

Weβre getting a little existential this week. Letβs play a little game.
Before you do your next task, whether itβs replying to an email, compiling a report, or trying to find that one slide you swear was in the folder, ask yourself this:
βCould AI be doing this instead of me?β
If the answer is even a hesitant maybe, then stop. Donβt do it. Not yet. Because hereβs the reality: if youβre still approaching every task like itβs 2019, youβre wasting time, burning energy, and quite possibly working yourself out of a job.
Your level of job automation sits somewhere along this line:

Accordingly, any of these tools are at your disposal to do more work, better and faster.
This article is your permission slip (and subtle warning) to stop doing low-leverage work manually. Below are 10 deceptively simple questions to help you figure out whether you should be doing a task or whether itβs time to let the machines take over.
1. Could AI realistically do this?
Weβre not talking about βin theoryβ AI. No sci-fi here. Weβre talking practical, generally-available, off-the-shelf intelligence. If the task is structured, rule-based, or based on information already online or in your systems, thereβs a high chance AI can do it, or at least help.
Still writing templated emails from scratch? Still summarising Zoom calls manually? These are examples of tasks that wonβt be required much longer.
2. Whatβs the actual logic behind this task?
AI doesnβt do magic; it does logic. If a task can be broken into conditions, steps, or triggersβlike βIf X, then send Yββthen itβs a strong automation candidate. You donβt need AI to take your job; you need it to take your flowcharts and run with them.
Zapier is a great platform for this, but there are a few in the market. On Zapier, though, you have to create the workflow step by step.
The next two years or so will deliver generally available solutions that will create the workflow for you from just a prompt.
3. Am I essential, or just available?
Hereβs a hard truth: some of the tasks you do arenβt yours because youβre brilliant. Theyβre yours because you were in the room when the question was asked.
AI doesnβt need to eat lunch or go to meetings. So if youβre the bottleneck only because no oneβs written the automation yetβ¦ that might not last.
4. Am I repeating myself?
If youβve done something more than twice, and you can describe the steps, itβs likely automatable. If you can describe it in a prompt, even better.
Repetition is the red flag. AI doesnβt complain about boring work, but it also doesnβt forget the steps.
5. Can I write a decent prompt for this?
Try it. If you can type, βSummarise this call into bullet points for a status updateβ or βCreate a draft email responding to this request with a polite no,β then youβve got a workflow.
Prompt fluency is fast becoming more important than Excel formulas. (You should still know a VLOOKUP though.)
6. Is this task creative, administrative, or analytical?
AI eats admin for breakfast. Itβs also getting worryingly good at analysis. Creativity?
Depends what you mean. If your βcreativeβ task is assembling a social media calendar or turning notes into a blog post, GPT-4 can do 80% of it.
That remaining 20% containing your judgement, voice, and timing is where you stay valuable.
7. Is there already a tool that does this?
Before hiring consultants or declaring that your job is βAI-proof,β open a browser. Thereβs probably a GPT plug-in, a Zapier automation, a Notion widget, a Copilot feature, or a Chrome extension that solves the problem.
You donβt need to build AI. You need to use it.
8. Could I prototype an AI-powered version in one hour?
Try it. One hour, max. ChatGPT plus your inbox plus a spreadsheet. Thatβs the new MVP.
If it works, great. Refine it. If it doesnβt, youβve learned something. Either way, you're not doing it manually anymore.
9. Could AI do the prep, and I do the judgement?
The best use of AI today is not full automation. Itβs partial delegation.
Let it do the donkey work, draft the brief, pull the stats, summarise the call, generate the outline. You can come in and make the decisions that matter.
Youβll move faster, think sharper, and frankly, work less.
10. Am I doing this because it's valuable or just because it's familiar?
Humans love rituals and are weirdly loyal to inefficiency. But AI isnβt sentimental. Itβs ruthlessly pragmatic.
If your job relies on habits instead of outcomes, AI is going to feel like a threat. But if you focus on results, it becomes your sidekick.
So ask yourself: are you adding value, or just doing whatβs always been done?
Hereβs the truth:
Thereβs good news in all this.
You donβt need to be replaced by AI to be made redundant by it. If your time is spent on tasks that AI could do faster, cheaper, or betterβand you donβt adaptβyouβre making a compelling case to be automated out of relevance.
The upside? You can do something about it. Right now. Today.
Start by asking better questions before you hit send, open a spreadsheet, or write that slide deck. Companies are not doing this in any meaningful way, yet. You could be the catalyst.
Could AI be doing this?
If the answer is yes, youβre not just saving time. Youβre futureproofing your career.
Temrel is a software engineering consultancy that helps organisations implement the kinds of Agentic AI automations weβre talking about above. Want to know more?
Reply to this email
Book a discovery call
