AI agent marketing just took another step from pilot chatter to commercial infrastructure. Datasea Intelligent Technology Ltd. (NASDAQ: DTSS) said on Sept. 17 that its Chinese operating entities signed cooperation agreements and pre-order arrangements with five customers, including Beijing Judongjiujiu Technology Co., Ltd. The work spans digital marketing, content generation, intelligent ad placement and SaaS service applications.
What the deals actually include
All five customers have started receiving services, with an aggregate delivered value of about US$1.10 million. Beijing Judongjiujiu Technology Co., Ltd. represented the largest share at roughly US$0.51 million.
The anticipated fees range from approximately US$0.45 million to US$1.49 million per customer per month. Across all five arrangements, that totals an expected US$2.24 million to US$7.45 million in monthly service volume. If the top end were sustained for a full year, the potential annualized service volume would be about US$89.4 million.
That is a big number, but it is not revenue. Datasea’s release explicitly states the amounts are not minimum purchase commitments, guaranteed revenue or backlog. Actual results will depend on customer usage, platform billing, ad placement execution, settlement and accounting treatment.
Why usage-based billing changes the conversation
The most relevant detail for growth teams is the billing structure. The customers use what Datasea calls a “recharge, consumption and service” model, drawing down AI service credits as they buy content generation, creative material, intelligent ad placement, analytics and optimization.
This matters because performance marketing already runs on consumption: ad spend, impressions, creative tests and optimization all scale with usage. If AI agent services plug into that same operating rhythm, they stop being a software line item and start being judged like a channel.
How to evaluate the opportunity
Datasea describes the business as a three-layer system. It is a useful checklist for any marketer vetting AI agent vendors.
- Platform layer: accounts, model invocation, system interfaces, data review and technical integration.
- Execution layer: content generation, digital marketing, ad placement, analytics, SaaS services and usage billing.
- Scenario layer: beauty, health and wellness, lifestyle services, retail, apparel and digital marketing use cases.
CEO Zhixin Liu said AI agents can combine content generation, placement strategy, execution, analytics and optimization into one loop. He also reiterated a previously disclosed Rule 10b5-1 plan to purchase up to US$1.0 million of Datasea shares with personal funds.
Leadership confidence is nice, but the operating lesson remains: pre-orders and usage estimates are early signals. Your real evaluation should track actual consumption, creative output, cost per acquisition and whether the agent improves what a human-led campaign already does.
Start small. Pick one campaign or one product line, set a baseline for CPA and creative turnaround, then compare the AI agent’s usage-based cost against that baseline. If the loop does not create clear operating leverage, it is just another SaaS subscription with extra steps.
Source: Cision PR Newswire
