Since Feishu helped AI truly integrate into the organization, Mingyue Health Technology Group has begun re-examining corporate management and gaining a renewed understanding of organizational structure itself.
In Qingdao's West Coast, in 1968, a company named "Mingyue Seaweed" was born.
In the more than half-century since, it evolved from a near-bankrupt old factory into the world leader in the alginate industry, capturing over 40% of the domestic market and more than 30% internationally, with products sold in over 100 countries and regions.
In recent years, this 58-year-old “time-honored brand” is showing fresh industrial vitality, with the responsibility for new business development entrusted to Zhang Jian.
Zhang Jian carries prominent labels: post-80s entrepreneur and second-generation successor. But unlike many peers in similar roles, he remains clear-headed standing on the foundation built by his predecessors:
His father laid out a large ship, but that vessel must now sail into entirely uncharted waters. Markets are changing, customers are changing, and competitive logic is shifting. Can methods that once worked reliably still carry them to their next destination?
Someone else might stick to conventional ways. But Zhang Jian made a decision that surprised many:
He launched an AI transformation pilot within the health technology group under Mingyue Group, spending four months integrating R&D, production, and sales across the entire chain—boosting efficiency tenfold.
Customer response cycles were compressed from three days to two hours; new product development shortened from one month to one or two weeks; and in just one month, he conducted more interviews than in all previous years combined. More importantly, within this 58-year-old enterprise, AI genuinely began entering and functioning within the organization.
Not long ago, I had an in-depth discussion with Zhang Jian about the company’s AI transformation:
Why did Mingyue Health Technology Group decide to push for comprehensive AI transformation? What challenges and obstacles emerged during the process?
As the company's top leader, how did he overcome these hurdles and lead the organization toward further AI advancement?
In this era of rapidly accelerating AI waves, what value and insights can Zhang Jian and Mingyue Health Technology Group’s AI transformation offer?
Drawing from my conversation with Zhang Jian, each of these questions found its answer. This “inward-facing” corporate AI transformation can be summarized as “four battles.”
The Breakthrough Battle:
The Boss Trapped by Information
The story begins with a “leader’s transformative reflection.”
In September 2005, Zhang Jian led a bold business transformation within Mingyue Health Technology Group, shifting one of its core divisions from the maternal-infant consumer sector to the adult consumer market.
The strategic decision itself was sound.
Judging from overall market trends, the maternal-infant market was shrinking while the adult health market was growing—so the direction was correct.
But after several months, Zhang Jian discovered a frustrating yet predictable situation:
The direction changed, the battle was being fought, but the tactics remained outdated.
“You say one thing, they do another—the team continued operating under the old logic even in the adult market, whose commercial principles are clearly different,” Zhang recalled. “I tried every method available—communication, performance evaluations, salary adjustments—even increased pay for some managers hoping they’d take the lead. The result? Minimal impact.”
The deeper challenge lay in information flow.
Zhang realized that as the company’s decision-maker, obtaining frontline feedback on a product required层层 inquiries; validating a sudden business idea meant waiting for various departments to conduct separate investigations and then convene meetings—losing one or two weeks in the process.
“Sometimes, even as the number-one person in a company, your internal perspective can be extremely limited. The higher you sit, the farther information drifts from you. Often, unless you proactively ask or call a meeting, you may be the last person in the company to know the truth.”
What troubled him most was that these bottlenecks weren’t primarily at the grassroots level.
Frontline employees actually adapted quickly to new things. The real roadblock came from certain middle managers—some mid-level leaders, having long been detached from frontline operations, relied on aggregated information and experience-based judgment; others subconsciously believed that maintaining information asymmetry was key to securing their positions.
This isn’t unique to one company—many enterprises face similar issues.
All conventional management tools had failed. Zhang realized this wasn’t a people problem—it was a flaw in the “old map” itself.
“The organization can no longer rely on ‘people’ to relay information, hold meetings for alignment, or layer upon layered reporting. This model itself is the bottleneck.”
Zhang Jian, Chairman and CEO of Mingyue Health Technology Group
It was at this moment that he started looking outward—to bosses who were genuinely profiting from AI, and to organizations moving faster and operating differently.
