AI in marketing is already part of many teams' daily work, but it still does not cover what they need: in Spain, 74% of marketers say they need more personalised content than they can produce, and only 57% are turning to AI to close that gap (Salesforce, State of Marketing 2026). The difference between using AI to write and using it to work better is connecting it to the team's tools and data. Here is how we do it in our own marketing team.
How is AI used in marketing?
AI in marketing is used in four areas: creative and content, data and analysis, SEO and visibility in AI answers, and team operations. In all of them the most useful move is not asking for a piece of text, but connecting the AI to the tools the team already uses so it prepares the work and the team decides.
| Area | What the AI does | What the team decides |
|---|---|---|
| Creative and content | Proposes variants of assets, scripts and copy from a brief | Which variant fits the brand and the message |
| Data and analysis | Prepares the weekly report and flags the metrics that move | What to act on, and in what order |
| SEO and AI visibility | Cross-checks crawl data and prepares the technical issues | What gets fixed, and in what order |
| Operations | Queries and summarises the state of the team's tasks | Who does what, and when |
Marketing · Four flows
And where it stops, in each one.
Creative and content
The AI proposes variants from the brief. The design team decides and adjusts.
Data and analysis
The AI prepares the report and flags what moves. The team decides what to do.
SEO and AI visibility
The AI cross-checks the crawl and prepares the issues. The team validates before the ticket moves.
Operations
The AI queries and summarises task status. Each owner reviews their own.
Source: Ironhack for Business.
SEO is one of the areas changing fastest: 89% of Spanish marketers say AI is transforming their SEO strategy, and 91% have already started optimising for AI-generated answers on platforms like ChatGPT and Google AI Overview (Salesforce, State of Marketing 2026).
How we use AI in Ironhack's marketing
In our team, AI applied to digital marketing is not a separate tool: it is connected to the tools we use every day, in many cases through MCP, the protocol that lets AI work with other applications. These are the flows we have running.
Creative and video
For creative work we use Figma connected to Claude, and combine ChatGPT and Claude depending on the task. The AI proposes variants and adaptations from the brief, and the design team decides and adjusts in Figma. For video, Claude prepares the script and the direction, Higgsfield generates the pieces, and the team checks brand and message before publishing.
Data and campaigns
Every week, Claude connected to our results prepares the review dashboards, and the team uses them to decide in the meeting. Metabase is connected to Claude over MCP, with an internal skill for querying the data, and our Meta Ads accounts send alerts to Slack when something falls outside the normal range.
Content and technical SEO
We update the content of DatoCMS, our content management system, from Claude using files that map each CMS field, so every change lands in the right one. On technical SEO, Claude is connected to ClickUp and Screaming Frog: it cross-checks the crawl against the development tickets and prepares the review, and the SEO team validates it before the ticket moves.
Process
Claude is also connected to ClickUp for the team's process management: it queries task status and prepares summaries, and each owner reviews what affects their work.
The pattern repeats across every flow: the AI prepares and watches, and the team decides.
Why connect AI to your marketing data?
Connecting AI to your marketing data is what turns it from a writer into support for deciding: without data it can only write, and with data it can watch campaigns, prepare reports and flag what changes. That access is the bottleneck today: in Spain, only 55% of marketers have full access to service data, 51% to sales data and 45% to commerce data (Salesforce, State of Marketing 2026).
The missing context shows with customers: 74% struggle to respond quickly because they cannot reach the context they need, according to the same report.
The prudent way to start any AI flow in marketing is to connect a single source with read-only permissions. The AI can query and summarise, but not change campaigns or data, and the permissions widen only once the flow has proved itself.
When AI starts acting on that data, apply the rules for AI agents.
What AI should you use for marketing?
The best AI for marketing is the one that connects to your tools (design, CMS, ad platforms, analytics and project management) and your team knows how to use. Claude, ChatGPT or Microsoft Copilot can all do the job: the difference is usually more about the connections and the daily use than about the model.
Before buying another tool, check which licences you already pay for and what is being used without approval. In our AI consulting we start with an audit of the tools circulating in the company, shadow AI included.
We compare the two most used assistants in Claude vs ChatGPT. For landing pages and internal tools without coding, see our guide to Lovable. And if marketing shares processes with the sales team, read AI for sales.
How do you train a marketing team in AI?
A marketing team learns AI by practising on its own flows and tools, with clear rules of use. The need is real: in Spain, 45% of marketers do not know how to adapt their strategies to the widespread use of AI (Salesforce, State of Marketing 2026).
Our hands-on workshops on productivity can be run per department and work on the tools your company already has (Claude, Copilot or ChatGPT), with the team's real use cases. Every team leaves with prompt templates and a repeatable demo agent built on its own case.
The training comes with governance: a use policy, a list of approved tools and a record of use. And it is measured by tasks, not attendance: the hours each flow takes before the training and 60 to 90 days later. We set the method out in the guide on the ROI of training.
Where to start with AI in marketing: 4 steps
- Pick a repetitive flow. The weekly results review or preparing creative variants are good candidates. Measure how long it takes today.
- Connect a single data source. Start with read-only permissions and widen them once the flow works.
- Decide who reviews, and train the team. Every AI result has an owner who validates it before anything is published or acted on.
- Measure before you scale. Compare the hours the flow took before and 60 to 90 days later. For a first indicative estimate, use the savings calculator.
When a procedure works, save it as a Claude skill so the whole team repeats it the same way.
Next step
If you want to know which flow in your marketing team is the best candidate to start with, we will come back with a starting point in under 1 working day.