R Roy.Tong

ROY TONG · RESEARCHING HOW AI BECOMES PRODUCTS, SYSTEMS AND MARKETS

Turning emerging technology into real products, systems and markets

I study how AI moves into real tasks and the physical world. Right now I am focused on one structural change: the software consumer is shifting from humans to agents.

Know which agent calls actually create value. Measure real agent operations — not just tool calls. First product: a free 7-day audit for capability providers.

Free 7-day Capability Audit Current campaign: Agent Capability Economy Agent Capability Index 30-second overview

CURRENT FOCUS

Agent Capability Economy

When agents become software consumers, how will software be discovered, chosen, invoked, metered, and paid for? This is the structural reconstruction of the software economy — the intersection of the long-term brand and the current campaign.

About the campaign → View the capability map

Why this campaign: the industry already has fragmented discussions of Skills, MCP, API, CLI, runtimes, marketplaces, and automated transactions — but no complete question. Measurement is the gap that must be solved first.

Long-term brand
→ How AI becomes products, systems and markets
→ Structural change: agents become software consumers
→ New question: how software is discovered, metered and paid for

THREE ASSETS

Growth comes from three assets, not constant posting

Traffic solves the first sight. Authority solves why to believe. Identity solves why to follow Roy.

TRAFFIC ASSET

Agent Capability Index

A public map of the agent software world: agent-invocable software organized by capability, with provider, interface, pricing, and sources. Worth bookmarking even if you disagree with the thesis.

Enter the map → AUTHORITY ASSET

AgentMeasure

Open measurement infrastructure for the Agent Capability Economy: Reach → Choice → Use → Utility → Value. Measure discovery, choice, use, and value before building metering and marketplaces.

View the project ↗ IDENTITY ASSET

Essays · About · X

Converting asset awareness into attention on Roy: long-term judgments, research method, and personal background. Public assets earn value first; the assets then transfer reputation to Roy.

Get to know Roy →

RESEARCH · FACTS

Fact research

Research answers "what is actually happening": facts, companies, products, technical paths, and market change. The ecosystem map and data ledger of the current campaign live here.

CURRENT FOCUS

Agent Capability Economy

The agent software ecosystem map, Capability Index data ledger, monthly updates, and source ledger — the only campaign running at full intensity.

Enter the hub ↗
01

Embodied Intelligence

How robots move from technical demos into real tasks, product definitions, user value, and viable business models.

Robotics
02

Multimodal Interaction

How voice, vision, space, motion, and environmental sensing reshape interaction across AI wearables and spatial computing.

Interaction
03

On-device AI

How models, chips, power, privacy, and local data define the product boundary of AI terminals, AI boxes, and edge systems.

Edge Compute
04

Agents

How agents acquire context, choose capabilities, call software, complete tasks, and reshape software infrastructure and economics.

Agent Systems

ESSAYS · INTERPRETATION

Interpretation

Essays answer "what the facts mean" — long-term judgments open to discussion and refutation.

Browse all writing →
FLAGSHIP ESSAY

When the Software Consumer Becomes an Agent

The old measurement chain (installs, seats, page views) breaks for agent consumers; why measurement precedes payment.

After AI Native, the basic unit of the product changes

From pages, features, and apps toward intent, tasks, capabilities, and outcomes.

After RAG, agents need Context Recommendation

Complex agents must dynamically decide what context, tools, state, and rules to inject on each turn.

Every Agent Usage Number Is Self-Reported

A field audit of six real agent-usage claims: independence beats volume, and nobody publishes the unit definition.

PROJECTS · PROOF

Proof

Projects answer "can this judgment become something real". I do not optimize for count; campaign projects come first.

FLAGSHIP PROJECT

AgentMeasure

Open measurement infrastructure for the Agent Capability Economy: Reach → Choice → Use → Utility → Value. Measure how agents discover, choose, use, and derive value from software capabilities before building metering, marketplaces, or CaaS on top.
01

Agent Capability Index

A public map of the agent software world and the factual foundation of the current campaign. Open contribution: add a capability, correct an error, claim your provider profile.

Campaign ↗
02

AgentMeasure Benchmark

Evaluating the current agent software ecosystem by measurement quality rather than hype.

Benchmark ↗
03

User Demand Research (SURE)

Turn interviews, reviews, and behavior records into auditable and refutable demand evidence.

Tool ↗
04

iRead Research Monitor

Discover sources continuously and generate evidence-bounded daily, weekly, and monthly reports.

Tool ↗

ABOUT

About me

For nearly a decade, I have built products across AI, software, and intelligent hardware, with several startup chapters along the way. I study not only how technology is built, but what products, user behavior, and business structures a technology change ultimately reshapes.

More about me → X / Twitter ↗ Contact me →

Building software, measurement, or commerce for the agent era? Let’s talk.

If you are building agent products, capability markets, metering and payments, or complex zero-to-one products, bring the task, user, and hardest unknown.

Contact Roy