AI-enabled marketing
AI makes marketing faster. It doesn't make it right. I use AI throughout the work — research, drafting, reporting, the systems that run in the background — with a senior marketer deciding what's worth doing and what ships.
What I build
The difference worth paying for isn't prompting. It's the systems that keep running when nobody is watching them.
Intelligence
A pipeline that watches your industry — news, competitors, regulation — filters it, checks it against the source, and delivers what actually matters on a schedule.
Operations
The unglamorous plumbing: routing, enrichment, lifecycle stages, and the classification work that otherwise eats an afternoon a week.
Reporting
Reporting that assembles itself from the source systems each month, instead of being rebuilt by hand and drifting from the truth a little more each month.
The honest version
It helps
It costs you
The rule
Use AI for the volume. Keep a human on anything a customer will judge you by.
Readiness
Most businesses are further along than they think in one area and nowhere in another. This is the order I look at it.
Which recurring marketing tasks cost the most hours for the least judgement. That is where automation pays first.
Where to startAI is only as useful as what it can see. A half-populated CRM limits what is possible more than the model does.
The constraintThe lines you will not cross — and writing them down before anyone is tempted, not after.
The boundaryOne system, shipped and running, before the second one is designed. Automation that is never finished saves nothing.
The workWhat AI can't do
A language model will agree with you. Ask it whether your positioning is good and it will find reasons that it is. That agreeableness is the single most expensive property of these tools in a marketing context, because strategy is exactly the place where you need someone to tell you no.
It will also invent things. Statistics, quotes, case studies, dates — presented with the same confidence as the true parts. I have caught this in my own systems more than once, which is why anything published from my pipelines is checked against the original source before it ships. If a claim cannot be traced, it gets cut rather than softened.
And it does not know your customers. It knows what has been written about businesses like yours. That is a useful starting point and a terrible finishing one.
Common questions
No, and anyone selling that is selling you a problem for later. AI is very good at volume — drafting, summarising, scanning, assembling. It is poor at the things a marketing manager is actually paid for: deciding what matters, knowing what will offend a customer, judging when a campaign is wrong, and being accountable when it is. It replaces the hours, not the judgement.
Research and competitive scans, first drafts of everything, automated intelligence digests that watch an industry and tell me what changed, CRM and reporting pipelines that assemble themselves, and internal tooling that removes admin from the week. All of it is checked by a person before it reaches a customer.
Not if I am editing it. I keep a written list of the words and constructions that give AI writing away, and I lint against it. A first draft from a model is a starting point with a recognisable accent; removing that accent is editing work, and it is the part most people skip.
Client material goes into tools with no-training terms, and never into a consumer chatbot. Anything sensitive stays out of a model entirely. I am happy to walk through exactly which tools touch which data before we start — it is a fair question and it should have a specific answer, not a reassurance.
Next step
Half an hour on a call and I can usually name the two things in your marketing week that should stop being done by hand.