“I studied 100 bosses who’ve successfully made money using AI—how they used AI to transform their companies, manage organizations, even embed AI into their products to create profitable offerings.”
He noticed a common thread: behind these companies’ AI applications stood Feishu.
“I was intrigued, so I began researching Feishu.”
After his research, he made a decision that shocked everyone: uninstalling Office across the entire company.
Why such a radical move? Because he realized AI was fundamentally different from any prior technological shift.
Over past decades, the company had weathered waves of reform and opening-up, internet adoption, and digitalization—each representing upgrades at the tool level: replacing pen and paper with spreadsheets, turning manual processes into systemized workflows. Information flow improved, communication costs dropped, and decision-making accelerated.
But AI wasn’t an upgrade—it was a transformation.
“It completely changes our understanding of corporate management and organizational structure.” An individual’s capability boundary expands infinitely—a newcomer can achieve 60–70% professional proficiency in a short time, and experienced professionals can compress a decade-long growth cycle into under three years.
“In the past, we spent two months creating a market research plan. Now, most of it can be done in two days.”
Zhang’s judgment was straightforward: “If we don’t act today, we’ll be obsolete in three to five years.”
The “Genetic Replacement” Battle:
First, Reshape the Boss
In March 2006, Zhang Jian proactively reached out to Feishu and launched a pilot within a subsidiary of the group.
But he did something radically different from most corporate leaders:
Instead of delegating the task to the IT department, he immersed himself first.
During the first two months of transformation, Zhang ranked first in the company for AI token consumption. He personally mastered every tool, every system, every line of code, and every technical detail.
“Why did I dive in myself? It’s simple. If the top leader doesn’t get hands-on and build real intuition, this effort simply won’t move forward. If you yourself can’t use it or explain it clearly, how can you expect your team to believe?”
Zhang summarized his personal experience with AI into three elements:
First, eyes—information transparency.
In the past, understanding frontline operations required层层 questioning and waiting for reports. Now, all business data and customer communication records are centralized on Feishu’s platform.
“The biggest change isn’t speed—it’s peace of mind,” Zhang added.
“I can access the information I need instantly, without going through anyone or any hierarchy.”
By simply opening Feishu’s smart Q&A and asking, “How is a particular business performing recently?” the system delivers answers in seconds, clearly citing sources—from meeting minutes on a specific date, to a salesperson’s client notes, to shipment data from a specific batch.
For the first time, AI allows a CEO to understand their company in near real-time.
Second, hands and feet—offload trivial tasks to AI.
The most typical example is interviewing.
“In the past, I could barely interview a few people a year because preparation took too long,” Zhang said. Each resume required careful reading, and interview questions needed tailoring to each candidate’s background—an extremely time-consuming process.
Now, AI helps him in three ways:
Automatically generating targeted interview questions based on candidate profiles and job requirements;
Providing scoring criteria—what constitutes a good versus poor answer;
After the interview, automatically generating match assessments and recommendations for second rounds based on intelligent meeting summaries.
“In one month, I’ve interviewed more people than in all previous years combined. And the quality is higher—I make better hiring decisions.”
Zhang emphasized that hiring is a CEO’s top priority. “We always said we didn’t have time to review resumes or conduct interviews, but AI has made this much easier.”
Third, an advisor—helping the boss fully develop ideas.
This may be the value Zhang values most.
“CEOs often make decisions alone, without anyone truly helping them think ideas through.”
But now, it’s different. Zhang shared a scenario: he frequently has “flash-of-insight” moments—say, suddenly envisioning a new product direction. In the past, turning this idea into action required coordinating multiple departments for research, only meeting again one or two weeks later.
Now, here’s what he does:
Open AI and spend five minutes verbally outlining the idea. AI responds by asking him 10 to 20 follow-up questions: “Who are the competitors? Who’s the target audience? What’s the pricing and channel strategy?”
Some he can answer immediately; others he hadn’t considered. AI honestly records the gaps, then within 15 minutes automatically pulls public data across the web to generate a complete report covering competitors, pricing, channels, and risk assessment.
