Your Job Is Already Being Done by Claude. You Just Haven't Been Told Yet.

Back to Blog

Let's skip the pleasantries. While you were debating whether AI is "just a fad," Anthropic quietly published data showing that 60% of work at their own company now runs through Claude — up from 28% just twelve months ago. Engineers there are merging 67% more pull requests per day. Productivity is up 50%.

And that's not some aspirational marketing figure. That's measured, internal, operational data from a company eating its own cooking.

Now here's the part that should make you uncomfortable: Claude's use cases page — titled with the almost patronisingly cheerful "Get inspired by what you can do with Claude" — reads less like a feature list and more like a job description. Your job description.

The Use Cases That Should Keep You Up at Night

Anthropic's own showcase includes Claude doing contract redlining and negotiation, identifying 19 strategic edits in a vendor agreement — removing auto-renewal traps, capping liability, ensuring data ownership. The kind of work a mid-level in-house lawyer bills at £250 an hour to do over three days. Claude does it during a coffee break.

There's workflow improvement planning, where Claude takes your messy process documentation, identifies bottlenecks, maps inputs and outputs, defines success metrics, and hands you back a structured implementation plan. That's the work of a £600/day operations consultant.

There's career path planning, where it analyses your CV against target roles and produces an action roadmap with skills tracking. That's what career coaches charge £150 a session for.

And there's the one that should really sting: "Explore what Claude can do for you" — a use case where you simply tell Claude what your job is, and it tells you exactly how it can do chunks of it. The tool is literally mapping its own capability against your role. In real time. To your face.

The Numbers Don't Lie — But They Do Terrify

Anthropic's own research found that for computer and maths workers, large language models are theoretically capable of handling 94% of their tasks. Currently, Claude covers about 33% of those in observed professional use. That gap is not a comfort — it's a runway. The 33% is where we are today. The 94% is where we're headed.

Dario Amodei, Anthropic's CEO, has publicly stated that AI could disrupt half of entry-level white-collar work and push unemployment as high as 10–20% within one to five years. This isn't a doom-mongering blogger. This is the CEO of the company building the tool, telling you what it's going to do.

Employee concern about AI job loss has surged from 28% in 2024 to 40% in 2026. The other 60% either haven't been paying attention or are already using Claude themselves.

The Real Question

The question isn't whether AI will change your job. It already has. The question is whether you're the person using Claude to do the work of three people, or you're one of the three people being made redundant.

Anthropic's own data shows a roughly 50/50 split between augmentation and automation — 52% of Claude usage involves augmented tasks where humans and AI work together, and the rest is pure automation. But that split is moving. Fast.

Every time Claude completes 20 consecutive actions before needing human input — doubled from 10 just six months ago — that augmentation line slides further toward full automation.

What This Means for You

If you're in a role that involves reviewing documents, writing reports, analysing data, managing workflows, drafting communications, or planning strategy — Claude can already do meaningful portions of your job. Not in theory. Right now. With a £15/month subscription.

The companies that understand this aren't waiting for you to catch up. They're already restructuring.

The "Get inspired" tagline on Claude's use cases page is perhaps the most British understatement in Silicon Valley history. You shouldn't be inspired. You should be mobilised.