Shen Xu / 徐慎
I am a product builder, entrepreneur, strategist, and independent researcher with more than 20 years of experience across digital products, technology, marketing, venture building, SaaS, and AI-native systems.
My product work did not begin with generative AI. It spans social and digital platforms, mobile applications, O2O products, consumer software, startup fundraising, enterprise digital strategy, venture-backed technology, SaaS commercialization, AI-native products, and agent-mediated decision systems.
Across these technology cycles, the underlying product questions have remained relatively consistent:
How are products discovered? How do they acquire users? How is trust established? How do they enter a consideration set? Why are they selected? And how does selection become a commercial action?
As AI systems increasingly participate in discovery, evaluation, recommendation, selection, and execution, these questions are extending from human users to AI agents.
Before the mobile and generative-AI eras, my work already focused on digital platforms, social media, online communities, user acquisition, digital distribution, and the way technology changes the relationship between brands, products, and users.
I also published an earlier Chinese-language book, 《社群營銷在中國》, focused on social media, community marketing, digital platforms, and the Chinese Internet environment.
That work examined how social platforms, communities, content, and digital distribution influence:
In retrospect, this represents an earlier stage of a question that continues through my current work:
Who controls discovery, distribution, visibility, recommendation, and ultimately decision influence?
That question later moved from social platforms to search engines, and now to AI systems and agents.
During the early mobile-Internet period, I participated in founding and operating Shanghai Vanilla Technology (上海香草科技).
The company developed mobile products including Banmake (斑马客) and Xiangge (相格).
Banmake was an early mobile and O2O product exploring the relationship between:
The product explored how physical commercial environments could be connected with digital identity, engagement, distribution, and user acquisition.
This work took place before QR-code-driven O2O became a mainstream commercial model in China.
It gave me early experience in designing products around the path:
Physical Environment → Digital Interface → User Interaction → Commercial Conversion
Xiangge was a consumer photography and mobile-application product.
The product expanded across multiple platforms and related products exceeded:
5 million cumulative downloads
Shanghai Vanilla Technology received strategic investment from Qihoo 360.
During this period, the company also received acquisition interest from Baidu.
My work covered product development, growth, distribution, community operations, commercialization, fundraising, and company building.
This period gave me direct experience building and scaling a consumer technology product rather than working only on marketing around an existing product.
I later served as Strategy Director at DigitasLBi Greater China, part of Publicis Groupe.
This stage expanded my perspective from startup product building into enterprise-scale digital strategy.
My work connected areas such as:
The experience provided a different view of the same underlying problem I had encountered as an entrepreneur:
How do people discover, evaluate, interact with, and ultimately choose products and brands in increasingly digital environments?
It also strengthened the connection between product thinking, marketing, technology, and commercial growth.
I later founded and served as CEO of TUNA.
TUNA operated across technology, digital products, marketing technology, creative systems, and enterprise commercialization.
The company received angel-stage investment from Tianqi Amoeba (天奇阿米巴).
This gave me direct experience across:
TUNA connected entrepreneurial product work with larger enterprise projects and commercialization.
Technology-related work connected with this entrepreneurial period also received financial support associated with the Shanghai Municipal Science and Technology Commission.
This added another layer to my product experience:
Product Development → Intellectual Property → Commercialization → Institutional Support
I later worked with Yueliu (阅流), a venture-backed technology company that had progressed through multiple financing rounds.
My work focused on areas including:
During an approximately four-month growth period, the platform reached more than:
300,000 users
The experience complemented my earlier founder-stage work by giving me exposure to growth inside a more mature venture-backed technology company.
I have also provided advisory support to businesses associated with later-stage venture-backed technology companies, including work connected with Zhejiang Xin Zailing Technology (浙江新再灵科技).
The advisory work included areas such as:
Combined with my own startup experience, this gave me exposure to different stages of venture development, from early fundraising to later-stage technology-company growth.
