China Merchants Bank (CIHKY): AI Assistant Now Covers Every Golden Sunflower RM

Summary
China Merchants Bank's relationship managers use an AI assistant to pick clients and draft proposals; first-half transaction value per client rose 35.82%, revenue impact unquantified.
China Merchants Bank (CIHKY) told its 2026-08-30 interim results call that its AI assistant for relationship managers is now in the hands of every relationship manager covering Golden Sunflower clients, and that in the first half of the year the average number of effective outreach customers per RM rose 14.65% while average transaction value per client rose 35.82%. The bank did not convert either percentage into any revenue line [1].
What the assistant does, and who actually uses it
China Merchants Bank is a Chinese joint-stock commercial bank best known for its retail franchise. Part of its revenue is the interest spread between deposits and loans; the rest is fee and commission income from distributing wealth-management products, funds and insurance, and from credit cards and settlement services. Golden Sunflower is the tier label it applies to retail clients with larger asset balances. Those clients are covered one-on-one by a relationship manager, while clients with still larger balances are served by private banking advisers.
On the call the tool was named RM Assistant, alongside newly launched private banking AI tools. The people using it are those relationship managers and advisers, not the clients. Working from the client book a manager already owns, the assistant sorts out who should be contacted first, drafts the asset allocation, protection and wealth-succession proposal, and drafts what the manager will say. Reaching the client and walking them through the plan remains the manager's own work. AI Xiao Ban, mentioned on the same call, is aimed at external fund, wealth-management and insurance partner institutions according to the bank's announcement, and is not part of this manager-facing toolset.
How the disclosure moved from March to August
The assistant first appeared with numbers attached on the 2026-03-30 annual results call: over 10,000 assistants were already in the hands of Sunflower relationship managers and corporate client managers, the customer reach ratio had improved by around 14%, and management filed those changes under macro-side changes and said the work was still in process [2].
By the 2026-08-30 interim call, coverage had widened from some relationship managers to all bank-wide relationship managers responsible for Golden Sunflower clients, and the newly launched private banking AI tools folded protection and wealth succession into the same set of recommendations. Asked what AI had actually done for the business, management offered this as its first example [1].
What the two percentages measure, and what they do not
Both figures are bounded by a period and a population: the first half of 2026, and the relationship managers covering Golden Sunflower clients. Inside that scope, effective outreach customers per RM rose 14.65% and average transaction value per client rose 35.82% [1]. The bank did not define what counts as effective outreach, and gave no base levels, client counts or transaction amounts. So the figures measure the output of the relationship managers and the transaction intensity of the clients they cover; no revenue increment can be read out of them.
They also cannot be set against the roughly 14% from March to claim acceleration. The March figure was an improvement in a reach ratio with neither a period nor a base behind it; the August figure is effective outreach customers per RM over the first half. The denominators are different [2].
Where the chain would land in the financials
Follow the operating chain and it runs this way: the more clients a relationship manager can effectively reach, and the larger each client's transactions, the larger the distribution and retained balances of wealth products, which end up in wealth-management fee and commission income. The bank's disclosure for the same period shows broad wealth-management income and retail AUM moving in that same direction [1].
For now that path is only directionally related. Management never converted 14.65% or 35.82% into a contribution to any revenue line. The Golden Sunflower and above client base was itself expanding over the same period, and distribution of non-money-market funds and trusts was recovering; either development on its own is enough to lift both the outreach base and per-client transaction value. Group operating income and net profit attributable to shareholders each grew far more slowly than these two percentages, which says that even if the transmission holds, its size has not been separately quantified.
What is confirmed, and what would settle the rest
The operating change can be confirmed: the assistant has become standard equipment across China Merchants Bank's retail wealth line, and the relationship managers using it reach more clients and book higher transaction value per client [1]. What is not settled is what that is worth. Judging how much wealth-management fee and commission income this operating improvement contributed would take a separate disclosure of the wealth-product sales or fee income these relationship managers generated, or the absolute bases behind effective outreach customers per RM and average transaction value per client.
Application assessment
- AI Assistant for Client Relationship Managers | Business position: core operations | Adoption stage: limited production | Scope: multi-business or multi-region | Value type: revenue growth
Sources
[1] Drillr - China Merchants Bank Co., Ltd. (CIHKY) - 2026-08-30 - earnings call
"In the first half of the year, the average effective outreach customers per RM increased by 14.65%, with average transaction value per client increased by 35.82%, generating good effects."
[2] Drillr - China Merchants Bank Co., Ltd. (CIHKY) - 2026-03-30 - earnings call
"And we have over 10,000 assistants for our Sunflower relationship manager and also for how our corporate manager help them to improve the customer recharge ratio by around 14%."