The ActivationGrid method / Research standards
Real-world activity is messy. Announcements are partial. Roles overlap. Good intelligence needs a method that tells you what the evidence establishes, what remains unknown and why a record was published.
Discover
Identify the source
Structure
Keep roles distinct
Verify
Check each claim
Human review
Record the decision
Publish
Apply publication rules
Maintain
Review new evidence
The central rule
A plausible claim is not the same thing as a supported claim.
01 / The process
Each stage has a different job. Discovery creates a candidate. Evidence supports specific facts. Human review makes a judgement. Publication is a separate, controlled action.
We look for real-world activations, fairs, pop-ups, installations and experiences documented by public sources. A lead is a starting point, not a verified record.
At this stage
Potential activity with an identifiable source
We organise supported details such as the title, date, location, format, participation mechanics and relevant organisations. A brand, venue, organiser and sponsor are not interchangeable roles.
At this stage
Structured claims with traceable context
Key factual fields are checked against recorded sources. Unclear dates, relationships or roles remain unconfirmed rather than being completed from inference.
At this stage
Verified fields and explicitly unresolved gaps
Human review and the publication gate assess record readiness, source quality and exceptional cases. Automated extraction alone cannot approve or publish a record.
At this stage
A review decision and an audit trail
Only records that pass the applicable publication conditions enter the public dataset. Evidence-linked context becomes available to explore through records, research views and connected tools.
At this stage
A published record with its supporting context
Markets change and mistakes can happen. Corrections can be requested, claims can be reconsidered and the documented history helps distinguish later changes from the original observation.
At this stage
A dataset that can be challenged and improved
02 / Confidence and limits
It is a statement about the recorded evidence and applicable review checks, not a promise that we saw the experience ourselves or measured its business impact.
Coverage is selective and expanding from the UAE. The absence of a record is not evidence that an activation did not occur.
03 / A real example
Editions, the 2025 art and design fair at Dubai Design District, offered a useful test of this principle. The venue listing supported the event details, while a separate official source identified Art Dubai Group as the organiser.
Our human-approved review treated the fair as an independently organised public exhibition. It did not invent a consumer brand or claim that unidentified sponsors did not exist.
Explore the published Editions record ↗Case note / Editions 2025
Separate organiser evidence
Editions official organiser page ↗Art Dubai Group identified in the reviewed evidence.
Publication decision
Approved with a documented scope exception
Unconfirmed sponsors remained unconfirmed.
04 / The principles
If published evidence does not identify a commercial brand or a participant's role, we do not assign one simply to fill a field.
AI may assist with discovery and organisation. Human decisions and rule-based gates control verification and publication.
Missing evidence, unconfirmed parties and uneven geographic coverage should not disappear behind a polished interface.
A public intelligence record is a documented account, not an untouchable statement. Challenges and new evidence matter.
05 / Accountability
Corrections are welcome. Tell us which claim needs attention and share the relevant supporting source so it can be reviewed.
