NeuroAI NEUROAINEUROAI.SITE
China AI, reported clearly
1,300+ verified stories on models, robotics, BCI and the business of AI.
Browse stories
ESC

Alipay's finance bot Zhi Xiaobao reached 59 million users by answering fund questions, not by picking stocks

Ant Group's Zhi Xiaobao lives inside Alipay and answers investment and insurance questions at scale. The user count is verified by QuestMobile; the 'advice' it gives is bounded on purpose.

2026-11-11 · 1099 words · NeuroAI
Alipay's finance bot Zhi Xiaobao reached 59 million users by answering fund questions, not by picking stocks

A first-time fund buyer types "why did my balanced fund drop this week" into a chat box inside the Alipay app. Seconds later she gets a plain-language answer, a holdings breakdown, and a reminder that past performance is not a promise. No call to a wealth manager, no sales pitch — just a reply inside the same app where she already holds the product.

That chat box is Zhi Xiaobao (支小宝), Ant Group's finance assistant, and by mid-2024 it had become one of the most-used consumer AI products in China's money world. Its growth tells a quieter story than the headline models: the winning move was answering questions about money people already had, not telling them what to buy next.

Key takeaways

  • Zhi Xiaobao (支小宝) is Ant Group's AI finance assistant inside Alipay (支付宝), built on the Ant Lingming (百灵) large model, focused on fund and insurance questions.
  • Independent researcher QuestMobile measured 59.08 million monthly active users as of June 2024, with 5.59 million daily uses and about 2.8 uses per user that month.
  • Ant said an earlier version had been tried by about 12 million users before 2.0 entered external testing in the second quarter of 2024, and by its 27 May tech day cumulative users who had used Zhi Xiaobao reached 43 million.
  • It runs inside Alipay's Ant Fortune (蚂蚁财富) and Ant Insurance (蚂蚁保) channels, supports text and voice, and covers market interpretation, holdings analysis, insurance matching and investor education.
  • Ant Fortune reported that users it served deeply had 5% more rational asset allocation and 60% lower frequent-trading rates than users it did not serve.

A bot that lives where the money already is

Most consumer AI products have to win attention from scratch. Zhi Xiaobao did not. It sits inside Alipay, the super-app where hundreds of millions of Chinese users already buy funds, pay bills and hold insurance. When someone opens the Ant Fortune or Ant Insurance tab, the assistant is one tap away, in text or voice. That distribution advantage — not a smarter model — is the reason a finance bot reached tens of millions faster than most standalone apps.

The assistant is deliberately narrow. It answers questions about funds and insurance, explains a market move, breaks down what a user holds, helps match an insurance product to a need, and teaches basic concepts. It is not a broker and it does not execute trades on its own. Ant has described it as the finance-field application that first reached scaled, customer-facing deployment — meaning real users, not a pilot.

What the verified numbers show

The clearest outside measurement comes from QuestMobile, cited by CCTV Finance, Eastmoney and Sichuan Daily. As of June 2024, Zhi Xiaobao's monthly active users reached 59.08 million, people used it 5.59 million times a day, and the average user opened it about 2.8 times that month. Those are independent figures, not Ant's own counts.

Ant's own disclosures fill in the earlier ramp. The 2.0 version entered external testing in the second quarter of 2024 and had been tried by about 12 million users in the first quarter. By Ant's 27 May tech day, people who had ever used Zhi Xiaobao totalled 43 million. The jump from 43 million cumulative in May to 59 million monthly actives in June is consistent with a product going from trial to habit inside Alipay's traffic.

The assistant's engine is the Ant Lingming (百灵) model, and Ant says the 2.0 version's finance knowledge store reached the 10-billion scale, running on a multi-agent framework called agentuniverse (多智能体协同) that simulates specialist roles — analysis, review, reflection — to check its own answers. On the insurance side it offers smart underwriting, judging insurability from a user's health, and smart claims that can fully automate review for simple outpatient policies.

Where it helps, and where it stops

The most cited benefit is behavioural, not informational. Ant Fortune reported that users Zhi Xiaobao served deeply built 5% more rational asset allocations and traded 60% less frequently than users it did not serve — a nudging effect that, if real at scale, matters more than any single answer. A calmer investor who holds a balanced plan is worth more to a platform than one who churns.

The boundaries are just as important. The bot works off the user's actual holdings, so its "market interpretation" and "this week's gain" answers are grounded in data the platform already has. It flags risks — fund-size shifts, manager changes — rather than promising returns. China News noted that during 2024's shaky equity rebound, the hottest questions were "market outlook" and "this week's earnings," and the assistant was built to answer while reminding users of risk.

According to Li An, chief scientist at BrainNet (脑机网), a finance assistant earns trust only when it is allowed to say "I can't decide that for you," and the real value of such tools lies in keeping the human in the loop for financial decisions.

Li An's comments reflect his own views and do not represent NeuroAI's editorial position or an endorsement by BrainNet (脑机网).

Honest limitations

The headline figure — 59.08 million monthly actives in June 2024 — is from QuestMobile and is the most reliable number here, but it measures usage inside Alipay's finance tabs, not paid advisory relationships, and it predates any later-year update, so it should not be read as a current total. Ant's 43-million cumulative and 12-million trial figures are company disclosures, not independently verified, and "cumulative users" and "monthly actives" are different metrics that should not be added together.

The 5% and 60% behavioural claims come from Ant Fortune and describe correlation among users the assistant "served deeply" — the company does not publish the methodology, sample size or whether the comparison controlled for who opts into such tools (typically already more engaged users). Treat them as directional, not proven causation. Finally, this is a China-only, Alipay-bound product; its model, compliance wrapper and distribution cannot be assumed to transfer to other markets or to unregulated advice.

What readers can do now

  1. If you hold funds or insurance inside Alipay, open Ant Fortune or Ant Insurance and try the assistant with a real holding question — it is strongest on "explain what I own," weakest on predictions.
  2. Use it to slow yourself down, not to replace judgement: ask for a holdings breakdown before any buy, and ignore any answer that sounds like a forecast.
  3. Keep the distinction clear — a finance assistant that answers questions is not a licensed adviser, and no chat output should substitute for the risk disclosure on the actual product page.

Related coverage

Cite this article

NeuroAI (脑机网), "Alipay's finance bot Zhi Xiaobao reached 59 million users by answering fund questions, not by picking stocks," 11 November 2026, https://neuroai.site/a/na-app-alipay-zhixiaobao

Machine-readable: BibTeX · RIS · JSON · Markdown

More in “AI in Action” → · Back to home · Markdown version

China AI, reported clearly
1,300+ verified stories on models, robotics, BCI and the business of AI.
Browse stories