-on, how big would the team be? What roles would exist? What would you never build in-house?
When I have asked publishers this question over the past year, the answer is consistently a team a fraction of the size of the one they have. The gap between that answer and the current org chart is, in my view, **where the real AI opportunity sits**. More articles is the surface. The organisation underneath is where the economics get rewritten.
We are early in this shift, and I am as much in the middle of it as anyone. mOOnshot digital is not a finished case study. It is a live one. Revenue is still rebuilding. The operational model is not done. But the direction has become clearer to me with every quarter that passes, and with every piece I write on [the future of media](/writing/future-media/). Treating AI as a content tool puts the business in a content-volume race that is already hard to win. Treating [AI as an operating system](/glossary/ai-as-operating-system/), the architectural stance that AI should be the foundation a business is built on rather than a tool bolted onto existing workflows, rebuilds the business underneath the content. That, from my own experience, is where the durable advantage seems to sit.
This essay is one principle inside the wider [Zero-Base Operations](/glossary/zero-base-operations/) framework — my shorthand for justifying every process, tool, and hire from zero with AI as part of the foundation rather than an add-on. The full method lives in [its own piece](/writing/bootstrapping/zero-base-operations/).
Dinner in Dubai ended with a few people asking what tools we use. The tools are the easy part to share. The harder and more interesting question, the one I am still working through myself, is what you would stop doing, stop hiring for, and stop buying if you were starting the organisation today. Once that question is answered honestly, the tools tend to follow.
I'm publishing essays like this one to start conversations. If you found this article interesting, or if you have questions or suggestions, I'd love to hear from you. You can connect with me on [LinkedIn](https://www.linkedin.com/in/simonbeauloye/) or [X](https://x.com/simonbeauloye), or write to me directly at [hello@simonbeauloye.com](mailto:hello@simonbeauloye.com).
## See also
- [Site index](https://simonbeauloye.com/llms.txt)
- [Full corpus](https://simonbeauloye.com/llms-full.txt)
- [Pillar index (future-media)](https://simonbeauloye.com/llms/future-media/llms.txt)
- [Pillar hub (future-media)](https://simonbeauloye.com/writing/future-media/)
- [AI-use policy](https://simonbeauloye.com/ai-policy.txt)
### Related essays
- [We rebuilt our publishing operation around AI agents. Here's what actually happened.](https://simonbeauloye.com/writing/ai-publishing/ai-rebuild-retrospective/)
- [The bottleneck shift: why AI in publishing is now a human-time problem](https://simonbeauloye.com/writing/ai-publishing/ai-publishing-bottleneck-shift/)
- [Zero-Base Operations: How to build a bootstrapped business with AI as the operating system](https://simonbeauloye.com/writing/bootstrapping/zero-base-operations/)
### Glossary terms referenced
- [Generative Engine Optimisation (GEO)](https://simonbeauloye.com/glossary/geo/) — The practice of structuring a website so AI answer engines (ChatGPT, Claude, Perplexity, Google AI Overviews) can ingest, ground, and cite its content reliably.
- [Zero-Base Operations](https://simonbeauloye.com/glossary/zero-base-operations/) — Zero-Base Operations is Simon Beauloye's framework for building businesses by justifying every process, tool, and hire from zero, with AI as the foundation rather than an add-on.
- [AI as Operating System](https://simonbeauloye.com/glossary/ai-as-operating-system/) — AI as Operating System is the architectural stance that AI should be the foundation a business is built on, not a tool bolted onto existing workflows.
- [The bottleneck shift](https://simonbeauloye.com/glossary/bottleneck-shift/) — The bottleneck shift is Simon Beauloye's framing for what happens after AI compresses one stage of a workflow: the constraint doesn't disappear, it moves.