Digital marketing and AI for teams that want to do it themselves
AI does not replace digital marketing. But it can support creative work and take recurring tasks off your team, from the monthly report to the twentieth ad variant. I show you how that works day to day, and I stay reachable when something jams.
Costs nothing.
Open the one that fits best. Underneath you will find what is usually behind it and where I would start.
It is rarely one single channel. More often demand has shifted while your measurement stood still. The reports show clicks and reach, but not which enquiry came from where. So budget wanders from channel to channel without anyone seeing the effect.
With a look at your measurement: does it count enquiries or only clicks? That is settled within days. Until it is, I would not change a single campaign, because every further euro is a bet.
Capacity is rarely the problem. Usually the projects are cut too large. When every round of development is expensive, you pack as much into it as you can to make it worthwhile. But large projects take longer, fail more often and answer the real question last. Once a round becomes cheap, the small cut pays off again.
With the smallest version that lets you test whether the idea holds. It should be running in two weeks. The finished product comes afterwards, and it often looks different from the plan.
Usually it is less about reach than about a picture of you. Anyone applying wants to know beforehand what it is like to work with you, and a job ad rarely answers that. As long as that picture is missing, more ad budget only spreads an empty message.
With the question of who in your company wants to be visible and is allowed to be. Without that, every channel becomes a chore and falls asleep after three months.
Nothing here? Then write me two sentences about what is pressing.
Four tasks from everyday marketing. You give the AI a brief and your material, and back comes a first draft. Whether it is any good depends on a few tricks, and those are listed under every example. They are what I teach your team.
Here is the campaign export for August. Turn it into the monthly report for management: the essentials at a glance and what we should change.
campaigns_august.xlsx · 279 rows
date channel campaign cost clicks leads 2026-08-01 google brand_exact 84.20 312 6 2026-08-01 google generic_broad 143.90 655 0 2026-08-01 meta retargeting 121.50 1,877 2 2026-08-01 meta lookalike_1pct 88.10 2,410 0 2026-08-01 linkedin dach_decision_makers 95.00 104 4 2026-08-02 google brand_exact 79.60 298 5 2026-08-02 google generic_broad 151.20 702 1 …
Monthly report, August
279 rows analysed
412 Enquiries ▲ 9% vs. July
45,66 € per enquiry ▼ 6% vs. July
Cost per enquiry by channel
LinkedIn 30,16 €
Google 37,45 €
Meta 98,31 €
What stands out
Suggestion
Scale both campaigns back and test the freed budget on LinkedIn for four weeks. Then compare again.
The trickNever let AI estimate numbers, let it calculate them. Good tools show their working, and someone checks it before the report goes out.
Prepared examples with invented data, and the results are shortened. They run without a network connection and without AI in the background, so the page stays fast.
12 creatives · meta_ads_q3.csv
What the strong creatives have in common
12 creatives · attributes read from the images
| Attribute | Click rate with | without |
|---|---|---|
| Person in the image | 1,9 % | 0,8 % |
| Six words of text at most | 1,8 % | 0,7 % |
| Warm background | 1,5 % | 1,1 % |
| Product front and centre | 0,9 % | 1,6 % |
What stands out
Creatives with a person and little text lead the field. The product on its own pulls the weakest, however beautifully it is shot.
For the next brief
A person cropped close, six words at most, warm background. The product waits for the landing page.
Twelve creatives are a small base. Treat this as a hypothesis for the next test, not as a rule.
customer_interviews.docx · 8 interviews
Interview 3 · Purchasing, mechanical engineering “I always get three quotes. I read the first one properly and skim the others.” “Price on request means too expensive, as far as I am concerned.” Interview 5 · Technical lead “If I have to call someone just to get a price, I am out.” Interview 7 · Purchasing, supplier “I hardly go to trade fairs any more. I find what matters online first.” …
Persona from 8 interviews
Katrin, 46
Heads purchasing at a mechanical engineering company.
Our idea is an ad that says “Book your consultation now”. Would you click?
Probably not. A meeting means I only learn the price after an hour of talking. Show me a price range and I will happily talk.
based on interviews 3 and 5
Which trade magazine do you read?
Your notes do not say. You would have to ask your real customers.
transcript_jana.txt · 24 minutes
[03:12] I started here as a career changer. Before that I was a pastry chef. On my first day I thought: I will never manage this. [03:40] My foreman said: tolerances are like cake bases. Measure and you do not have to guess. [11:05] What keeps me here? That I can flag a mistake before it gets expensive. …
LinkedIn post
Jana was a pastry chef before she joined our assembly team. On her first day she thought: I will never manage this. Today she says: “Tolerances are like cake bases. Measure and you do not have to guess.”
Careers page
Plenty of us come from completely different trades. Here you can flag a mistake before it gets expensive, and someone will teach you how to measure.
Three short videos
In all three I work the same way: set it up, show it, stay reachable.
With AI an idea turns into a tool that runs. Nobody needs to know how to code: I set everything up, and your team builds the first one together with me.
Ideas can be tested in days instead of quarters, and your team can do the testing.
First I check whether your measurement captures what really counts. Then we develop a strategy that fits the market and the budget, and set up reports that AI writes from your exports.
Your team reads the numbers itself and acts on them, with no need for me as an interpreter.
I set up social media channels that fit your company, along with routines that do not overwhelm a single person. With AI one conversation becomes material for many posts.
Your team posts regularly and in its own voice, even when the day is full.
I combine marketing expertise with technical curiosity. I have worked in digital marketing for fourteen years, most of that time responsible for teams, and I always wanted to know what sits behind the tools.
One is worth little without the other. People who only know the tools automate the wrong tasks. People who only know marketing never get past trying things out. This website came about the same way: built with AI, without me being a developer. So I have walked the path your team is meant to walk.
For three years I taught at a university. Teaching is a fixed part of the work for me.
Roles, projects and numbers gladly in conversation.
With every step I do less and your team does more.
Measurement, prompts, templates, routines. What gets built stays, even when I am gone.
On a real project, not on an example. Workshops without a project of your own fizzle out.
You do it. I am reachable when something jams.
Your team develops the routines further on its own. Whoever joins next learns it from you, not from me.
After every step you have something in your hands that you can use. Never just a document.
So that you know from the start where you stand with me.
I do not run communities and I do not work through editorial calendars. I set the channels up and train your team, which then posts for itself. Whoever hands the posting over hands over their own voice as well.
Every project has a handover goal. Sparring after that, gladly, even for years. But I do not permanently replace anyone on your team.
On site in Hamburg, remote everywhere else.
One conversation to work out whether we are a fit. If we are not, I will say so.
Request an intro callCosts nothing.