AI marketing consultant: what AI does in your ads, feeds and reports
Everyone wants to do something with AI. In marketing it pays off mostly in routine work: search terms, feeds, reports and ad variants. The choices, your brand and your margin stay with a human.

What is AI in marketing?
AI in marketing is the use of smart systems for marketing work: choosing bids and audiences, making copy and images, improving product data and writing reports. In practice you work with three layers at once.
The platform's AI
Google and Meta decide more and more themselves. That AI is strong, but it follows the goal you give it, and the platform's own.
For example: Smart Bidding, Performance Max, AI Max and Advantage+.
AI assistants
Language models that help with writing, analysis and research. Fast, but they do not know your account or your margin.
For example: ChatGPT, Claude and Gemini for a first draft of a text or an analysis.
Your own agents on your data
Agents connected to your accounts and feeds, with your rules and a human who approves. This is where the difference is.
For example: my crew: Leo, Stack, Pitch, Tally and Hoot. See AI agents.
The gap between using AI and getting results is almost always in layer three: AI that knows your own data, with oversight.
Sound familiar?
“Everyone wants to do something with AI now”
You want to know what pays off and what is hype, before any budget goes to it.
“Google and Meta automate more and more”
Performance Max, AI Max and Advantage+ decide a lot themselves. What do you still steer?
“You do not want AI putting things live on its own”
Rightly so. A wrong price in an ad or a budget that is too high happens fast.
Where AI in marketing already works
The platforms already use a lot of AI themselves. You make the difference with what you put in and what you check. For each part you see which of my crew looks along.
Google Ads

- What the platform does itself
- Smart Bidding, Performance Max and AI Max pick search terms, copy and landing pages themselves (Google).
- What I add
- Steering on margin instead of revenue, going through search terms every week and excluding what costs money. Pitch proposes, I decide.
Shopping and product feeds

- What the platform does itself
- Google reads your feed and turns it into ads. What is missing or wrong costs impressions.
- What I add
- Checking prices, stock and disapprovals every day, and improving titles and attributes. That is Stack's job.
Facebook and Instagram Ads

- What the platform does itself
- Advantage+ chooses audiences, placements and combinations of image and text itself.
- What I add
- Making variants within your brand rules and checking every text for price and promise before it goes live. That is Leo's work.
Getting found in AI

- What the platform does itself
- ChatGPT, Gemini and the AI overviews in Google name brands they come across in reliable places.
- What I add
- Putting the same customer questions every month and measuring whether you are mentioned. Tally measures; Bob Crombag does the SEO, in the Netherlands.
Measurement

- What the platform does itself
- GA4 and the platforms fill gaps in measurement with models. Handy, but not always right.
- What I add
- Watching that conversions and values stay right, and flagging when something drops out. Also Tally.
Reporting

- What the platform does itself
- Google Ads and Analytics recently got their own AI assistants that answer questions (Google).
- What I add
- A weekly report with what stands out, where every number is checked against the source. Hoot writes it.
How the crew works and which permissions each agent has is on AI agents.
What AI should not do in your marketing
Let AI do
- Going through lists: search terms, products, disapprovals
- Spotting deviations in costs, prices or measurement
- First versions of copy and variants
- Writing reports, with a check on every number
Keep with a human
- Your strategy and the choice of where budget goes
- Your brand: tone, promises and what you do not say
- Anything irreversible, such as raising a budget or putting a campaign live
- The conversation with you
Is your data ready for it?
AI is only as good as what you put in. Before I automate anything, I check four things. If one is missing, I start there.
Conversions are right
Does your measurement count real purchases and inquiries, without duplicate or missed conversions? Tracking and measurement
Value and margin are known
Does the platform know what an order earns after cost of goods? Otherwise AI steers on revenue instead of profit. How to calculate ROAS
Your product feed is complete
Titles, attributes, prices and stock that are right. The feed is what Shopping, Performance Max and ChatGPT ads read. Product feeds
Access and permissions are sorted
Who may do what in which account? Agents get only the permissions they need.
Google says so itself: start with simple tasks that come back often, such as the weekly report (Google).
How it works

A trial run first
Every new task first runs as a trial, so I see what the AI does before it matters.

Then live
Only then does the task go live, with checks on live data instead of on a file.

Final check and approval
Anything irreversible, I approve myself. The decisions stay with me.
What it costs
If you already work with me on Google Ads, Shopping or Meta, I simply use my own tools in the management. Collaboration from €2,000 per month, see how I work and pricing.
Want agents on your own systems, or want to automate outside marketing as well? Then it starts with the scan. A custom build project then costs €5,000 to €15,000, maintenance €500 to €1,250 per month. For Google Ads you can also start right away with one of my Google Ads MCPs, from €2,000.
Prices excl. VAT.
Who it suits and who it does not
Good fit
- Webshops and lead generators with a media budget from €5,000 per month
- Marketers who want to know what AI does for their work, without hype
- Teams that lose time on reports, feeds and search terms
Less of a fit
- Anyone looking for a workshop or course: this is execution, not training
- Anyone who wants AI to do everything itself, without oversight
Frequently asked questions
What is AI in marketing?
AI in marketing is the use of language models and smart algorithms for marketing work: bids and audiences in ad platforms, making copy and images, improving product data and writing reports. It pays off most in recurring routine work.
Will AI take over my job?
No. AI helps with routine work such as reports, checks and product data, leaving more time for what really matters. The setup, the choices and the final check stay with me, and nothing goes live without me having reviewed it myself.
What is the difference from Performance Max or Advantage+?
That is the platform's own AI. It works well when you give it the right signals, such as margin and reliable conversions. My own tools check what the platform does and steer toward your goal.
Is my data safe?
By default my tools get read-only access, and a human approves irreversible steps before they are carried out. We agree up front which data is used for what.
Which AI do you use yourself?
For writing and analysis I work with language models such as Claude and Gemini. On top of that I use my own connections to Google Ads, Merchant Center, GA4 and Meta. Which model fits best differs per task; I choose that per agent.
What is agentic marketing?
Agentic marketing means AI agents carry out multi-step marketing work themselves, such as going through search terms or writing a report, within limits you set. With me, a human approves everything that goes live or costs money.
Can AI manage my Google Ads?
Partly. Google's AI already decides a lot by itself, and agents can do routine work. Strategy, budget choices and checking what the platform does remain human work. That is also where the difference in results comes from.
Does that cost extra?
Not if you already work with me: my tools are part of the management. If you want agents on your own systems, it starts with the AI automation scan of €2,950.
Sources
- HubSpot, From output to outcomes, 2026
- HubSpot, State of Marketing 2026
- Eurostat, 20% of EU enterprises use AI technologies, December 2025
- Google, about AI Max for Search campaigns
- Google, channel reporting in Performance Max
- Google, Ads Advisor and Analytics Advisor
- Google, agentic AI in marketing
Updated November 18, 2026.
Start with the AI automation scan.
Then you know what it returns.