From 2025 to 2026, I worked on China market marketing and growth for TSplus, a French B2B software company focused on remote access, application delivery, and related enterprise-software products.
My work included:
This experience connected my earlier work in product, entrepreneurship, digital strategy, and growth with the commercialization of a mature international software product in China.
It also became one of the practical bridges between my earlier search and visibility work and my later research into AI Visibility and Agentic Search.
Across my founder, operating, growth, and advisory work, I have participated in technology businesses at different stages of company development.
Selected experience includes:
Shanghai Vanilla Technology
Strategic investment from Qihoo 360, alongside the development and scaling of consumer mobile products.
TUNA
Founder and CEO; received angel-stage investment from Tianqi Amoeba.
Yueliu
Operating and growth experience inside a venture-backed technology company that had already progressed through multiple financing rounds.
Later-stage technology advisory
Advisory work involving venture-backed technology businesses, including work connected with Zhejiang Xin Zailing Technology.
My venture experience therefore includes both sides of technology commercialization:
building and financing my own ventures
and
working inside or advising venture-backed technology companies at later stages of development.
My product work has developed alongside writing and research.
The technologies have changed, but many of the underlying questions are connected.
My earlier Chinese-language publishing work examined social media, communities, digital platforms, and digital marketing in China.
That work focused on how digital platforms change:
This represented an earlier phase of my interest in how digital systems mediate decisions.
My later work moved toward how brands, companies, products, people, and other entities are represented inside AI systems.
The central questions include:
This became the foundation of my work on AI Visibility.
Amazon:
https://www.amazon.com/dp/B0H4XBSYQD
I later developed the Decision Authority Economy (DAE) framework.
It examines what changes when people delegate parts of:
Search → Comparison → Recommendation → Selection → Execution
to AI systems.
The central issue is no longer simply information visibility.
It is the redistribution of decision authority.
Amazon:
https://www.amazon.com/dp/B0GY65JZ3H
My work on Agentic Search Optimization (ASO) extends this thinking from AI visibility toward AI-mediated selection.
Traditional search optimization focuses primarily on whether information can rank and be discovered.
Agentic Search introduces a different problem:
An AI system may retrieve many entities but eventually select only a small number of them.
Therefore:
Retrieval is not the same as citation.
Citation is not the same as recommendation.
Recommendation is not the same as selection.
ASO focuses on that transition.
Amazon:
https://www.amazon.com/dp/B0H2Y7MYNR
My 2026 book:
From Search to Agentic Commerce: Attribution Collapse, Agentic Search, Decision Authority, and A2A Markets
extends the same trajectory into commerce.
It examines what happens when AI moves beyond information retrieval and increasingly participates in:
Discovery → Evaluation → Selection → Transaction
The longer-term implication is the emergence of markets in which AI agents act on behalf of people and organizations.
Publisher:
https://elivabooks.com/en/book/book-5876480398
Amazon:
https://www.amazon.com/dp/9999353231
My current product work increasingly translates these research ideas into executable systems, open-source frameworks, interactive tools, plugins, and AI skills.
The Agentic Search Optimization Framework is an open-source framework and interactive evaluation system for assessing whether brands, products, companies, services, and other entities are prepared for AI-agent-mediated discovery and selection.
The framework asks a question that is different from conventional visibility measurement.
Not simply:
Is the entity visible to AI?
But:
Can an AI agent discover, understand, verify, qualify, select, and act on the entity?
The framework is structured around six stages:
Discover → Understand → Verify → Qualify → Select → Act
The current implementation includes:
The framework has been publicly released as an open-source project under the MIT License.
Live Demo:
https://aso-framework.vercel.app/
GitHub:
https://github.com/ShenXuAkaEkstasis/agentic-search-optimization
Release:
ASO Framework v1.0 / v1.0.0
The project converts part of my Agentic Search research into an operational framework that can be assessed, tested, compared, and extended.
