Accountable autonomy, in writing.

What can Marca's AI agents do on their own?

Agents work alone only where the work is internal, reversible and easy to check.

• Research and monitoring: audiences, competitors, news, trends, keywords and mentions.
• First drafts of posts, scripts, articles, releases and reports. A draft is not a decision.
• Platform versions of approved work: sizes, subtitles, formats.
• Scheduling content that has already been approved.
• Routine replies to comments and messages, using wording approved in advance.
• Reporting and data summaries.
• Checking their own work against the brief and brand guidelines you approve, and against platform requirements.

What must a senior Marca person approve?

Anything with consequences goes to a named senior person first.

• Anything public. Every post, video, page and reply that is not pre-approved wording.
• Spend. Any proposal that commits budget, from a boosted post to a creator fee.
• Press. Every media pitch, press release, statement and journalist reply.
• Brand changes. Any change to positioning, tone, visual identity, messaging or strategy.
• Sensitive conversations. Complaints, and anything touching religion, politics, health, money or personal data.

What only the client can approve

• Anything published under your brand or in an executive's name.
• Any spend of your money.
• Any use of a real person's face, voice or likeness.
• Every response in a crisis.

For government and enterprise clients, your own approval chain sits inside the workflow. Work moves to the next approver only when the last one signs.

What is never done, by anyone?

• Publishing content nobody has approved.
• Fake reviews, bought followers or manufactured engagement.
• Creating a likeness of a real person without their explicit written consent.
• Sharing one client's data or material with another.
• Automated replies during a crisis.

The more public, financial, personal or irreversible an action, the more human approval it needs.

How does Marca handle client data?

• Least-privilege access. People and agents get only the access the work needs.
• Your brief, your guidelines. Agents work from the brief and brand guidelines you approve, and a person reviews their work before it reaches you.
• Confidentiality. A mutual NDA is available before you share a brief.
• Incidents. If an incident affects your data, we tell you promptly, within the time set in our contract with you.

Where is data stored?

Website enquiries submitted through marca.ae are hosted on enterprise-grade cloud infrastructure and may be stored outside the UAE, as explained in our Privacy Policy. Only authorised Marca team members and service providers can access them. We keep enquiries for 24 months from the date of the request and then delete them, or sooner if you ask us to.

Client engagement data is handled under the terms agreed with each client. Where we can, your data stays on your own platforms and accounts. If you have data-residency requirements, we'll discuss them on a call.

Who owns the accounts and the work?

You do. Social profiles, ad accounts, analytics, domains and asset libraries stay registered to your organisation. We work inside them with delegated access, never as the owner. Approved content and assets are yours, subject to our agreement with you and any third-party licence terms. Because the accounts are yours, you control our access and can remove it at any time.

Which AI does Marca use?

Our agents draft, research, generate images and video, and check their work against your brief, using a mix of commercial AI models and our own workflows, chosen per task. On this site we describe our providers by category, such as AI model providers, cloud hosting, email and scheduling, rather than by name, because models change and we don't imply endorsements.

The current list of AI providers and sub-processors that would touch your data is available to clients and prospective clients under NDA.

How does Marca disclose AI?

• We say plainly that AI agents do the work and senior people approve it.
• Our chat and SDR agents are designed to identify themselves as AI in their first message and whenever asked.
• The agent characters on our site are illustrations of roles, not people. They carry role codenames, not human names.
• We label AI-generated or AI-altered media wherever the law or the platform requires it, and wherever leaving it unlabelled could mislead.

What are Marca's rules on synthetic media and consent?

Synthetic media means images, video or audio created or altered by AI. It is useful, and it is easy to misuse. Our rules:

1. Verified likeness only. We build a model of a real person's face or voice only after verifying their identity and receiving their explicit written consent for that specific use.
2. Consent has limits. Consent covers a stated purpose, a set of channels and a period of time. It can be withdrawn, and we stop when it is.
3. No unconsented likeness. We never create a synthetic likeness of a government official, public figure or any other real person without their verified, explicit written consent, and for an official, their entity's approval too. This includes satire and concept work.
4. No fake customers. We don't present AI-generated people as real customers, employees or testimonials.
5. Disclosure. Synthetic media is labelled where the law or the platform requires it, and wherever a viewer could otherwise be misled.
6. Creators are checked. Where real creators are involved, we confirm they hold the permits the law requires before anything is published.

Report misuse of any likeness linked to Marca to hello@marca.ae. A senior person reviews every report.

How does Marca approach UAE and Saudi regulation?

• UAE data protection. Marca is licensed in the Sharjah Media City (Shams) free zone, and we process personal data in line with the UAE Personal Data Protection Law (Federal Decree-Law No. 45 of 2021). Where a client is based in the DIFC or ADGM and those free zones' data protection rules apply to its data, we follow them as the client's contract requires.
• Saudi data protection. Saudi Arabia's Personal Data Protection Law, overseen by the Saudi Data and Artificial Intelligence Authority (SDAIA), covers the personal data of people in the Kingdom, not only Saudi clients. Where it applies, we follow it, including its rules on transfers outside the Kingdom. Where a client requires National Cybersecurity Authority controls, we agree in the contract which controls apply and how they are met.
• Government data. Government data follows the client entity's own rules and classification. Theirs take precedence over ours.
• Advertising content. Since 1 February 2026, individuals who publish advertising content, paid or unpaid, on social media or other digital platforms from within the UAE need an Advertiser Permit from the National Media Authority (formerly the UAE Media Council). The permit sits under the Media Law (Federal Decree-Law No. 55 of 2023). People promoting their own business on their own personal account are exempt. Visiting creators can get a permit only through an agency licensed in the UAE and accredited by the National Media Authority. It is valid for three months and can be extended to a maximum of six. Marca does not issue permits or register creators. We ask any creator who will publish or promote a client's advertising for their permit number, and check it before their content runs. In Saudi Arabia, individuals who advertise on social media need a Mawthooq licence from the General Authority for Media Regulation (GAMR), and we check that too. AI-generated ads follow the same content and labelling rules as any other ad. Ads for regulated sectors such as health, real estate and finance also need the sector regulator's approval before they run. Details: nma.gov.ae.

What doesn't Marca do?

• We don't publish anything a person hasn't approved, either piece by piece or as wording approved in advance.
• We don't buy followers, write fake reviews or run engagement pods.
• We don't create anyone's likeness without explicit written consent.
• We don't own your accounts, and we don't lock you in.
• We don't use your material to serve other clients.
• We don't let agents speak to journalists, handle complaints or reply in a crisis.
• We don't promise guaranteed rankings, guaranteed AI citations or guaranteed coverage.
• We don't claim results we can't show.