Six years ago I wrote that COVID would push every business online. It did. The question I get now is a different one: will AI replace my marketing team?
Wrong question. The better one: which parts of your marketing should never have been manual in the first place?
Where AI actually earns its keep
Plenty of places, and I use it daily. AI earns its keep in volume work. These are the tasks that eat hours and produce no thinking:
- Research. Market scans, competitor sweeps, category reading that used to take days.
- Reporting. Weekly performance reads without a manual export every Monday.
- Creative volume. Writing and testing 40 ad variants instead of 4.
- Lead handling. Scoring and routing leads the moment they land, not the next morning.
- Nurture. Sequences that used to sit in a backlog for weeks.
- Connecting the dots. Stitching ad platforms, CRM and analytics into one view instead of five tabs.
That's the 60%. It was always mechanical. It just used to need a person.
Where it doesn't
Strategy. And underneath strategy, the thing everything else rests on: actually understanding your customer.
Ask AI about your audience and you'll get a demographic profile that reads well and sounds right. What it cannot tell you is whether that profile holds. That only comes from having run campaigns against it, watched what actually converted, and been wrong a few times along the way.
Demography has to be tried and tested, not researched. Two brands can both sell to "male, 25-34, metro India" and still need completely opposite creative, because one audience buys on aspiration and the other buys on proof. No model tells you which is which. Someone has to spend budget finding out.
That's what experience buys you. If you've pulled something off in a particular domain, you carry things no dataset holds: which channel quietly stopped working last year, which segment looks small on paper but closes fastest, what your competitor already tried before it failed.
Research is leverage, not the answer
This is the distinction most people miss. AI gives you a genuine head start: faster inputs, wider reading, cleaner data. That's real leverage and I'd never give it up.
But leverage isn't a decision. Everything after the research is a judgement call. What to bet on. What to cut. When to hold through a bad week, and when to trust the market over the dashboard. Judgement gets built by being accountable for outcomes, not by reading about them.
AI only does more of whatever you point it at. Point it at the wrong audience and it will produce the wrong campaign faster, at scale, and in a confident tone. The output is only ever as good as the person behind the machine.
The compounding effect
The businesses pulling ahead right now aren't the ones with the best AI tools. They're the ones who automated the mechanical layer, freed up their thinking hours, and spent those hours on strategy and execution. That gap compounds every quarter.
What this means for your business
- Audit where your team's hours actually go before buying a single tool.
- Automate the repeatable layer first: research, reporting, lead routing, nurture. It pays back fastest.
- Treat AI's audience research as a hypothesis, never a conclusion. Test it small before you scale it.
- Put someone with domain experience on the execution. That's where the money is won or lost.
The winners in 2026 won't be the businesses that adopted AI. They'll be the ones that knew what to keep human.
Want this working for your business?
I help founders and marketing leaders turn ad budgets into measurable pipeline. Clear brief or vague challenge, either works.