Shanghai, China – October 12, 2026 – In the first half of AI, the competition was about model capabilities; in the second half, it is about delivering results. Recently, Microsoft released Microsoft-Decision-1, a model that does not generate dialogue but only outputs structured decision results, covering routing, classification, prioritization, validation, and workflow control.
Decision-1 represents the specialized small-model route. Such models have parameter scales far lower than general-purpose large models, and their unit inference costs are naturally lower. General-purpose large models, meanwhile, are also continuing to cut prices along another path. Recently, OpenAI released GPT-6 Sol and GPT-6 Luna, with prices about 50% lower than the previous generation; Google’s Gemini Flash series is running at promotional prices through the end of 2026.
For model vendors, price cuts mean intensifying competition; for application-layer companies, they mean continued cost concessions. But the “application layer” is not a homogeneous whole: whether cost declines can be converted into profit depends on whether the results of the business can be measured and whether customers are willing to pay for results.
By this standard, enterprise marketing is currently one of the areas with the shortest path to monetization—marketing budgets are already directly tied to campaign results, customers already have established habits for measuring input-output, and once AI increases lead volume and conversion rates, willingness to pay is immediately reflected.
Marketingforce Turns Model Costs into Billable Results
Hong Kong-listed Marketingforce (2556.HK) is an enterprise-level AI application service provider. Its products are centered on AI agents and serve two types of customers.
For small and medium-sized enterprises, it provides full-link, full-scenario one-stop AI marketing applications: agent clusters automatically generate marketing content, cover mainstream traffic ecosystems, and acquire business leads based on the traffic logic of large-model recommendation entry points, mainly solving the problems of high customer acquisition costs and low conversion rates for SMEs. For large enterprises, it provides enterprise-level AI Agent application clusters and an AI Agent development and management platform, using its self-developed Tforce vertical marketing large model and enterprise-level agents to cover all-channel customer acquisition, customer management, automated marketing, customer service, R&D, and training. In terms of industries, it covers consumer retail, automotive, finance, healthcare, cultural tourism, manufacturing, education, the internet, and more.
In terms of product lines, the marketing side includes GTM Agent, Mirror World, AI CMS, AI KOX, AI CDP, Agentic DAM, and T-GEO, which connects the entire process “from asking questions on AI application platforms to obtaining answers”; the sales side includes AI-native CRM, AI solutions, AI bidding assistant, and AI contract review; in addition, there are AI customer service covering both To C and To B, AI training assistants that automatically generate scenarios by industry, DaTa Agent for business attribution analysis, EVA automated overseas customer acquisition and Nora AI video marketing for foreign trade, and AI R&D agents for research and development.
Marketingforce believes that the core value of enterprise AI applications lies not in simply consuming underlying model Tokens, but in transforming underlying computing power and model capabilities through enterprise business scenarios into quantifiable “scenario Tokens” that can sustainably generate value. Only when AI applications shift AI capabilities toward AI being able to create deterministic business results do they have commercial value.
AI application business revenue exceeded 1.1 billion yuan in the first half
As model prices continue to fall, the growth space for AI applications will open up faster. The previous generation of software was priced by functional modules and seats, while this generation of AI applications is more often priced by the workload replaced and the results delivered: the former’s price depends on the vendor’s pricing power, while the latter depends on the customer’s confirmation of results. This shift determines how much of the profit released by falling model prices an AI application company can retain.Marketingforce’s performance data for the first half of 2026 reflects the change in revenue structure under this logic.
In the first half of 2026, Marketingforce’s total revenue was RMB 1.960 billion, up 111.2% year on year. Among this, revenue from AI application business was RMB 1.128 billion, up 123.7% year on year, rising to 57.5% of total revenue, with a gross margin of 76.6% (84.9% after excluding the newly added agent all-in-one machine business); revenue from precision marketing services was RMB 832 million, up 96.3% year on year.
At the same time, in the first half of this year, Marketingforce’s number of customers increased by 25.9% year on year, and average monthly revenue per user increased by 80.0% year on year, with customer value continuing to be released.
Marketingforce deeply combines underlying computing power and model capabilities with enterprise data, industry knowledge, and business scenarios, converting general Tokens into billable, reusable “scenario Tokens.” The deeper customers use the products, the richer the business knowledge accumulated by agents, and the more pronounced the application effects, making renewals, expansions, and long-term cooperation more likely, thus forming a growth flywheel.
But the cheaper large models become, the less gentle competition will be. Application vendors that merely wrap and call models will quickly have their price differentials erased by price cuts; only companies that truly master data, platforms, scenarios, and delivery capabilities have the chance to turn cost dividends into profit. Marketingforce is currently betting on full-stack capabilities of “model + data + platform + scenarios,” and is trying to build barriers with years of accumulated enterprise service experience, industry knowledge assets, a mature AI application platform, and an expanding customer base.
At least judging from its first-half financial report, Marketingforce has already proven that enterprise-level AI applications are not only capable of telling stories; results have already been delivered. The real test is: as model prices continue to fall, can it continuously turn technological capabilities into customer results, and then continuously turn customer results into revenue and profit on its financial statements.
About Marketingforce
Marketingforce isThe World’s Leading AI Application Platform. Founded in 2009, the company operates over 30 branches worldwide. It has filed more than 800 software copyrights and patents in the fields of AI and digital intelligence. It has been honored with the Second Prize of the National Science and Technology Progress Award and the First Prize of the Shanghai Science and Technology Progress Award, and has twice been featured as a benchmark enterprise in AI-enabled digital transformation in the special coverage of the China News“Two Sessions” edition.
Focusing on artificial intelligence, Marketingforce provides competitive products and services, continuously creating value for its clients. To date, Marketingforce has served over 210,000 enterprises across various industries, including retail and consumer goods, automotive, finance, B2B manufacturing, healthcare, enterprise services, and cross-border e-commerce.
Media ContactCompany Name: Marketingforce Management LtdContact Person: Yanbin.gaoEmail: Send EmailCity: ShanghaiCountry: ChinaWebsite: https://www.marketingforce.com/about/aboutus.html
