Why does every AI rollout feel like it’s really about cutting people instead of making the product better?

– Senior Director of Customer Operations

Why are companies using AI to cut costs instead of create better products?

Why does every AI rollout feel like it’s really about cutting people instead of making the product better?

– Senior Director of Customer Operations

Because a lot of them are.

Not all of them. Let’s not be children. AI can make products better. It can improve search, discovery, personalization, support, accessibility, localization, workflow speed, recommendations, moderation, packaging, and decision-making. It can remove stupid work from smart people’s lives. It can help teams build things they couldn’t build before.

But that’s not how most companies are leading with it. Most companies are leading with cost. That’s the part everyone can feel, even when the press release is wearing innovation cologne. The CEO says transformation. The product deck says intelligent experiences. The internal memo says operational efficiency. The CFO hears margin expansion. The customer gets a chatbot that apologizes with the emotional range of a parking meter.

The question isn’t whether AI can create better products. It for sure can. The question is whether companies have the taste, patience, and product discipline to use it that way. A lot don’t. 

The Spreadsheet Wins First

Cost cuts show up faster than better products. That’s the boring answer, and boring answers are usually where the bodies are buried.

Headcount reduction goes in a model. Support deflection goes in a model. Faster content tagging goes in a model. Reduced QA hours go in a model. Fewer contractors go in a model. Automated copy, automated clips, automated reports, automated workflows, automated everything. Better product quality is harder. So is lower churn. So is higher trust. So is better discovery. So is a support experience that doesn’t make customers want to throw their phone into the ocean.

Those things matter, but they don’t always fit neatly into the next earnings call. They require connecting product quality to business outcomes over time. That takes discipline. It also requires admitting the product may be mediocre because leadership made it that way, not because the team lacked automation. Nobody wants that meeting, so AI gets framed around efficiency because efficiency is easier to explain.

Can we answer tickets with fewer agents? Can we generate marketing copy with fewer writers? Can we produce more assets with fewer editors? Can we move faster with fewer people? Sometimes the answer is yes. But “can we?” isn’t the same as “should we?” That’s where companies keep stepping on the rake.

Most AI Strategies Are Labor Strategies in a Hoodie

A real AI strategy starts with the customer. What problem are we solving? What experience can we improve? What decision can we make easier? What friction can we remove? What capability can we build that wasn’t possible before?

A lot of corporate AI strategy starts somewhere else: who can we replace? That’s not an AI strategy. That’s a labor strategy with better branding.

This is why so many AI rollouts feel cheap. The company isn’t using AI to make the experience meaningfully better. It’s using AI to make the old experience less expensive to operate. There’s a difference, and customers can feel it immediately.

An AI tool that helps a customer find the right show faster, understand a bill, manage subscriptions, customize recommendations, get better support, or discover something they’ll actually care about is product improvement. An AI chatbot that blocks the customer from reaching a human while confidently misunderstanding the problem is a cost-cutting moat with a smiley face.

Companies know the difference. They just prefer the cheaper one.

The Customer Can Feel the Spreadsheet

Customers aren’t dumb. They may not know whether a company is using a large language model, a rules-based bot, a third-party support platform, or three interns in a trench coat, but they know when the experience gets worse.

They know when customer service becomes a maze. They know when search stops understanding intent. They know when recommendations get weirder. They know when content feels mass-produced. They know when personalization becomes creepy instead of useful. They know when a brand starts sounding like every other brand because everyone’s using the same tools to generate the same flavorless corporate pudding.

That’s the danger. AI can help companies scale taste, but it can also help them scale sameness. Right now, too many companies are choosing sameness because it’s easier to operationalize. They’re filling apps, feeds, emails, help centers, product pages, and customer journeys with output that’s technically acceptable and emotionally dead.

The machine can generate words, images, summaries, recommendations, responses, and workflows. What it can’t do is decide whether anyone should care. That’s still leadership’s job, and bad news, leadership keeps trying to outsource it to the machine.

AI Should Remove the Stupid Work, Not the Human Judgment

The best use of AI isn’t replacing the human layer. It’s removing the stupid layer.

Every company has stupid layers. Internal workflows that waste time. Data trapped in systems nobody uses. Manual review processes that could be assisted. Customer journeys that require too many clicks. Search experiences that can’t understand natural language. Recommendation systems that confuse “you watched one cooking show” with “please show me poultry documentaries until I die.”

AI should attack that stuff. It should make products easier to use, support more useful, discovery more intelligent, and teams more effective. It should help people see patterns they couldn’t see before and compress the distance between customer intent and customer outcome.

