ATIYO Brain · Bring your own API key
An AI creative strategist should know your evidence—or say it does not.
The ATIYO Brain is a read-only strategist grounded in the workspace's recorded brand context, briefs, outcomes, and learnings. An administrator connects an OpenRouter, fal.ai, or Kie.ai key and controls provider spend.
Direct answer
What is a grounded AI creative strategist?
A grounded AI creative strategist answers from a defined body of brand evidence instead of treating a generic model response as a customer fact. In ATIYO, the Brain reads the workspace's Brand Bible, briefs, recorded winners, failures, outcomes, and learnings; it cites the records behind an answer and should state when evidence is missing. It is read-only by design: it can advise, compare, summarize, and suggest, but it does not change the workspace, buy media, or create performance data.
01
Why a generic chatbot is not the brand record
A general model can help with language and pattern generation, but it does not automatically know which claims are approved, which angle failed, what a prior brief contained, or how your team defined a winner.
Copying a few notes into a prompt creates a temporary slice of context. A grounded workflow keeps the durable records—audiences, desires, angles, briefs, outcomes, and learnings—connected to the work. The model can then point back to those records instead of asking the team to trust an unsupported summary.
02
What the ATIYO Brain reads
The Brain uses the current workspace's Brand Bible and recorded creative evidence. Tenant boundaries remain intact: one workspace's content is not used as another workspace's context.
| Context | Useful question | Important limitation |
|---|---|---|
| Brand Bible | Which audience desire and angle fit this brief? | Only as accurate as the saved direction |
| Briefs | Have we already made a similar concept? | A missing brief cannot be cited |
| Recorded winners | Which supported patterns recur? | Winner labels need a real team decision |
| Failures and outcomes | What should we avoid repeating? | Correlation is not automatic causation |
| Learnings | What should the next brief carry forward? | Human judgment still owns the action |
03
Questions that fit a grounded strategist
Ask for comparisons, gaps, patterns, and brief implications that can be traced to saved evidence. The quality of the answer depends on the quality and completeness of the records.
- Which recorded winners address this avatar, and what angles do they use?
- Have we tested this objection before, and what did the team record?
- Which briefs lack a clear proof mechanism or call to action?
- What supported learning should shape the next three concepts?
- Where is the evidence too thin to recommend a direction?
04
Bring your own model access
An ATIYO workspace administrator supplies an OpenRouter, fal.ai, or Kie.ai key. The key is not included in ATIYO's subscription, so the administrator controls provider choice and spend.
OpenRouter documents an API that routes requests to supported models. Model availability, capabilities, pricing, and provider terms can change, so administrators should review current OpenRouter and model-provider documentation before choosing a model.
Keep API keys out of briefs, comments, exports, and shared files. Only an authorized workspace administrator should configure the key through the product setting intended for it.
05
The human remains the decision owner
The Brain is read-only and does not approve claims, change statuses, launch ads, or interpret private performance systems it cannot see. A responsible teammate must review the cited evidence and choose the action.
This boundary is a product feature, not a missing automation. Creative strategy contains commercial, legal, brand, and measurement judgments. Advice can be fast; responsibility should remain explicit.
06
How to make the Brain more useful in seven days
Do not begin by asking for more output. Begin by improving the workspace evidence the strategist is allowed to use.
- Complete the Brand Bible with current audiences, desires, and angles.
- Add the briefs the team is actively producing.
- Record real outcomes and the source used to make each decision.
- Write concise learnings with caveats when the evidence is limited.
- Ask the Brain to cite the records behind one upcoming brief decision.
Frequently asked questions
Questions about this workflow
Does ATIYO include AI model usage?
No. The Brain is included in ATIYO plans, but an administrator brings an OpenRouter, fal.ai, or Kie.ai key and controls provider spend.
Can the Brain change workspace records?
No. It is read-only by design. It can advise from available workspace evidence but does not change records or take actions.
Will the Brain invent a result when data is missing?
It is designed to cite the workspace records behind its answer and to state when there is not enough evidence. Teams should still review every answer and keep the underlying records accurate.
Primary and official sources
Sources used in this guide
External product facts were checked against the organizations’ own documentation. Features can change; confirm current details before making a purchase or campaign decision.
- OpenRouter Quickstart Official OpenRouter API setup and request documentation.
- ATIYO creative operations ATIYO's public description of its connected creative workflow.
- ATIYO company facts Verified product limits, company description, and contact.
Move the plan out of scattered sheets
Run the roadmap, briefs, assets, and learnings in ATIYO.
ATIYO keeps the brand context and production decisions connected. It does not buy media, connect to ad accounts, or invent performance results.