From September 22 to 24, 2026, the Yunqi Conference was held at the Hangzhou International Expo Center under the theme "AI put to work". The event featured four themed halls — computing, intelligence, creation and industry — and more than 120 forums, drawing over 2,000 experts from technology and industry and more than 1,000 exhibiting and participating companies. From AI chips and super-node servers to cloud computing, large models, agents and embodied intelligence, AI is moving faster along the chain from technology to value and into real industry.

As a global digital trust provider, eSign.AI returned to the Yunqi Conference with two core products at Booth 3-C02 in Hall 3: eSign.AI, its AI e-signature and contract lifecycle management platform, and VeriAgent.AI, trusted digital-human infrastructure.

eSign.AI: letting contracts work on their own

eSign.AI positions itself as "AI-powered signing, trust connecting the world", aiming to be a leading global AI e-signature and contract lifecycle management platform. Unlike e-signature products that treat AI as an add-on, eSign.AI embeds AI across the whole contract workflow — drafting, review, comparison, signing, performance reminders and archiving.

AI contract drafting turns "starting from a blank page" into "editing from a draft". AI contract review flags risk and compliance issues, freeing legal teams from reading every clause. Version comparison automatically spots differences between drafts so nothing is missed. Built-in enterprise e-signature completes online signing in three minutes. Performance tracking and renewal reminders keep contracts from being forgotten after signature. And the upgraded AI Contract Agent 2.0 carries the contract through its full lifecycle, end to end.

These capabilities are not demos. Our intelligent contract review reaches 98% accuracy, about ten percentage points ahead of general-purpose large models, with more than 10 million AI calls a day. Behind the numbers sits a contract model we have built in-house since 2023 and a hybrid architecture of "six vertical models plus one general model", training specialised models for extraction and risk detection rather than chasing size for its own sake. Commercially, AI-related revenue passed 100 million CNY in 2025 — AI in e-signature has closed the loop, not just appeared on a product page.

VeriAgent.AI: when an agent needs a digital identity

If eSign.AI answers how AI can do contracts better, VeriAgent.AI answers a more prior question: as AI agents begin to execute critical tasks independently, an enterprise needs to confirm not only who the person is, but who the agent is, on whose behalf it acts, and whether its behaviour is verifiable, traceable and non-repudiable.

In 2026 agents are moving from demos into core enterprise scenarios, and security incidents are arriving at the same pace. A Dark Reading reader survey found that nearly half of cybersecurity professionals rank agentic AI among the most dangerous attack vectors of 2026. Agent identity and behaviour governance is shifting from a technical topic to a regulatory one.

Built by eSign.AI together with DBAPPSecurity and the Intellectual Property Institute of the China Industrial Control Systems Cyber Emergency Response Team, VeriAgent.AI is a trusted execution platform for enterprise AI systems, filling a market gap in trusted agent management. Around identity confirmation, intent authorisation and operation attestation, it builds end-to-end protection: tamper-proof digital identities for every agent, with lifecycle management and guardian binding that integrate with enterprise IAM through certificate services and mutual authentication; dynamic, fine-grained permissions with delegated execution, quota control, overreach circuit-breakers and anomaly detection; security certification and admission for the Skills agents rely on, via static scanning, dynamic sandbox analysis, security ratings and digital signatures; and full-chain behaviour auditing from instruction input and chain of thought to tool calls and output, with cryptographic signatures and third-party timestamps to prevent log tampering.

In one line: after twenty years securing the electronic acts of people, VeriAgent.AI moves the same capability onto agents.

From people signing to people and agents signing together

Industry research puts China's enterprise-grade AI agent market at about 3.1 billion CNY in 2024, growing to about 5.7 billion in 2025 and projected to reach 59.1 billion by 2029, a compound annual growth rate above 60%. As the market expands, the compliance bar rises with it, and traceability remains the floor of trust: who did what, when, and under whose authorisation must all be verifiable.

This is exactly what eSign.AI has built over more than twenty years. As a provider ranked first in China's electronic contract market for five consecutive years, eSign.AI obtained a MIIT-licensed CA certificate in September 2024, becoming the first e-signature company to apply for and receive the licence under the new rules. Today eSign.AI's global service network covers e-signature compliance capability in more than 100 countries and regions, with deep localisation partnerships in over 50.

From "make signing easier, make trust simpler" to "make AI behaviour traceable and accountability definable", the subject of eSign.AI's trust services is extending from people to agents, and from contracts to the entire AI execution chain. The same discipline that keeps a contract's record tamper-proof is what makes an agent's actions auditable — a problem we explore further in how AI reshapes contract management.

Event: 2026 Yunqi Conference, Hangzhou International Expo Center, September 22-24, 2026. eSign.AI exhibited at Booth 3-C02, Hall 3.