Turning losses into profits, revenue up 3.2%, the cost structure revolution behind Minglue Technology's 2025 annual report

Ask AI · How can AI technology help Minglue Technology achieve a cost-structure revolution?

Over the past three years, China’s professional services industry has undergone a brutal reshuffling. On one side, leading clients have tightened budgets and upgraded their ROI demands; on the other, technology change driven by AI continues to reshape delivery logic. Against this backdrop, the old growth model of “stacking headcount and doing projects” has become increasingly unsustainable. Only companies that can complete self-restructuring in the AI wave can find new growth anchors in this changing landscape.

The full-year performance for fiscal year 2025 recently released by Minglue Technology (2718.HK) provides a sample worth reading closely. During the reporting period, the company achieved operating revenue of RMB 1.43B, up 3.2% year over year; gross profit reached RMB 790 million, up 10.8%; and adjusted net profit was RMB 42.04 million, turning from loss to profit—whereas in the same period of 2024, this figure was a loss of RMB 45.11 million.

On its own, the 3.2% revenue growth rate isn’t particularly impressive. But if you look at this number through the lens of profitability logic, you’ll understand what this annual report truly wants to convey: the reversal in profit isn’t achieved by increasing revenue scale, but by a systematic internal efficiency reconfiguration.

Structural changes behind three sets of numbers

The key signals in the financial report are hidden in the comparison of three sets of figures.

First set: revenue growth of 3.2%, gross profit growth of 10.8%. Gross profit growth is three times revenue growth, which means the company is selling something more “valuable”—or, with the same revenue scale, consuming less cost.

Second set: adjusted operating profit jumped from RMB 5.8 million in 2024 to RMB 24.98 million, while operating losses narrowed sharply from RMB 132 million to RMB 15.72 million. Behind this is AI’s systematic intervention in internal delivery processes—already starting to form a tangible leverage effect at the financial level.

Third set: In its first year, the Agentic Services business segment generated revenue of over RMB 100 million, and among newly added major clients, more than 30% came from this segment. This is a commercial validation of a brand-new direction from 0 to 1, and it’s also an important source of improvement in gross margin structure.

Put these three sets of numbers together, and you can see a clear logic chain: higher-margin contributions from new businesses lead to a better gross margin structure; AI-driven internal transformation reduces the marginal cost of existing businesses. Together, these effects drive the company’s leap from losses to profitability.

How AI restructured delivery costs

Minglue Technology’s profit improvement first comes from a large-scale internal AI experiment.

During the reporting period, its subsidiary Miaozhen System spent 60 days to complete a full-chain Agentic AI transformation. This was a systematic reconfiguration of core business processes such as processing social media data, insight analysis, and report generation—while retaining necessary human-machine collaboration nodes, the full-chain AI automatic completion rate reached 90%, and the human-efficiency for producing deep-replay reports improved by 20 times.

For the marketing intelligence business, delivery efficiency improvements reached as much as 4 times, meaning that with the same delivery scale, the required human effort drops significantly. For the operations intelligence business, the time to resolve tickets was compressed by more than 30%, meaning that with the same number of customers, customer service costs decline.

Individually, these numbers are all impressive efficiency indicators. But only by looking at them together can you understand the underlying business logic: the cost structure of a professional services company is naturally constrained by the “headcount grows linearly” rule. Once scale expands, labor costs inevitably increase year over year. After AI is embedded, the marginal cost of delivering to new customers begins to decline systematically—this is the fundamental reason why the profit leverage forms.

It’s also worth noting that Minglue Technology has achieved 100% access to the DeepMiner platform for all employees. Employees can create and deploy dedicated business Agents on their own, gradually evolving internal standard operating paths from human execution to human-machine collaborative orchestration. This is not just a tool rollout, but an upgrade to the organization’s operating system level—it expands AI’s efficiency dividends from a few pilot scenarios to everyday operations across the entire organization.

Gross margin contribution from the new business segment

The second clue to the profitability improvement comes from Agentic Services’ gross margin contribution.

Unlike traditional SaaS software or project-based services, Agentic Services adopts an outcome-based pricing model: customers pay for quantifiable business results, not for how many tools are deployed or how many reports are produced. The core advantage of this model is that the service provider can use AI-driven automation to reduce marginal delivery costs while maintaining relatively high customer pricing.

According to research by Frost & Sullivan, Minglue Technology is China’s largest data intelligence application software provider by total revenue in 2024. Building on this base, the addition of Agentic Services is not only adding a new business line, but also grafting a new business logic onto the existing data intelligence foundation—upgrading from selling “insights” to selling “results,” and moving from one-time project contracts to long-term outcome-based collaboration.

The major client renewal rate reached 96%, confirming that this service upgrade is being accepted by the market. Customers’ willingness to renew, in essence, is “voting with real budgets”—a market signal that is more convincing than any survey data.

Profitability is the starting point, not the finish line

In the annual report, Minglue Technology’s founder and CEO Wu Minghui clearly stated that in 2025 the company completed a key transformation—from helping clients understand data to helping clients achieve results. The significance of Agentic Services lies exactly in pushing AI’s value from tool delivery to outcome delivery. This not only changes the logic by which customers procure professional services, but is also reshaping the entire way services are delivered and business models.

From a financial perspective, profitability in 2025 is more like an inflection point than an endpoint. As Agentic Services continues to scale up, as the DeepMiner platform continues to go deeper into organizational operations, and as more business scenarios complete AI transformation, the room for releasing profit leverage will theoretically continue to expand.

For companies exploring the proposition of “how AI can truly help enterprises reduce costs and improve efficiency,” Minglue Technology’s 2025 financial data provides a quantifiable and comparable real-world case—not a grand narrative about AI potential, but what can be seen on the income statement after AI is implemented. Perhaps that is the most deserving part of this annual report to be taken seriously.