The real lesson in Meta’s job cuts

Meta’s job cuts are a neat little warning label for the rest of the internet. AI is being used to make content teams smaller before it makes them better, and for most ecommerce brands, that is the wrong order. Leaders see faster output and a lower payroll, then they sprint straight to headcount cuts like they’ve found a shortcut through the maze. Quality comes later, if it shows up at all. The problem is not the technology. The problem is the habit of treating AI like a way to need fewer people instead of a way to do better work.

McKinsey estimates generative AI could add $2.6 trillion to $4.4 trillion annually across industries, but most of that value comes from process redesign, not headcount cuts. That distinction matters. Process redesign means changing how work gets done, who reviews it, and where human judgment sits in the workflow. Cutting people first does the opposite. It removes the people who know the brand, know the customer, and know when a page sounds polished but empty, which is a very specific kind of failure and a very common one.

Small and mid-size ecommerce teams feel this most. One person is often writing SEO pages in the morning, fixing email copy after lunch, updating landing pages before dinner, and rewriting product content whenever a new collection drops. AI can remove the busywork inside that load. It can draft faster, organize notes, and turn one brief into five variations. It cannot decide which angle fits the brand, which claims need proof, or which line should be cut because it sounds like every other store on the internet. That last one is a killer. Not dramatic, just quietly lethal.

This is the part leaders keep missing. AI is useful for output. It is useless for taste without a human editor. If you remove the editor, you do not get efficiency, you get faster average work. The same mistake shows up everywhere people confuse a quick answer with a good answer, whether they are trying to figure out how to screenshot on Mac or how to boil eggs. Speed is easy. Judgment is the work.

Why companies cut content teams first

Content work looks easy to automate from the outside because the output is visible and the process is hidden. Anyone can see a blog post, an email, a product page, or an ad. Fewer people can see the research, the editing passes, the brand checks, the SEO decisions, and the fact checking that sit behind it. That makes content look like typing. It is not typing. It is judgment, revision, and a long chain of small decisions that keep the work from sounding thin, wrong, or weirdly confident about things it does not know.

That visibility problem leads leaders to confuse volume with efficiency. If one person can produce more pages with AI, the easy conclusion is that the team can shrink. Fewer writers mean lower cost, so the spreadsheet looks cleaner. The downside is hidden until later, when the site fills up with generic copy, duplicate angles, weak product descriptions, and email campaigns that all sound like they came from the same template. The savings are immediate. The damage shows up in conversion, trust, and search performance, which is a much less charming surprise.

Ecommerce teams are under constant pressure to publish more with the same budget. More category pages. More emails. More ads. More SEO content. More landing pages for promotions, launches, bundles, seasonal pushes, and retention flows. The work never stops because the store never stops changing. A lean team can use AI to keep up, but management often hears that as a reason to ask for even more output, not a reason to protect the people who keep the work sharp. The machine gets faster, then the calendar gets greedier. Very on brand for modern business.

Content is often the first place management tries to save money because it is seen as support work, not revenue work. That view is wrong, and it costs brands. Content shapes discovery, click-through, conversion, repeat purchase, and brand memory. A weak product page can lose a sale. A sloppy email can train customers to ignore future sends. A thin SEO page can waste months of traffic. The 2023 CMI survey found 54% of B2B marketers expected budget cuts or flat budgets, which explains the pressure, but it does not make the cuts smart. It just makes them common.

Why AI makes teams smaller before it makes them better

The pattern is predictable. AI drafts faster, leaders see the output, then they reduce staffing, and the team loses the people who knew what good looked like. That is how AI makes teams smaller before it makes them better. The speed gain is real, but it lands in the first draft, the outline, the repurposed version, and the rough rewrite. It does not land in strategy. It does not land in quality control. It does not land in the hard part, which is deciding what should be said and what should be left out.

The hidden cost of shrinking too early is easy to miss because the work still gets done. There are fewer editors, so weak lines survive. There is less fact checking, so errors slip into product pages and articles. There is less time for brand review, so the copy drifts toward generic language. There is less room to challenge bad ideas, so the team publishes more, but the work starts to blur together. That is how you end up with a site full of content that reads like it was written by someone who learned how to tie a tie by watching a five-second clip and never checked the mirror.

