// in this post
Most advice on how to use AI at work is too abstract to use. You don't need a philosophy of artificial intelligence. You need to know what to actually do with it on a Tuesday.
In practice, using AI well at work comes down to a handful of specific jobs you hand to AI, check, and reuse. We have published detailed guides to the ten most common ones. This page pulls them together, explains what each is for, and covers the one habit that decides whether AI makes you more productive or just makes more work for everyone around you.
How to use AI at work without creating workslop
A quick definition, because it matters. "Workslop" is the term that has stuck for AI output that looks finished but isn't: the plausible but hollow report, the summary that misses the one caveat that mattered, the email that commits you to something you never agreed. It does not save anyone time. It moves the work downstream to whoever has to catch the problem.
Using AI at work safely and well comes down to two habits, and neither is about the tool. Tell the AI what "good" looks like before it starts, and check what it gives you against that before you pass it on. Every use below works when you do those two things, and produces workslop when you don't.
Ten ways to use AI at work
Working with information
- Summarise long documents, threads and meetings. Pull the decisions, actions and open questions out of anything long, and check it before you trust it.
- Get up to speed on an unfamiliar topic. Go from nothing to functional on a new subject in twenty minutes, with sources you can verify.
Communicating
- Write and rewrite work emails. Draft the awkward ones, fix the tone, and keep it sounding like you.
- Draft customer and client replies. On-tone replies and a reusable template bank, without inventing a policy you do not have.
- Meeting prep, notes and action items. A tight agenda before the meeting, owners and actions after it.
Producing the work
- Turn messy notes into SOPs. Turn the thing you keep re-explaining into a document someone can actually follow.
- Build a first-draft slide deck. Get the story and structure right before you open PowerPoint.
- Spreadsheets: formulas, cleaning and analysis. Describe your columns and the result you want, and get the formula plus an explanation.
Automating the repetitive
- Write and debug scripts when you are not a developer. PowerShell, Python, SQL and regex for the routine jobs, the first real step into automating your own work.
Thinking
- Pressure-test your ideas and decisions. Get AI to argue against your own plan and find the holes before you commit.
How to use AI at work effectively: the method
The ten above are different jobs, but they share one shape. Learn this loop and you can apply AI to almost anything you do repeatedly:
- Describe the task and the constraints. Not "write an email", but who it is for, the tone, the facts, and what must not happen.
- Give it the raw material you already have. AI is at its best turning something you know into something that exists, rather than inventing from nothing.
- Write a prompt good enough to reuse, and save it. The second time you run it is where the time-saving compounds.
- Check the output against what "good" looks like. This is the step that prevents workslop. The tool produces; your judgement is what makes it worth keeping.
- Keep what works. When a prompt reliably gets you there, it has earned a place in how you work.
Effectiveness comes from steps one and four, the clear instruction and the honest check, far more than from anything clever in the middle.
Which AI tool should you use at work?
People expect the tool to be the answer. It rarely is. Any capable assistant, whether Claude, Copilot or Gemini, will do all ten of these jobs. What changes the result is the clarity of your instruction and the quality of your check, not the logo on the tab. Pick one tool, learn it properly, and resist switching every time something new launches.
FAQ
How do I start using AI at work?
Pick one task you do repeatedly, do it with AI this week, and check the result honestly. Build from there. One use you trust beats ten you have only read about.
How can I use AI at work safely?
Keep confidential and personal data out of consumer tools, and check every output before you act on it or pass it on. Each of the ten guides has a short section on how to stay safe for its specific task.
How do I use AI effectively at work?
Be specific about what you want and what "good" looks like, give it the material you already have, and verify the result. Effectiveness comes from the instruction and the check far more than from the tool.
What is "workslop"?
Workslop is AI-generated work that looks polished but lacks substance, which shifts the effort of fixing it onto whoever receives it. The fix is a proper check before you pass anything on, which is the habit every good use of AI is built around.
Where to start
Pick the one task that matches something on your to-do list today, and read that guide. Do it once, properly, end to end. That single rep teaches you more than reading all ten.
All ten guides, plus the AI Fluency course and the rest, are free and ungated at thehumanco.org/ai-resources. And if you are rolling AI out across a team and want it to stick rather than scatter, get in touch. Helping organisations build AI capability is the work I do.