The GEO Knowledge Test is an interactive assessment product designed to evaluate practical understanding of:
Rather than testing simple terminology recall, it uses scenario-based questions to examine whether users understand distinctions such as:
Mention vs Citation
Citation vs Recommendation
Visibility vs Selection
The project represents another direction in my product work:
Turning emerging AI-search concepts into interactive evaluation products rather than leaving them only as written theory.
Live Demo:
https://geo-quiz-vercel-v2.vercel.app/
GitHub:
https://github.com/ShenXuAkaEkstasis/geo-quiz-vercel-v2
The A2A Decision Layer is an executable prototype exploring AI-to-AI decision structures.
It examines systems in which AI participates directly in:
Discovery → Evaluation → Comparison → Selection → Transaction
rather than acting only as an information interface.
GitHub:
https://github.com/ShenXuAkaEkstasis/a2a-decision-layer
The AI Decision Mediation Framework explores how AI increasingly mediates human and organizational decisions.
The framework examines the transition from:
Information Retrieval → Recommendation → Evaluation → Selection → Execution
This project is closely connected to my work on the Decision Authority Economy.
GitHub:
https://github.com/ShenXuAkaEkstasis/ai-decision-mediation-framework
The DSH AI Shopping Assistant is a DeepSeek Harness plugin designed for product comparison and purchasing decisions.
Its workflow includes:
The objective is to move beyond simple product search toward structured AI-assisted purchasing decisions.
GitHub:
https://github.com/ShenXuAkaEkstasis/dsh-ai-shopping-assistant
The plugin is also indexed in the public DeepSeek Harness community plugin registry:
https://github.com/dshworks/awesome-dsh-plugins/blob/main/lists/usage-cost.md
The DSH AI SaaS Deal Finder is a DeepSeek Harness plugin designed to identify legitimate lower-cost SaaS and AI-service purchase options.
Instead of simply ranking advertised prices, it evaluates:
The underlying principle is:
The lowest advertised price is not necessarily the lowest usable price.
GitHub:
https://github.com/ShenXuAkaEkstasis/dsh-ai-saas-deal-finder
I have also published a portfolio of AI skills on Tencent SkillHub.
These skills explore how AI can participate in practical decision and production workflows across commerce, business, professional work, entrepreneurship, reputation, and creative production.
AI Shopping Assistant — Product Selection & Price Comparison
AI-assisted product-selection and comparison workflow.
SkillHub:
https://skillhub.cn/skills/user_59867446/ai-shopping-assistant
AI SaaS Deal Finder
AI skill for evaluating lower-cost SaaS and AI-service purchase options.
SkillHub:
https://skillhub.cn/skills/user_59867446/ai-saas-deal-finder
AI Grocery Shopping Assistant
AI-assisted grocery selection and shopping-decision workflow.
SkillHub:
https://skillhub.cn/skills/user_59867446/ai-grocery-shopping-assistant
Store Feasibility Analyzer
Decision-support skill for evaluating the feasibility of opening a store or small business location.
SkillHub:
https://skillhub.cn/skills/user_59867446/store-feasibility-analyzer
Upwork Job Evaluator
Decision-support skill for evaluating whether a freelance opportunity is worth bidding on.
SkillHub:
https://skillhub.cn/skills/user_59867446/upwork-job-evaluator
AI Startup Fundraising BP / Pitch Deck Assistant
AI skill for structuring and evaluating early-stage startup fundraising materials and pitch decks.
SkillHub:
https://skillhub.cn/skills/user_59867446/ai-startup-fundraising-bp-pitch-deck
Reputation Strategy Architect
AI-assisted strategy workflow for personal-brand, corporate-reputation, and influence growth planning.
SkillHub:
https://skillhub.cn/skills/user_59867446/reputation-strategy-architect
AI Lianhuanhua / Sequential-Art Generator
Creative AI workflow for generating structured Chinese-style sequential visual narratives.
SkillHub:
https://skillhub.cn/skills/user_59867446/lianhuanhua-generator
DepthMotion AI — Photo-to-3D Motion
AI skill for creating depth-based 3D motion effects from still photographs.