AI only improves output when humans stay in the loop long enough to train prompts, edit hard, and set standards. The Stanford and MIT study on call centers found generative AI improved productivity by 14%, with the biggest gains for less experienced workers. That is the point. AI helps when people are still there to guide it. It helps junior staff move faster, it helps experienced staff clear repetitive work, and it helps teams produce more without lowering the bar. It does not replace the people who know when a headline is vague, when a claim needs proof, or when a page sounds like it was written to satisfy a machine instead of a shopper.

This is why cutting too early is such a bad move. A smaller team can absolutely use AI well, but only if the team keeps enough experienced people to shape the output. Without that, the brand gets more content and less control. The work gets faster and flatter at the same time. That is not how you make a killing with content. It is how you fill the calendar and empty the standard.

What content teams actually do that AI cannot replace

AI can draft text fast. It cannot decide what deserves to exist. That job still belongs to a human who knows which topics matter this month, which angles fit the brand, which search intent is worth chasing, and which ideas should stay in the notebook. A good team does more than answer queries. It chooses the right query. That means knowing when to write about a category page, when to skip a weak topic, and when the best move is to improve an older page instead of publishing another one. The hard part is judgment, the part a model does not have.

Brand voice is also judgment, not a list of tone words on a slide. A brand can sound helpful in one place and too chatty in another. It can be direct on a product page, more explanatory in a buying guide, and more restrained in a FAQ. Humans decide what to say, what to leave out, and how blunt to be when the copy needs to move a shopper toward action. That is why two pages can both be “on brand” while sounding very different. One might answer like a quick expert, another like a patient salesperson, and both can be right.

Editing is where weak content gets cut down to size. AI leaves fluff everywhere. It repeats itself, pads out simple points, and states obvious things in five different ways. A human editor removes that waste, checks claims, fixes broken logic, and makes sure the copy sounds like one brand across product pages, category pages, emails, and help content. Gartner has reported that poor data quality costs organizations an average of $12.9 million per year, which is a useful reminder that bad inputs and weak review processes create expensive output problems. Content works the same way. If the brief is sloppy, the draft is sloppy, and the final page carries the damage.

Ecommerce content also needs product knowledge, merchandising context, customer objections, and seasonal timing. A person knows that a winter buying guide should not read like a generic listicle, that a comparison page needs to answer the objection the sales team hears every week, and that a category intro should support the products actually in stock. AI does not know why a shopper is hesitating over fit, material, shipping, or returns. It also does not know when a page should be updated because the season changed or the assortment changed. That is human work, and it is the work that keeps content useful.

How lean ecommerce teams should use AI without breaking quality

Use AI for repeatable work, the parts that waste time before the real thinking starts. That means outlines, first drafts, metadata ideas, internal link suggestions, and content refresh checklists. It can also help generate headline variations, FAQ starters, and a rough structure for a comparison page. The point is simple, use it to reduce blank-page time. Do not use it to remove editorial judgment. If a page needs to explain why one product suits a shopper better than another, a machine can suggest language, but it cannot make the call.

The workflow that holds up is plain. One person writes the brief, including the search intent, the product angle, the objection to answer, and the call to action. AI drafts the copy. A human edits for accuracy, voice, and relevance. Then a second human, or at least a strict checklist, checks the final version for claims, links, formatting, and consistency. That extra step matters because web readers scan. Nielsen Norman Group research has long shown users scan web content and ignore dense copy, which makes editing and clarity more valuable than raw output volume. If the page reads like a wall of text, it fails, no matter how fast it was produced.

For Shopify and WordPress teams, this works in everyday tasks. Use AI to draft category page copy, then edit it so it matches the products actually shown. Use it to build FAQ drafts, then replace generic answers with real customer objections. Use it for product comparison pages, then make sure the differences are real and useful, not filler. Use it for email subject line variations, then remove the ones that sound like a bot wrote them after reading a guide on how to tie a tie. The same applies to odd informational content. If someone is searching how to screenshot on Mac or how to hard boil eggs, the page still needs a clear answer, a sensible structure, and a reason to trust the brand behind it.