“Previously, I’d schedule meetings to brainstorm ideas. Now, AI first helps me fully flesh out my thinking, and then I bring a semi-finished proposal to the team. The efficiency difference is not even comparable.”
Zhang defined this “genetic replacement” as follows:
“AI didn’t just add a skill—it equipped me with eyes, hands and feet, and a permanently online advisor who dares to ask anything. Only then could I free up time to focus on studying markets, users, and future directions.”
The “Capability Reconstruction” Battle:
The Challenge Isn’t Changing Tools, But Changing People
But transforming just the CEO isn’t enough.
Zhang soon realized the real battle was at the organizational level.
“After introducing the new tools, we asked each department to re-map their business flows. As it turned out, some couldn’t even clearly articulate their own past workflows.”
Even trickier was strong “passive resistance” from some middle managers:
They didn’t refuse outright, nor did they actively engage—progress was extremely slow.
“Talk to them, and they agree with everything. But when it comes to action, they find excuses—‘I have a client meeting today,’ ‘I’m traveling tomorrow,’ ‘I’m swamped right now,’ ‘We’re already making progress’… In short, no time to map workflows, no time to learn.”
Zhang pinpointed the root cause: “AI eliminates information asymmetry and makes decision-making fully transparent. Middle managers who previously survived as ‘information gatekeepers’ suddenly realize their core competencies are no longer sufficient.”
This is the most genuine—and difficult—hurdle in AI transformation. It’s not a technical barrier, but one of mindset, power, and sense of relevance.
Many companies would choose to “replace personnel” at this point, but Zhang did not.
“I don’t believe a manager with years of experience has no value,” he said. “Tell me, can a product manager with ten years in manufacturing really be replaced? Their tacit knowledge, lessons learned from failures, and intuitive grasp of supply chains—these aren’t something large models can generate from scratch. These experiences are invaluable. What we lack is a way to turn that experience into reusable organizational capabilities.”
Zhang chose not to discard people casually, but instead actively help them reconstruct their capabilities—enabling individuals within the organization to leverage AI to multiply their effectiveness ten- or hundredfold and complete their transformation.
He took two key actions:
First, establishing a “powerful AI pioneer team.”
This team wasn’t a support function, but one led by the CEO with authority over process design and performance evaluation. Their mission: dive into every department to help managers and teams who want to transform but don’t know how—breaking down business flows, building systems, and training skills.
“As long as you’re willing, even if you don’t know where to start, we’ll provide support. For those less motivated, we’ll keep the organization moving forward—no one wants to fall behind, so they’ll naturally try to catch up.”
Second, using “strong enforcement” to drive middle managers’ role transition.
He mandated that all work must go through AI first—“AI first.” Before submitting reports, run your ideas through AI for challenge; before drafting proposals, let AI fill in blind spots.
“I tell the team: once you combine your existing skills and experience with AI, you’ll feel ‘omnipotent’—as if nothing is truly difficult anymore, and your work grows increasingly professional. This isn’t preaching—we’ve seen managers reach in four months a capability level that previously took three years.”
Then he added: “Of course, some people persist in using ‘don’t understand, can’t learn, no time’ as excuses. These managers will eventually be naturally identified by the system.”
The “Organizational Evolution” Battle:
From ‘One Person’s AI’ to ‘the Entire Organization’s Foundation’
When the toughest ice began to crack, a top-down transformation unfolded naturally.
Within the health group, there’s a frequently cited case: the “Multi-Dimensional Integrated Operations Platform.”
This “office middleware” consolidating all business entry points, data dashboards, and approval workflows wasn’t developed at great cost by an external team—it was built in-house by the AI Pioneer Team together with business staff using Feishu’s multi-dimensional tables.
Simulated data
In the past, even a minor system update required scheduling with software vendors and submitting requirements—waiting one to two months was routine.
Now, frontline employees who best understand operational pain points can build lightweight applications themselves in one or two weeks, refining them iteratively as they go.
“This is the fundamental difference between Feishu and many tools we’ve used before,” Zhang said.
Many past office tools operated under a control model—check-ins, approvals, task assignments—the core logic being “management.” But Feishu is a collaboration platform where information reaches peers directly under shared permissions, eliminating layers of transmission.
