AI or hire an admin? An operator's honest trade-off for a trades business.

The short answer

It's rarely one or the other. Automation is best at the repetitive, rules-based grind: capturing enquiries, chasing quotes, reminding customers, moving data between tools. A person is best at judgement, relationships and the messy exceptions. The smart move is to automate the grind first, then decide if you still need a hire, and for what.

What automation is genuinely good at

• The same task, every time. Capturing enquiries, sending booking confirmations, chasing unpaid invoices, following up quotes. It never forgets, never has an off day, never goes on leave.
• Speed and volume. It answers at 2am, handles ten enquiries at once, and does the boring bit instantly.
• Moving data between systems. The copy-paste between your job app, accounting and email that quietly eats hours. That's the definition of automatable.
• Fixed cost. A monthly tool cost that doesn't rise with volume, get sick, or need managing.

What a person is genuinely good at

• Judgement and exceptions. The job that doesn't fit the template, the customer who needs handling, the call about whether to squeeze someone in. A person reads the situation; a rule can't.
• Relationships. The regular who rings and wants to talk to someone who knows them. That loyalty is worth more than the ten minutes a bot would have saved.
• Chasing the messy stuff. The supplier who's stuffed up, the council form, the thing that needs a phone call and some tenacity.
• Owning it. A person notices when something's gone wrong and fixes it. Automation does exactly what you told it, including the wrong thing, until you catch it.

Don't ask 'AI or a person'. Ask 'which of my admin hours are the same every time, and which need a brain'. Automate the first list, hire for the second. The bill usually halves.

How the numbers usually shake out

• A part-time admin hire is real money every month, plus super, plus the time you spend managing them, plus the risk they leave and take the knowledge with them.
• Automating the repetitive load is a build cost then a small monthly, and it doesn't walk out the door.
• But automation can't cover the human-judgement work, so if that's most of your admin, a hire is the right call and no tool replaces it.
• The trap is paying a full wage for work that's 70% robotic. Strip out that 70% with automation and you either don't need the hire yet, or you get a much better one focused on the work that matters.

Common questions

Can automation fully replace my office person?

Usually not, and you probably wouldn't want it to. It replaces the repetitive slice of their job, not the judgement and the relationships. If someone spends their day on data entry and reminders, most of that can go; if they spend it handling customers and solving messes, it can't.

I'm drowning in admin. What do I do first?

Write down where your (or your partner's) admin hours actually go for a week. You'll see a clear split: the same-every-time tasks and the needs-a-brain tasks. Automate the first pile before you post a job ad. You often find the hire you thought you needed shrinks to a few hours.

Isn't hiring safer than trusting software?

Both carry risk. A hire can leave, get sick or make mistakes; a badly-set-up automation does the wrong thing at scale. The safe path is automating the boring, low-risk, repetitive work and keeping a human on anything with judgement or money attached.

Where I fit

Before you commit to a wage, it's worth seeing how much of your admin a system could just do. That's what I build: automations for the repetitive load, from $1,500 fixed, so any hire you make is for the work that genuinely needs a person. More in the AI for tradies cluster.