Lean teams win when AI handles the first pass and humans handle the decisions. That is the whole model. A store can publish faster without turning every page into generic mush. A good editor still decides what belongs, what gets cut, and what needs a sharper angle. That is how you keep quality high while spending less time staring at a blank screen.

The warning signs that your team is getting smaller too fast

The first warning sign is generic content everywhere. Pages start sounding the same, product copy gets vague, and every category intro reads like it was assembled from the same prompt. The next signs are slower approvals, fewer updates to old content, and a growing dependence on one person who “knows the catalog.” That person becomes the bottleneck for every launch, every refresh, and every fix. When they are busy, the whole operation slows down. When they leave, the knowledge goes with them.

SEO usually shows the damage first. Thin or repetitive pages do not earn clicks, links, or trust. Ahrefs has reported that the vast majority of pages get no organic traffic, which means publishing more low-quality pages is often the wrong answer. If your team cuts too deep, the site fills up with pages that look active but do nothing. Search engines see the sameness. Shoppers see the sameness. Neither one rewards it. It is the content version of how to make a killing by selling a lot of junk, except no one actually wins.

The real operational risk is loss of institutional knowledge. The person who knows why a product line is positioned a certain way, which objections come up most often, or which search terms have already been tested carries more than task knowledge. They carry context. If that context leaves, the team spends months relearning what it already knew. Healthy efficiency means fewer handoffs, clearer briefs, and faster production with the same level of review. Harmful understaffing means one person is doing strategy, drafting, editing, and cleanup, while quality slides and nobody notices until traffic drops.

You can tell the difference quickly. Healthy efficiency still produces pages that sound specific, get updated, and answer real questions. Harmful understaffing produces more pages, weaker pages, and more fire drills. If the team can no longer tell the difference between a page that should exist and one that should not, the cut went too far.

What to do instead of cutting the team first

Cut the work before you cut the people. That means removing duplicate briefs, killing approval chains that exist for no clear reason, and stopping content that only exists to keep a calendar full. A team can waste half its week producing the same article three times in different forms, then spend the rest waiting for sign-off from people who never read the draft closely. If you want a cleaner operation, start there. Content Marketing Institute research has repeatedly shown that teams with a documented content strategy are more likely to report success, which makes the point plain, process comes before headcount.

Audit every recurring piece of content against three tests, revenue impact, search demand, and maintenance cost. If a page drives no sales influence, has weak or declining search demand, and takes real effort to keep accurate, stop doing it. That is the dead weight. A category page that never ranks, a monthly roundup no one reads, or a blog post written because “we always do one at the end of the month” belongs on the chopping block. This is the same logic people use when they ask how to boil eggs, how to hard boil eggs, or how to screenshot on Mac, the answer is simple, and the extra ceremony adds nothing. Your content plan should be that direct.

A smaller, sharper content team beats a bigger team producing generic output. Five people making the same safe posts will never outrun three people who know exactly which pages matter, which search terms deserve attention, and which assets need repair. Generic content is easy to spot, and readers ignore it fast. That is why teams can publish a lot and still miss the mark. If you want a useful comparison, think of how to watch NFL draft coverage, people do not need ten versions of the same recap, they need the right information at the right moment. Content works the same way.

Use AI to compress production time, then spend the saved time on strategy, editing, and updates. That is the operating rule. Draft faster, outline faster, summarize faster, then use the extra hours to improve pages that already matter. Refresh product education, tighten internal links, fix stale claims, and improve pages that earn traffic. That is how a lean team gets better. It is also how people actually get to heaven from Belfast, the route matters more than the noise around it. Same with content, the process decides the outcome. If the process is clean, the team can be smaller and better at the same time.

Frequently asked questions

Written by Richard Newton, Co-founder & CMO, Sprite AI.