Run every client from one place. Let clients approve in your portal.

No. Each client is a separate workspace, and access is enforced in the database for every request, not only in the interface.

Yes, as client approvers or viewers in your branded portal. They do not see your other clients.

Each automation slot includes the AI cost of its own runs within fair use, including up to 4 (hourly), 2 (12 h) or 1 (24 h) Trend Radar concepts a day. Extra concepts, final creative renders, posts to X, chat and on-demand actions use credits. On agency plans you can cap credits per client as well as per agency.

You can bill clients for your service and for the portal. Reselling 1ads accounts as a separate product needs a written agreement.

Keep using them. Advantage+, Performance Max and AI Max are free, they use signals no outside tool can see, and 1ads does not bid instead of them. What they do not do is judge themselves: each platform optimises and reports on its own attribution, inside its own walls. 1ads sits above them. It joins ad data with your store and analytics, judges every channel on your targets and, where your store data allows, on profit after COGS, refunds and refused COD parcels, flags when tracking breaks or delivery stops, proposes moving budget between channels, and helps you test whether a channel really adds sales (for example with a brand-search holdout). If you advertise on one channel, your tracking is clean and you are happy with the results, you probably do not need us. Our own model says a shop spending about €2k a month on Meta alone does not pay back.

No. The AI can read, analyse, draft and propose. It has no permission to apply anything. A change runs only after an authorised person approves it. In the app, an approval needs multi-factor authentication within the last 15 minutes. From email or Slack, the approver confirms with a one-time code; approval links expire and work only once. During the closed access program there is no auto-approval at all, with no exceptions in any contract. Before an approved change runs, 1ads re-reads the live campaign; if something changed since the proposal, it goes back for approval. Hard limits, such as how much a budget may change in a day, apply even to approved changes. The trade-off is speed: at 3 a.m. a person has to approve, so tools with autonomous rules react faster.

Two different things carry that name.

Your ads' return inside 1ads: we start from profit, not revenue. Profit on ad spend (POAS) is net revenue minus COGS, shipping, payment fees, refunds and refused COD parcels, divided by ad spend. We never add up conversions reported by different platforms. Each channel then gets an incrementality factor from tests and cited research, and every result shows a range and a confidence grade from A (a recent controlled test) to D (estimates only).

The value of paying for 1ads: the calculator adds what you can verify first (tools you cancel, then net hours saved × your loaded hourly cost), then problems caught earlier and a cautious range for ad efficiency, and subtracts the plan, add-ons, AI credits, the hours you spend setting it up and the cost of any holdout. It shows the net value and payback first, then value-to-cost, for the conservative and expected cases, with and without ad improvement.

Three reasons. First, platform approvals: Google, Meta, LinkedIn and TikTok review apps like ours before granting full access, so channels arrive progressively, and we would rather onboard you when your channels work. Second, we want to measure results with every customer (hours, tools cancelled, one guided test) instead of guessing them. Third, contracts and invoices are manual during the closed access program: an Order Form you sign, invoices paid by bank transfer. So a person reviews each request, usually with a short call, and we may say "not yet" if we are not a good fit for you.

Your account and customer data are stored in the EU, on our own on-premises Kubernetes cluster in a data centre in Romania (not a public cloud). We run the cluster ourselves, under our FirstIT brand. To answer a chat request or run an analysis, the excerpts needed for that task are sent to our AI provider, Anthropic, which processes them in the United States under EU-approved transfer safeguards (the mechanism is stated for each provider in the Sub-processors list). Images are generated by OpenAI, also in the United States, and transactional email is sent through Amazon SES (Amazon Web Services EMEA SARL, EU region). Under their commercial terms, neither AI provider trains its models on this data. Until zero data retention is confirmed, Anthropic deletes API inputs and outputs within 30 days, except content flagged by its safety systems (kept up to 2 years) and where the law requires; OpenAI may keep abuse-monitoring logs for up to 30 days. Tokens for the platforms we connect directly are encrypted in our vault and never sent to an AI model; channels connected through data partners are authorised with that partner. Each customer's data is kept separate. The sub-processor list, with locations, is public, and we give 30 days' notice before adding one. An EU-only AI processing option is being evaluated; it is not available yet.