That’s the prize. But that prize requires companies to think like product builders instead of procurement departments. Procurement asks how much this can save. Product asks what this can make possible. Most companies are stuck on the first question, which is why the results feel small.

Efficiency Isn’t Innovation

There’s nothing inherently wrong with cutting costs. Some workflows are bloated. Some tasks should be automated. Some old operating models deserve to be taken behind the barn. Companies should absolutely use AI to reduce waste, speed up operations, and remove low-value work.

But stop pretending every efficiency project is innovation. Innovation creates new value. Efficiency protects margin. Both matter, but they’re not the same thing.

This is the lie inside a lot of AI theater. Companies announce AI initiatives as if they’re reinventing the customer experience, when what they’re really doing is reducing the cost of running the same experience. That may be good business. It may even be necessary business. But it’s not a product breakthrough.

If the customer doesn’t feel the improvement, don’t call it transformation. Call it what it is: expense management.

Media Companies Should Be More Worried Than Most

For media and entertainment companies, the temptation is obvious. AI can cut production costs, localize content, generate clips, write metadata, assist dubbing, improve ad targeting, build promo assets, summarize scripts, automate customer service, and probably sit through three strategy meetings while producing the same outcome with less coffee.

Great. But media companies need to be careful because their product isn’t just distribution. It’s taste, trust, identity, emotion, and habit. When AI is used carelessly, it can make the company more efficient while making the brand less distinct. That’s a terrible trade.

Nobody subscribes because your metadata pipeline got cheaper. Nobody loves a service because its chatbot deflected more tickets. Nobody builds a viewing habit because the thumbnails were generated at scale by something trained on everyone else’s thumbnails.

AI can improve the media experience, but only if it’s pointed at things viewers actually feel: discovery, personalization, packaging, search, accessibility, localization, support, and smarter ways to connect people with the content they’ll care about. Use AI there and you might build a better product. Use it only to hollow out teams and you’ll build a cheaper version of the same mediocre experience.

Congratulations on the savings. Enjoy the churn.

A Quick Note for Media Teams Trying Not to Screw This Up

This is also why we published The Streaming Wars Guide to AI & The Modern Media Workflow.

Media companies don’t need another AI hype document. They need a practical way to understand where AI actually helps: faster metadata, better tagging, cleaner packaging, fewer manual steps, stronger distribution, smarter monetization, and less operational drag.

The useful question isn’t “where can we shove AI?”

It’s “where can AI improve how work gets done without sanding the product into generic mush?”

That’s what the guide is built for: media teams trying to move faster, reduce friction, and create more value from the same content and people before the spreadsheet crowd turns AI into another headcount exercise.

The Real Goal Should Be Talent Density

Here’s where this gets uncomfortable. The goal shouldn’t be “use AI so we need fewer smart people.” The goal should be “use AI so smart people spend less time doing dumb work.”

That’s a very different operating model. Good companies will use AI to raise talent density. They’ll automate repetitive tasks, compress workflows, improve decision-making, and give strong teams more leverage. They’ll keep humans where judgment matters: product taste, creative direction, customer empathy, brand voice, trust, and strategy.

Bad companies will use AI to justify headcount cuts, then wonder why everything feels thinner six months later. This is the pattern. The spreadsheet improves first. The product deteriorates later. Then leadership calls it a market challenge.

No, Brad. You replaced the people who understood the customer with a workflow tool and a quarterly savings target.

The Test Is Whether Customers Would Notice

Every AI initiative should have to answer one ugly question: would the customer notice if this worked?

Not the board. Not the consultant. Not the CTO. Not the executive trying to sound dangerous on a panel. The customer. Would they find something faster? Would they get better help? Would they understand the product more easily? Would they trust the service more? Would the content feel more relevant? Would the experience feel less exhausting?

If the answer is yes, keep going. If the answer is “we can reduce operating expense by 12%,” fine. That may be useful. But call it what it is. That’s cost reduction. Cost reduction isn’t evil. Pretending it’s innovation is.

Skip Says

AI shouldn’t be judged as cost cutting versus product improvement. The best companies will do both. They’ll use AI to remove waste, reduce stupid work, speed up teams, and make the product better for customers. The problem is that too many companies are stopping at the first half.

They’re treating AI as a socially acceptable way to say layoffs, then dressing it up as innovation. That may help margins in the short term, but it doesn’t automatically create better products, stronger brands, happier customers, or more durable businesses.

AI exposes the difference between companies with product vision and companies with expense targets. The lazy AI strategy asks who can be replaced. The better one asks what can be made possible.

That difference is the whole game.

Ask Me Anything

Whether you’re fed up, fired up, or just want the truth behind the trends, send me your questions using this form. Anonymity guaranteed. Bullshit not included.