ROESER · AI Visibility Operations

From industrial proof to AI recommendation.

A private evidence view for ROESER's 90-day programme. Observed public signals are shown now; model-specific scores remain unassigned until the frozen 30-prompt panel has been run directly and archived.

Review the baseline See the 90-day plan

Evidence dated 17 July 2026 · no invented scores · human approval before every external action.

3D AI core sphere routing company processes into an automated digital workforce

Observed baseline

Strong company substance. Weak category discovery.

These figures are observable starting signals, not an AI Visibility Index score. They establish where the controlled measurement begins.

Ranking keywords

Germany snapshot · 17 Jul 2026

373

Estimated monthly organic visits

Directional third-party estimate

147

Referring domains

230 total backlinks observed

Relevant ROESER AI mentions

In the sampled public category datasets

Direct ChatGPT, Claude, Gemini, Perplexity and Google AI tests are not represented as completed. Authenticated testing access is not connected in this workspace.

Five-dimensional index

Every score needs a denominator and evidence.

The dashboard refuses false precision. Each dimension stays pending until the relevant answer set, sources and assessment rules are complete.

AI Understanding

Pending direct run

Are ROESER's category, capabilities, scale, markets and differentiators represented correctly?

9 approved fact checkpoints × branded prompt panel

AI Citation

Pending direct run

Do relevant answers cite accurate ROESER-controlled or authoritative independent evidence?

Valid citations ÷ relevant answers tested

AI Recommendation

Pending direct run

Is ROESER included in eligible buyer recommendations with a defensible rationale?

Eligible prompts recommending ROESER ÷ eligible prompts

AI Authority

Observed foundation

How strong, diverse and consistent is the independent evidence supporting ROESER?

Source diversity, relevance, consistency and provenance

AI Readiness

Audit pending

Can AI crawlers discover, parse and connect ROESER's approved evidence across DE and EN?

Technical, structural, language and evidence checks

Evidence register

The baseline is a traceable record, not a slide.

Every observation carries a source class and evidence state. Public claims remain blocked until the underlying fact and customer approval are present.

Signal	Observation	Source	State
Official company scale	~400 employees; 400M+ decorations/year	ROESER company page	Verified first-party
Organic discovery	23 ranking keywords; branded terms dominate	DataForSEO DE snapshot	Observed
Category search	Absent from first 20 for tested DE category query	Google SERP dataset	Observed
Authority base	230 backlinks from 147 referring domains	DataForSEO backlink snapshot	Observed
Entity consistency	External scale data conflicts with official facts	Cross-source comparison	Action required
AI recommendation presence	No relevant mention in sampled public datasets	Seven category/brand datasets	Provisional

Frozen prompt architecture

A · Branded understanding

6 prompts

What does ROESER do? · Which materials and markets? · Can it support production scale?

Frozen · direct evidence pending

B · Category discovery

8 prompts

Packaging-decoration specialists · combined technologies · glass finishing

Frozen · public sample observed

C · Buyer recommendation

8 prompts

Who should a prestige cosmetics buyer consider? · Who combines coating, print, stamp and laser?

Frozen · direct evidence pending

D · Capability evidence

8 prompts

Process selection · quality controls · sustainability · partner criteria

Frozen · direct evidence pending

90-day work package

Measure, build, distribute and prove.

One programme, four overlapping phases and a weekly approval rhythm. ROESER buys defined outcomes and finished work—not a software licence or unlimited retainer.

Weeks 1–2

Measure & ground

Run the fixed panel directly
Approve company facts and sources
Reconcile entity conflicts
Archive every answer and citation

Weeks 2–4

Decide & architect

Prioritise buyer questions
Define Complete Decoration clearly
Approve evidence backlog
Set acceptance and stop rules

Weeks 4–10

Build & distribute

Improve answer-ready evidence
Align German and English
Correct approved profiles
Publish useful authority assets

Weeks 6–12

Operate & prove

Measure weekly
Record answer changes
Close evidence gaps
Deliver before/after review

Weekly operating log

What changed—and what did not.

The log separates completed work, model movement and evidence gaps so reporting cannot imply causation that has not been established.

17 Jul 2026

Public-data baseline

Completed

Search, backlink and sampled AI-answer observations archived.

Direct multi-model run

Blocked

Requires approved test conditions and connected model access.

Entity correction

Ready

Reconcile official company scale and external profile data.

Evidence backlog

Ready

Prioritise answer-ready proof across DE and EN.

HEART guardrails

Human-directed recommendation operations.

Agents research, compare, draft and measure. People retain responsibility for facts, publication, outreach, partnerships, spending and customer commitments.

Human benefit

Prioritise buyer and customer usefulness.

Evidence

Separate verified facts from hypotheses.

Approval

No consequential external action without sign-off.

Respect

No impersonation, manipulation or fabricated familiarity.

Transparent intent

Disclose commercial purpose and limitations.

Ready to start the controlled ROESER baseline?

Approve the prompt panel, evidence rules and competitor set. We will then replace every pending state with directly captured, reproducible evidence.

Approve programme scope thorsten@thorstenkeller.ai