You measure EU AI Act readiness by tracking a few concrete things: how many AI systems your team uses and whether each is inventoried and risk-classified, how much AI-generated content carries the required disclosure, how many staff have completed AI literacy training, and how quickly issues are resolved. A dashboard of these indicators shows leadership progress and gives regulators and clients evidence.
The EU AI Act, formally Regulation (EU) 2024/1689, is the European Union’s law on artificial intelligence. It entered into force in August 2024 and applies in stages. It sorts AI uses by risk: some practices are prohibited, high-risk systems face strict requirements, certain systems have transparency duties, and general-purpose AI model providers have their own obligations. It also requires organisations that deploy AI to take measures for adequate AI literacy among their staff. This article is not legal advice; for your specific obligations, work with a qualified lawyer.
Why should marketing and PR teams care?
Most marketing uses of AI, such as drafting copy or generating images, are not high-risk under the Act. But marketing is where transparency obligations show up most visibly. Chatbots must tell people they are talking to an AI, and deepfakes and certain AI-generated content must be disclosed as such, with the details set out in the Act and in guidance from the European Commission.
There is a reputational side too. Journalists, clients and job candidates increasingly ask companies how they use AI. A PR team that can answer with facts, backed by an inventory and a policy, is in a stronger position than one that improvises.
Which KPIs show real readiness?
Good indicators are simple, countable and owned by someone. A starting set for a marketing or communications department:
- AI inventory coverage: share of AI tools in use that are recorded with owner, purpose, data used and risk category.
- Risk classification completion: share of inventoried systems reviewed by legal or compliance.
- Disclosure rate: share of published AI-generated or AI-manipulated content that carries the agreed label, checked by sample.
- Chatbot transparency: every customer-facing bot tells users they are interacting with an AI, checked monthly.
- AI literacy: share of staff using AI tools who have completed internal training in the last 12 months.
- Vendor documentation: share of AI vendors that have provided compliance information or contract clauses.
- Incident handling: number of AI-related issues reported and median time to resolve.
How do you build the AI inventory?
Start with a survey and a spending report. Ask every team which AI tools they use, including free ones and features inside existing software, such as AI in Canva, HubSpot or Microsoft 365. Cross-check against finance records for subscriptions.
Record each system in a shared table: name, vendor, owner, purpose, data inputs, whether output reaches the public, and a first risk guess. Shadow use is common, so repeat the survey every quarter. The inventory is the foundation for every other KPI; if it is incomplete, the other numbers look better than reality.
What should the dashboard look like?
Keep it to one page for leadership, with drill-down for the compliance team. A Looker Studio or Power BI report connected to the inventory sheet and training records works for most mid-size companies.
Suggested layout
- Top row: four headline percentages, namely inventory coverage, classification completion, disclosure rate and literacy completion.
- Middle: a table of systems by risk category and owner, with open actions.
- Bottom: trend lines per quarter and a list of open incidents.
Add a dated note on what changed in the law or guidance since the last review. The Act phases in over several years, and codes of practice and guidelines continue to appear.
How do you check disclosure in practice?
Write a short labelling rule first: which content gets a label, what wording, and where it appears. For example, realistic generated images of people or events carry a visible caption, and AI-assisted text in press materials follows the company’s stated policy.
Then sample. Each month, pick a fixed number of published items, such as 20 social posts and five press assets, and check them against the rule. The disclosure rate is the share that complied. Sampling is cheap and honest; self-reporting tends to overstate compliance.
Does measuring readiness create business value?
It does, in three ways. It reduces the chance of fines and platform enforcement. It shortens client and procurement questionnaires, because the answers already exist. And it often reveals duplicate subscriptions and unused tools, which saves money.
Track those benefits too: time to answer an AI questionnaire from a client, number of tools retired after inventory, and media enquiries answered with documented facts.
Who should own the readiness KPIs?
Give each indicator a named owner. The marketing operations lead usually owns the inventory, the content lead owns disclosure sampling, HR or learning owns literacy training, and legal owns risk classification.
Review them together each quarter in a short meeting with a fixed agenda: changes in the inventory, sampling results, training gaps and any new guidance. Ownership turns a compliance project into routine work, which is what regulators expect to see.
How Alien Road applies EU AI Act readiness work
In a typical PR or marketing engagement, we start by building the AI inventory for the marketing and communications function, using a survey and a review of tools in active campaigns. We then help draft a labelling rule and a short internal AI use policy for legal review by the client’s counsel.
Our ongoing role is measurement and communication: monthly disclosure sampling across the content we produce, a shared readiness dashboard, and prepared answers for media and client questions about AI use. We do not give legal opinions; we make sure the facts are organised so the client’s lawyers and leadership can act on them.
Where to start this week
Send a five-question survey to your marketing and communications team asking which AI tools they used in the last month and for what. Put the answers into a shared sheet with an owner for each tool.
Then check your website chatbot, if you have one. Does it say clearly, at the start, that the user is talking to an AI? If not, that is a quick fix you can make before any dashboard exists.
Keep reading: PR Services · Digital PR