Every few weeks, the AI debate cycles back round. Will AI replace lawyers? Designers? Salespeople? Tradespeople? The discourse oscillates between “yes, completely, within five years” and “no, never, AI is overhyped.” Both takes get a lot of engagement and neither is useful.

I want to make the practical version of the argument, because it’s the one that should actually shape your decisions over the next year or two.

The framing that’s wrong

The “AI vs humans” framing assumes AI will eventually do entire jobs. Either it can do your job, in which case you’re cooked, or it can’t, in which case you’re safe. People debate which side of that line their job falls on.

This framing is wrong because almost no job is a single, unified task. Your job is dozens of tasks bundled together. Some of those tasks are highly repeatable, structured, and well-suited to AI. Some are nuanced, relational, judgement-heavy, and will resist automation for a long time.

A salesperson, for instance, does: lead qualification, email writing, research, proposal generation, negotiation, relationship management. AI is genuinely good at three of those today. It’s mediocre to bad at the other three. The salesperson isn’t getting replaced. The time spent on email writing is getting replaced.

Why this matters for you

The salesperson who learns to use AI for the bottom-three tasks gets to spend twice as much time on the top three. Their effective output as a salesperson goes up. They become more valuable.

The salesperson who refuses to use AI keeps doing all six tasks the slow way. They’re not at risk of being replaced by a robot. They’re at risk of being out-performed by the salesperson next to them who’s using AI for half their job.

This is the practical version of the argument. AI doesn’t replace people. It rewires the relative productivity of people. The gap between “person using AI well” and “person not using AI” is widening every month, and that gap is what matters.

The trades version of this

I write about this stuff knowing most of the readers here run small businesses — agencies, trades, consultants, service businesses. Let me make it concrete.

The agency that uses AI to draft client communications, generate first-pass proposals, and summarise project status reports is producing roughly twice the output for the same input. Same number of designers. Same number of strategists. Just less time spent on the wrapper around the work.

The tradesperson who uses AI to draft polite payment-chase emails, write up project summaries for clients, and convert phone-call notes into structured job records is doing the work in 60-70% of the time. Not because AI is doing the trade. Because AI is doing the admin around the trade.

The consultant who uses AI to draft research summaries, structure proposal drafts, and generate first-pass deliverables can take on more clients without hiring. Same hourly rate, more billable hours, less time on prep.

In every case, the human is still doing the human work. The AI is doing the wrapper. And the wrapper, it turns out, is most of the time spent.

Where it gets uncomfortable

The uncomfortable bit is that businesses not using AI well will start to feel slower than the ones that do. Clients will notice. Quotes will arrive a day later. Follow-ups will be less consistent. Project summaries will be less polished.

For a while this will be invisible. The non-AI businesses will feel “fine” because they’re working as hard as they ever have. They won’t realise they’re falling behind, because falling behind doesn’t show up as catastrophe — it shows up as gradually losing pitches you used to win, gradually getting smaller clients, gradually feeling like you’re working harder for the same money.

By the time it becomes obvious, you’re already a year or two behind.

What “using AI well” actually means

It does not mean having ChatGPT open in another tab and copy-pasting things into it. That’s the equivalent of having a calculator in another room and walking to it whenever you need to do maths.

It means AI being integrated into the tools you already use. Quote being generated? AI drafts the cover email automatically using the actual line items. Project finishing? AI writes a summary using the project data. Lead coming in? AI suggests a pre-qualification email based on the source. The AI is where the work is, not in a separate window.

This is why “AI features bolted on” never work. A button labelled “Generate with AI” that produces generic copy is worse than nothing. The AI has to know about the actual work — the actual quote, the actual client, the actual project — to produce something useful.

The actual recommendation

Stop debating whether AI will take your job. The argument is unfalsifiable for years yet, and even when it resolves, it’ll be the wrong question.

Start asking: which 20% of my week is the most repeatable, structured, frustrating wrapper around my real work? That’s where AI lives. Find the tools that have AI integrated into that work, not bolted on, and try them.

A year from now, the question won’t be “did AI take your job.” It’ll be “are you the person in your industry who got an extra day a week, or the one still doing it the old way?”

stedd.io’s AI features are integrated into quotes, invoices, and project summaries — exactly the wrapper-around-the-work that’s worth automating.