An advertising agency built on AI
We are an agency too, but we do most of the production with AI. That is why we can also take on work that falls outside the scope of a traditional agency.
- Principles
- 3
- Steps in how we work
- 5
- Disciplines
- 6
- Days to permanent deletion
- 30
Three principles
Every decision is made by these three sentences. Each comes with what it means in day-to-day work.
-
People have the final say
Every step that spends money, goes live or changes your brand voice comes to you for approval. You set the thresholds.
- Any budget change above the spend threshold is put to you.
- Everything that goes live passes through our team first, then you.
- For the first two weeks every output passes through you; after that you loosen the threshold.
-
Every decision can be traced
We record which work was done on which data. You can look back without waiting for the month-end report.
- Every decision, input and output is stored in a way that cannot be altered.
- In advertising, the reason for every change is written in the decision log.
- Access is role-based, and every access is logged.
-
Leaving should be easy
You can stop any job with a single switch. Every asset we produce is yours, and you take all of it with you when you leave.
- There is no commitment on the Starter and Growth plans.
- You can export all assets and the brand memory.
- On request, records are permanently deleted within 30 days.
AI does, the team supervises, you decide
AI agents carry out most of the production. But no job goes live on its own.
AI does
- Turns a brief into a plan and a draft
- Produces copy and visual variations
- Runs ad tests and the publishing order
- Collects channel data and catches deviations
Our team supervises
- Teaches the system your brand and applies corrections
- Reviews every action over the threshold first
- Handles errors and exceptions
- Sends you the report with written commentary
You decide
- Set the approval thresholds
- Approve spend, publishing and changes of voice
- Stop anything with a single switch if you wish
- Take everything with you when you leave
Steps that spend money or go live ask for human approval by default: first from our team, then from you. For work that runs without approval, every action can be reversed and is traced with its reason.
How we work in five steps
From the first call to the first report, you know from the start what will happen.
Introduction
30 minutes, free
We listen to your business, your goal and how your marketing works today. This call is not a sales pitch, it is a situation check.
Scope
In writing
We give you in writing which items we will take on, the delivery schedule and the monthly amount.
Brand training
For the first two weeks every output passes through you
We teach both our team and our system your brand voice, your visual language and your past campaigns. Your corrections are learned for good.
Production
Six disciplines in parallel
Work that needs approval comes to you, and the rest moves ahead without waiting.
Measurement and correction
Weekly, monthly or daily
We bring the results together in one table and say plainly what worked. The next period's plan is built on this data.
The system that runs alongside the team
Not one big model, but agents that each specialise in a narrow task and hand work to one another.
Agent runner
Queues tasks, resolves dependencies and rolls back on failure.
What it means for youIf one step fails, the job is not left half done; it is rolled back.
Brand memory
Tone, glossary, visual rules and past campaigns are kept in one place.
What it means for youYou make a correction once and it applies on every channel.
Integration layer
Two-way connections with ad accounts, CMS, CRM and social channels.
What it means for youWork runs on your own accounts.
Approval engine
Stops any action over the defined thresholds and routes it first to our team, then to you.
What it means for youA spend over the threshold never happens on its own.
Measurement pipeline
Turns channel data into shared metrics and catches deviations.
What it means for youYou hear about an unexpected swing without waiting for the report.
Audit log
Every decision, input and output is stored in a way that cannot be altered.
What it means for youYou can see the reason for a decision afterwards too.
What happens to your data
Data takes a different path in paid agency work and in the free panel. We set the two side by side.
Paid agency work
- Model training
- Client data does not go into general model training.
- Your brand memory
- Used only in your own work.
- Where it is stored
- Servers inside the European Union. For enterprise agreements it can also stay in Türkiye.
- Access
- Viewer, approver and admin roles; every access is logged.
Free brand identity panel
- Model training
- The brand details you enter are sent to Google's AI service. Google may use this content to improve its own services.
- Detail
- The KVKK privacy notice (in Turkish) sets out which data is processed and how to have it deleted.
Is my data used for model training?
In paid agency work, no: client data does not go into general model training, and your brand memory is used only in your own work. The free brand identity panel is different: the brand details you enter are sent to Google's AI service, and Google may use this content to improve its own services. The details are in the KVKK privacy notice (in Turkish).
Where is data stored?
On servers inside the European Union. For enterprise agreements we can also keep the data in Türkiye.
Who can access what?
Access is role-based: viewer, approver and admin roles are kept separate. Every access is written to the audit log.
What happens to my data if I leave?
Everything we produce is yours. You export all assets and the brand memory, and on request records are permanently deleted within 30 days.
What if the AI gets something wrong?
Steps that spend money or go live ask for human approval by default: first from our team, then from you. For work that runs without approval, every action can be reversed and is traced with its reason.