Today was my first day at a new company. Not the first time I've started a job, but the first time I did it with an AI assistant handling the logistics while I focused on the actual work.
Here's what that looked like.
The morning routine that wasn't
My first day started like any other Monday: coffee, open the laptop, check messages. But instead of scrambling through onboarding paperwork, I had an AI that had already pulled together everything I needed.
The first request was simple enough: "I need to fill out my employee personal data form. What do they need?"
My assistant pulled the data from my personal files — tax ID, address, health insurance info, bank details — and had it ready in seconds. Not a generic list of "things you might need." The actual numbers, formatted the way a German HR department expects. IBAN with spaces, Steuer-ID, Krankenkasse name and number. All of it.
In a normal first day, I'd spend an hour digging through old emails, PDFs, and password managers to find this stuff. Today it took about two minutes.
Getting into the codebase
Next up: repo access. The company uses GitLab for their main group, and I needed to clone two services I'd be working on.
Again, the assistant handled it. Set up the SSH alias, cloned the repos, verified access. The only thing I had to do was type my passphrase. Everything else — SSH config, key path, clone URLs, verifying the remotes — was already sorted.
Two repos cloned, code open in the editor, and I hadn't even finished my coffee.
The calendar, without the chaos
First days usually mean a flood of calendar invites. Orientation calls, team introductions, setup meetings. My assistant kept track of all of it through the day, making sure I knew what was next without me having to constantly check Google Calendar.
Not rocket science. But when you're trying to absorb a new codebase, a new team, a new domain — having someone (or something) watch the clock for you? That's the kind of small thing that makes a real difference.
Why this matters
What I found interesting is mundane: an AI assistant removed about two hours of friction from my first day. Finding documents. Setting up access. Keeping time. These tasks aren't hard, just annoying. And when you're starting a new job, the last thing you need is annoyance.
The real value was in doing things I didn't want to waste time on, so I could focus on what mattered: understanding the architecture, meeting the team, figuring out where the bugs live.
What I actually did today
For the record: I reviewed the engineering roadmap, explored two microservices (one for search, one for pricing), studied the infrastructure setup, and got oriented with the team's workflow. The kind of stuff you're supposed to do on day one.
The personal data, the repos, the calendar? That was handled in the background, the way a good colleague would grab you a coffee while you're setting up your desk.
What's next
I don't know yet if this is how every first day will look from now on. Probably not — most people don't have an AI assistant configured to their personal files and workflow. But it's a glimpse of something practical: not AI doing your job, but AI clearing the path so you can actually do your job.
That's the version of AI I find interesting: the one that makes sure you have time to write good code, not the one that writes it for you.