SkillHub:
https://skillhub.cn/skills/user_59867446/depthmotion-ai
DepthMotion AI is an AI-powered photo-to-cinematic-3D-motion system.
It combines:
to transform still photographs into depth-based motion experiences.
GitHub:
https://github.com/ShenXuAkaEkstasis/DepthMotion-AI
Entity Ranking Comparison is an implementation project connected to my research into entity importance in AI knowledge systems.
The work examines entity importance through two broad classes of signals:
Audience Evaluation
and
Structural Authority
The implementation accompanies my 2026 arXiv paper:
Representing Entity Importance in AI Knowledge Systems: A Dual-Signal Framework of Audience Evaluation and Structural Authority
Paper:
https://arxiv.org/abs/2607.20925
GitHub:
https://github.com/ShenXuAkaEkstasis/entity-ranking-comparison
Viewed as a continuous product trajectory, my work has moved through several technology cycles.
Content → Communities → Distribution → Attention
My early work examined how digital platforms reshape communication, user attention, and commercial decisions.
↓
Mobile Applications → QR Codes → O2O → Consumer Interaction
Products included Banmake and Xiangge.
Xiangge-related products exceeded 5 million cumulative downloads.
↓
Product → Company → Financing → Commercialization
This included Shanghai Vanilla Technology and TUNA.
↓
Product Thinking → Consumer Behavior → Enterprise Digital Transformation
This included my role as Strategy Director at DigitasLBi Greater China / Publicis Groupe.
↓
Technology Product → Growth → Market Adoption → Commercialization
This included Yueliu, TSplus, and advisory experience with venture-backed technology companies.
↓
Retrieval → Representation → Citation → Recommendation → Selection
This led to my work on AI Visibility, Entity Ranking, and Decision Authority.
↓
AI Agents → Evaluation → Comparison → Selection → Action
Current implementations include:
↓
The next stage examines AI systems participating not only in discovery and recommendation, but directly in transactions and machine-mediated markets.
Across these projects, my current work increasingly focuses on one extended decision chain:
Discovery → Understanding → Verification → Evaluation → Comparison → Selection → Action → Transaction
This is the implementation layer connecting my work on:
The technology has changed considerably over the past two decades.
The underlying product question has changed less:
Who controls the entry point, who enters the consideration set, who earns trust, and who ultimately gets selected?
In traditional digital markets, those decisions were primarily made by people interacting with platforms.
In agent-mediated markets, AI increasingly becomes part of the decision process itself.
Name: Shen Xu / 徐慎
Professional experience:
20+ years
Roles:
Founder · Co-founder · CEO · Strategy Director · Marketing & Growth Executive · Advisor · Independent Researcher · Product Builder
Product domains:
Digital platforms · Social media · Mobile applications · O2O · Consumer software · Enterprise technology · SaaS · AI-native products · Agent systems
Selected product result:
Xiangge-related products — 5M+ cumulative downloads
Selected venture / financing experience:
Qihoo 360 strategic investment in Shanghai Vanilla Technology · TUNA angel-stage investment · venture-backed scale-up experience · later-stage technology-company advisory
Enterprise strategy experience:
Strategy Director, DigitasLBi Greater China / Publicis Groupe
SaaS commercialization experience:
TSplus China market growth and localization
Current research and product focus:
AI Visibility · Agentic Search Optimization · Decision Authority · Entity Ranking · Agentic Commerce · A2A Markets
Current open-source and AI systems:
Agentic Search Optimization Framework · A2A Decision Layer · AI Decision Mediation Framework · DSH AI Shopping Assistant · DSH AI SaaS Deal Finder · DepthMotion AI · Entity Ranking Comparison
Interactive products:
GEO Knowledge Test · ASO Framework
GitHub:
https://github.com/ShenXuAkaEkstasis
Main profile:
https://shenxuakaekstasis.github.io/