A lot of teams are already using ChatGPT.
They use it to write emails. Polish documents. Create captions. Summarise notes.
That can save a little time. But there is a bigger question worth asking:
Has the work itself improved, or are we simply doing the same work with a new tab open?
Using AI is not the same as improving work
If every person prompts differently, checks differently and stores the output differently, the organisation has not created a better system. It has created several personal shortcuts.
Useful? Sometimes. Reusable? Not yet.
The real opportunity is to look at the full piece of work. Where does it begin? What information does the person need? Where do they get stuck? Who checks it? What happens to the output next?
That is where AI can become part of a better process instead of a clever side activity.
Start with the bottleneck, not the tool
Imagine a manager who spends every Friday preparing a project update. She collects messages from five people, copies numbers from two spreadsheets and rewrites everything into one report.
The tempting question is: Which AI tool should we buy?
A better first question is: Why is this report so difficult to produce?
Perhaps the inputs arrive in different formats. Perhaps nobody agrees on what the update should include. Perhaps the final report repeats information that already exists elsewhere.
If we skip those questions, AI may only help us produce a messy report faster.
A simple way to look at it
Take one repeated task and map four things:
- Input — What information enters the task?
- Judgement — What decisions need a person?
- Repetition — Which steps happen in almost the same way every time?
- Output — Who needs the result, and what do they need to do with it?
Now you can see where AI may assist, where a clearer template may be enough and where human judgement must stay in the loop.
Brief. Ground. Check. Reuse.
When AI does belong in the process, I use four simple reminders.
Brief
Tell it the job, the audience and the outcome. A vague request creates a vague result.
Ground
Give it the source material and business context it needs. Do not expect it to know your policy, customer or way of working.
Check
Decide who verifies facts, tone, judgement and risk before the output is used.
Reuse
Turn what worked into a shared prompt, checklist or workflow. The value grows when the team does not have to reinvent it every time.
The workplace question I would ask
Do not ask your team, Are you using AI?
Ask this instead:
Which repeated task is taking too much time, and what would a genuinely better version of that task look like?
That conversation will tell you far more than a list of tools ever will.
Less AI hype. More business impact.
