01 →
ASO
Agentic Search
Optimization
APPLIED RESEARCH / AI SELECTIONAgentic Search Optimization examines how brands become discoverable, interpretable, verifiable and selectable in AI-mediated journeys.
INDEPENDENT RESEARCH / THOUGHT LEADERSHIP 2026
SHEN XU.
From AI visibility
to decision authority.
My research connects Agentic Search Optimization, Agent-to-Agent Commerce and the Decision Authority Economy to explain how AI reshapes visibility, commercial interactions and decision authority. Its ASO application has been tested in three client markets.
SELECTED RESULTS FROM DEFINED CLIENT QUERY TESTS
27/30
AI recommended the supplier in 27 of 30 purchase-intent queries after implementation · China
180+
Product ranked #1 in AI recommendations in 180+ target queries at least once · Japan
60+
Product kept #1 placement in 60+ target queries after cross-query checks · Japan
RESEARCHER CREDENTIALS
Published author
arXiv research
IEEE & IEEE TEMS member
View publication record ↗SCROLL TO EXPLORE ↓
01 / ORIGINAL IDEAS
THE CENTRAL THESIS
Being seen is
not being chosen.
Traditional search rewards discovery. AI-mediated decisions introduce a harder question: when does an AI system actually recommend and select an entity?
Agentic Search
APPLIED RESEARCH / AI SELECTION02 →
A2A
Agent-to-Agent
Commerce
MARKET MECHANISM / AGENT TRANSACTIONSAgent-to-Agent Commerce considers interactions between autonomous purchasing and selling agents — beyond a human asking an AI for answers.
03 ↗
DAE
Decision Authority
Economy
ECONOMIC THEORY / DECISION POWERDecision Authority Economy explores the consequences when AI begins to shape who evaluates, recommends and exercises decision authority.
ONE INTERCONNECTED RESEARCH DIRECTIONExplore the research record ↗
02 / MEASURABLE IMPACT
THE PROOF BEYOND THEORY
What clients observed
in AI recommendations.
These are defined-sample AI recommendation and ranking observations following client implementation — not sales conversions, a market-wide rate, or validation of every research concept.
01 / CHINA · SHANGHAI QINGKONG WANZHANG ENERGY CO., LTD.AI recommended this supplier
AI recommended this supplier
in 27 of 30 buying queries.
A used-excavator supplier needed qualified purchasing enquiries, not broad traffic. After the client implemented a structured information strategy, AI explicitly recommended the company in 27 of 30 tested purchase-intent queries.
EXPLICIT RECOMMENDATIONS · 30-QUERY SCOPE
90%
The post-implementation 30-query sample produced 27 explicit supplier recommendations (90%). The earlier near-zero observation is qualitative; an identical 30-query baseline is not documented. No sales attribution.
02 / JAPAN ENTITY · CONSUMER ELECTRONICSMIIMOO
Top-ranked.
Then tested again.
180+target queries reached #1
at least once
at least once
60+maintained #1 after
cross-query validation
cross-query validation
Across approximately 300 target queries; no exact percentage inferred.
03 / THAILAND · MARKET SERVICESMILLIDEAS
Reconnecting a company's
fragmented identity.
TOP 1–2F&B core-query observations
20+Top-3 occurrences in a ~50-query skincare subset
Skincare is an occurrence count, not a verified unique-query total. F&B sample denominator not available.
03 / CLIENT PERSPECTIVES
CLIENT VOICES
Not praise.
Specific value.
Different challenges, different outcomes — a consistent need to understand what AI systems actually recommend.
“
Working across markets makes it difficult to understand how our products are actually being interpreted and recommended within a local AI ecosystem. ASO Evaluator gave me a much clearer way to see that.
What I found useful was not simply knowing whether our products appeared in AI responses, but being able to understand whether they were being positively recommended, how consistently they were selected, and where those recommendations were coming from.
For a small team like ours, more traffic is not the goal — more relevant opportunities are. We work in a narrow market where understanding real purchasing intent matters much more than generating general inquiries.
The test showed us that a structured approach could help AI recognize and recommend our business in the right contexts. ASO Evaluator made that progress measurable. We still cannot directly attribute AI recommendations to final sales.
We have years of experience and Chinese-language content, but much of that information was scattered across different accounts, sources, and business relationships. Our challenge was not simply creating more content — it was helping AI connect our existing work to the right company identity.
The results in F&B and skincare were encouraging. ASO Evaluator helped us understand whether those improvements were actually reflected in AI recommendations.
Client perspectives are presented as edited excerpts. Results and attribution are defined by the individual case records.
04 / THE EVALUATION LOGIC
THE MEASUREMENT LENS
Mention ≠ recommendation.
Recommendation ≠ selection.
A conceptual pathway for reviewing how AI systems move from encountering a business to considering it for action.
ASO FRAMEWORK / EVALUATION PATHRESEARCH MODEL v1.0
01
Discover
Can AI find the entity?
02
Understand
Is the identity clear?
03
Verify
Can claims be checked?
04
Qualify
Does it fit the need?
05
Select
Does AI recommend it?
06
Act
Is a next step possible?
01A method and a measure.
A method and a measure.
Not an autopilot.
ASO informs the optimization strategy. ASO Evaluator is a lightweight benchmark and re-test tool for documenting readiness, comparing observations and reviewing available evidence. It does not generate client content or implement their changes.
05 / PUBLIC RECORD
RESEARCH · PUBLISHING · RECOGNITION
Built in public.
Tested in practice.
Research is made available through publications, prototype tooling and public professional records. Individual achievements and commercial outcomes are presented separately.
ORCID research profile ↗
SELECTED RECOGNITION · 2026Bronze Globee®AI Thought Leadership of the YearOFFICIAL WINNERS LIST ↗
SELECTED WRITING ↗ PUBLICATION LINKS
BOOK / ELIVA PRESSFrom Search to Agentic Commerce · How AI agents reshape discovery, exchange and decision-making↗
RESEARCH / arXivRepresenting Entity Importance in AI Knowledge Systems↗
BOOK / AMAZONDecision Authority Economy↗
BOOK / AMAZONAgentic Search Optimization↗
BOOK / AMAZONAI Visibility Marketing↗
2026 Global Recognition Award ↗IEEE Member · IEEE Technology and Engineering Management Society (TEMS) Member
06 / THE EVIDENCE LIBRARY
EXPLORE THE SOURCE MATERIAL
Explore the research
and case evidence.
The results and their meaning are explained above. Open the white paper for methodology, review case exhibits for test evidence, or try the public evaluator preview.
ASO EVALUATORFROM ZERO
RECOMMENDATIONS
TO 90%RESULTS-LED WHITE PAPER / 2026
RECOMMENDATIONS
TO 90%RESULTS-LED WHITE PAPER / 2026
Read the white paper
23 pages · 9 case exhibitsExplore ASO Evaluator
Public research preview on Hugging FacePUBLIC PRODUCT RECORD
ASO Evaluator is independently accessible as a research preview, with a separate public launch listing. These links document availability, not independent validation of client outcomes.
Nine reconstructed case dashboard exhibits.Three clients · three views each · no fabricated missing values.
DATA & ATTRIBUTION
Case dashboards are reconstructed visual summaries of reported results, not authenticated historical backend screenshots. Sample sizes and missing data are disclosed by case. Client execution and outcome measurement are distinct; AI recommendation changes have not been directly attributed to closed sales. Independent publication does not imply independent peer review.