“I don’t want the organization to operate under control mode. Control is highly inefficient in today’s world, breeds laziness, and is meaningless. I want full data transparency, ultra-short decision pathways, and highly efficient collaboration—truly empowering frontline soldiers who ‘hear the gunfire’ to ‘command’ the battle.”
The deeper change lies in data governance logic.
Zhang used a vivid analogy: Excel and Feishu’s multi-dimensional tables represent entirely different application mindsets.
Excel operates on a “single-file, closed-source” logic—a complex spreadsheet is readable and editable only by its creator; anyone else touching it risks chaos.
Feishu’s multi-dimensional tables, however, follow a “unified data source, multi-view access” logic—all data resides in a single underlying database. The CEO views customer dimensions, sales views product categories—each sees their own view, without interference.
“This is called ‘de-formularization.’ In the past, the company depended on that one expert who could handle spreadsheets—when they went on vacation, the whole company waited. Now, data and capability reside within the organization itself.”
This shift enables a complete restructuring of organizational collaboration—from “serial” to “parallel.”
Previously, teams waited—waited for tasks, upstream inputs, notifications, meetings, emails. If upstream stalled, downstream froze.
Now, monthly goals and departmental plans are fully visible in multi-dimensional tables. On logging into Feishu each morning, all work entry points appear on a single page. No need to be reminded—systems prompt you. When a step gets stuck, everyone sees it immediately.
Zhang gave a concrete example: tracing a customer complaint across the entire chain.
In the past, resolving a customer complaint required assembling multiple departments and took several days. After AI integration, data ranging from raw material batches to production parameters, quality inspection reports, and shipping times can be retrieved instantly. AI even performs initial diagnostics, reducing response time from “days” to “minutes.”
“Previously, discussing a customer request involved contacting someone, who then contacted others—taking at least three days even at best. Now, the team delivers solutions within two hours. Customers perceive us as exceptionally professional and fast. In today’s hyper-competitive market, this is how you build trust and create barriers.”
Conclusion:
You Either Use AI to Outcompete Others,
Or Wait for Others to Use AI to Outcompete You
Toward the end of the interview, I asked Zhang Jian one final question: Do you still have AI anxiety?
He laughed: “Long gone. My only anxiety now is that people aren’t using it fast enough.”
He paused, then added:
“In today’s world, either you use AI to outcompete others, or you wait for others to use AI to outcompete you. I don’t want to be outcompeted—I want to be the one taking initiative.”
Reflecting on his AI transformation journey, Zhang offered several practical suggestions:
The top leader must personally dive in. “For the first two months, I ranked first in token usage across the company—I had to gain firsthand experience. Don’t delegate this to a subordinate unless they possess your highest authority and strategic understanding.”
Start with small, painful operational issues. “Don’t aim to solve big problems right away. Begin with customer service, proposal drafting, or data statistics—quick wins build momentum and intuition. Avoid launching with grand strategies.”
Don’t approach tools with skepticism. “I believe AI is 100 times smarter than us. If AI output seems flawed, first consider whether your usage method is wrong.”
Don’t stress over tool iterations. “Feishu understands AI far deeper than we do. I suspect their AI anxiety is worse than ours. Let them handle technological evolution—we focus on how to apply it.”
Fifty-eight years ago, his father Zhang Guofang took over a factory on the brink of collapse and built it into a global leader—by doing one thing thoroughly, transforming seaweed from being sold “by the ton” to “by the gram.”
Fifty-eight years later, Zhang Jian continues his father’s journey. He no longer faces existential crisis, but a deeper challenge: how can a large ship avoid falling behind in a new era?
His answer: give this 58-year-old enterprise a new organizational “gene.” Not waiting until the market pushes it against the wall, but embracing advanced technologies while still ahead—every industry can be reinvented.
Tools will continuously evolve, but what ultimately determines a company’s future is its ability to keep evolving.
From a single strand of seaweed to a billion-dollar industry, from the previous generation’s pursuit of perfect products to the current generation’s organizational reconstruction—perhaps the story of Mingyue Seaweed is answering a crucial question:
In the AI era, what enables China’s “hidden champion” enterprises to keep winning?