Ant Digital на WAIC 2026 представила Agentar 2.0 — «супер-фабрику» бизнес-агентов

Ant Digital Technologies (蚂蚁数科, the digital division of Ant Group) unveiled a "super-factory for business agents" and the Agentar 2.0 platform on July 17, 2026 at the WAIC 2026 conference in Shanghai — featuring almost 200 templates of job-level digital experts and hundreds of ready-to-plug agent tools.

What Agentar 2.0 is

Agentar 2.0 is Ant Digital's platform where the company assembles ready-made AI agents for specific work roles, so businesses don't have to build them from scratch. The first batch includes templates of job-level digital experts and subscribable Skills-level tools, each with its own professional knowledge, business process, and ability to call tools.

• Announcement date — July 17, 2026, WAIC 2026 in Shanghai
• Almost 200 pre-built templates of job-level digital experts
• Hundreds of subscribable Skills tools "out of the box"
• Three pillars of the approach: industry focus, scenario focus, ecosystem focus
• Foundation — Ant Group's industry vertical models and more than 20 years of technology experience

WAIC 2026 was held under the theme "Intelligent Partners, Shared Future," and, according to 36Kr's assessment, the industry's focus is shifting from the race for model size toward the value of its application in real business processes. Ant is betting on the APIfication of components so partners can quickly connect and build out solutions for their own industry.

Why it works in finance

In financial scenarios, Ant Digital claims agent answer accuracy of 90–95% and agreement with a human expert's answer above 85% — according to the company's vice president Sun Lei. Finance, in his words, sets up three barriers: security and compliance, professional accuracy (the hallucination problem), and effect verification — ROI.

Ant's answer is a combination of an industry vertical model and engineering "wrapping" tailored to the business process, plus the principle of "pay for results": the client pays for a verifiable effect — a rise in marketing conversion or a reduction in risk-control losses. The company draws on its own experience in payments, wealth management, and insurance.

"Better not to do it than to get it wrong," —

Sun Lei, vice president of Ant Digital Technologies, on financial regulators' requirements for AI deployment.

A Chinese version of Harness

Ant Digital wants to build a Chinese industry version of Harness — the engineering "wrapping" that turns a model's capabilities into a working business process. Where finance previously ran on the logic of "three workshops," the company has split Harness into four: intent recognition, planning, execution, and answer formulation. Intent recognition is key here: a client asks about insurance, but the real request is portfolio selection or wealth management.

Ant Digital is extending the approach beyond finance — into healthcare, energy, transportation, and aviation. For example, in solar energy the company uses a time-series model that, factoring in weather and on-site conditions, forecasts electricity output. An evaluation system plays a key role: in a project for one large city commercial bank, Ant ran 28 rounds of testing before confirming the agents were ready.

What it means

Ant Digital is shifting the competition from "model size" to its embedding into real business processes. The company's bet is not on a universal chatbot but on industry agents with verifiable ROI and pay-for-results pricing; over the next 6–12 months it promises to expand its agent matrix and gradually open up industry Harness capabilities.

Frequently Asked Questions

What is Agentar 2.0?

Agentar 2.0 is Ant Digital Technologies' platform for quickly launching business agents. The first batch includes almost 200 templates of digital experts for specific job roles and hundreds of subscribable Skills tools ready to work "out of the box."

What accuracy does Ant claim for its industry agents?

According to the company, in professional financial scenarios agent answer accuracy reaches 90–95%, and agreement of their answers with a human expert's answers is above 85%.