AI's entry into enterprises is not difficult due to the availability of tools, but rather due to whether it can truly take root and become operational.
At the 2026 Feishu Future Unlimited Conference and Doubao Work launch event, Feishu Chief Commercial Officer Lin Chan emphasized: through empowerment in four key areas—management, knowledge, processes, and talent—Feishu and Doubao Work help enterprises move AI adoption from isolated breakthroughs toward systematic deployment.
Below is the full transcript of her speech.
An individual who skillfully uses AI becomes an "AI Pioneer"
Friends who know me are aware that I usually come to give gifts at every product launch. Today is no exception—I’ve prepared three major gift packages for the finale.
However, what most of you care about is how Feishu helps advanced enterprises achieve AI transformation. I will share our related experiences so that after listening, you’ll clearly understand how to use AI effectively and well.
Earlier, while sitting in the audience listening to presentations, one point particularly resonated with me: those who best understand business pain points should be the first to adopt AI tools.
A few years ago, this scenario was actually quite rare.
In the past, when we visited enterprises, the most common conflict we saw was: business teams urgently requested solutions, while IT teams couldn't schedule development in time. By the time systems were finally launched, business needs had often already changed. Now, however, many "AI Pioneers" are emerging directly from frontline business roles. They not only use AI to solve practical problems but also share their effective strategies with colleagues around them.
Since 2023, Feishu has partnered with 792 enterprises to cultivate and certify nearly 90,000 "AI Pioneers."
Besides, nearly 1,800 job postings across approximately 400 companies include the same statement: priority given to applicants holding the Feishu AI Pioneer certificate.
This year, we have further upgraded the "AI Pioneer Competition," extending it directly to universities worldwide.
Feishu has joined forces with 108 partner companies to provide real-world challenges, senior mentors, and actual job offers, aiming to recruit future AI talent. Currently, over 4,000 teams from around the globe have entered the competition, with the grand finals taking place next month. We invite all company leaders to attend with your "Boss Fast Pass" and scout top talents.
Therefore, through the "AI Pioneer Competition," we already have a clear answer on how individuals can effectively use AI. Yet these individual practices led us to our next question: if every employee uses AI, does that mean the enterprise has successfully realized AI’s value? Allow me shortly to explain using an analogy.
For enterprises to use AI effectively, careful design is essential
An individual using AI is like driving a car—only needing to control the accelerator, brake, and steering wheel.
But an enterprise using AI is akin to managing an entire city's transportation network—requiring lanes, traffic lights, and speed limits. In other words, no matter how skilled the drivers are, they cannot single-handedly build a complete transportation system.
At its core, individuals using AI save their own time; enterprises using AI gain overall efficiency. However, there is often a gap between individual output and enterprise results, typically stuck on four types of issues.
In management, if leadership hasn’t personally adopted AI and merely delegates tasks to others; in knowledge, although many enterprises store vast amounts of data neatly in systems, this data remains inaccessible to people or AI; in processes, even though AI can efficiently complete certain steps, few consider who takes responsibility afterward; in talent, outdated standards are still used to evaluate new-era talent, making recruitment and retention difficult. Next, let's address each of these four challenges one by one.
Feishu services help enterprises advance AI from isolated successes to systematic deployment
First, let's discuss the management issue:
Over the past year, we've held face-to-face discussions with over 20,000 enterprise executives on the same topic—"how enterprises can implement AI"—including more than 6,400 founders or CEOs. I’d also like to ask the bosses here today: how many of you have personally built an Agent? Please raise your hands.
In traditional classrooms or large conferences, we usually assess effectiveness by "head-up rate." But in Feishu's AI workshops, we focus on the "head-down rate." A room full of executives running businesses worth tens or even hundreds of billions of dollars are all heads down, using "Doubao Work" to package Skills and execute workflows, immediately practicing solutions for their own company's real problems. The excitement many feel upon completing their projects resembles the thrill of beating a video game level.
One participant once told me specifically: "A good memory is no match for notes, and notes are no match for personally running a workflow. This approach is far more effective than simply listening to lectures as before." Many executives leave the course and immediately mobilize their teams into action—one such example being Wang Shoucheng, CEO of Yonghui Superstores.
Looking back over the past year, Mr. Wang spent roughly 300 days touring stores. The essence of retail isn’t just reflected in financial reports—it lies among shelves and aisles. Previously, he would walk and talk while store managers took notes. Yet issues spotted during one visit weren’t necessarily corrected the next time, and successful practices from one store rarely reached others. After attending the Feishu AI workshop, he and his team accelerated collaboration and developed an inspection Agent named "Xiaohui Store Visit." Initially designed to review revenue, profit, and loss data before entering a store, Mr. Wang quickly identified a logical flaw—he pointed out that store inspections should start with understanding the person, then the store.
Now, during inspections, sending a location to the "Xiaohui Store Visit" Agent instantly retrieves the store manager's profile: tenure at Yonghui, hometown, whether they’ve received bonuses. Thus, from the moment he enters, he can call the manager by name and chat casually, making them feel warmly acknowledged. During the visit, Mr. Wang wears the "Feishu Voice Recorder Bean," connected to "Xiaohui Store Visit," enabling real-time communication with the store manager. When issues arise, he takes photos and sends them instantly to the Agent. Within ten minutes of leaving the store, a report is generated, and corrective tasks are automatically assigned to specific roles. If he spots an excellent practice, saying “can be promoted nationwide” prompts "Xiaohui Store Visit" to generate a learning card and push it instantly to hundreds of store groups with one click.
Therefore, management empowerment means enabling leaders to personally operate AI and lead organizational transformation.
Having addressed management, let’s now examine the second challenge—knowledge.
Many enterprises don’t lack data or experience—they lack the ability to truly utilize these “core assets.”
At a dairy farm in Hebei operated by “Adopt a Cow,” a post-00s information officer faced the farm’s most pressing summer challenge—heat stress. Simply put, cows suffer heatstroke too, and milk production drops as temperatures rise. Back then, he had to visit 38 zones and physically observe the health of 7,000 cows. Yet the farm wasn’t lacking in data—over 1 million health records sat dormant across five different systems, difficult to extract. One million data points left unused equaled one million observations wasted.
After graduation, this young man returned to the farm as a full-time employee. Using Feishu, he integrated data from the five systems and upgraded the existing notification bot into an Agent colleague via "Doubao Work Companion," directly linking indicators for "breathing, rumination, and feeding" to detect cow heatstroke. Now, the health status of all cows can be monitored from a smartphone. Once anomalies are detected, the Agent immediately alerts veterinarians.
Reflecting on this experience, this post-00s employee once said something memorable: "Combining Feishu with AI allowed a research topic from my graduation thesis to find a concrete solution in my daily work."
In reality, transforming data from idle system entries into actionable business knowledge requires two things:
First, deeper accumulation of knowledge. Before adopting Feishu, internal collaboration tools in many enterprises were used for only 10–20 minutes per day on average. After adopting Feishu, this increased to over 200 minutes, with some reaching 300 minutes—meaning more than 10 times the amount of business knowledge is now recorded in real time.
Second, optimized knowledge retrieval experience—not only convenient for humans, but also directly accessible by AI. Over the past six months, enterprise AI Agents have retrieved knowledge within Feishu over 40 billion times, with available documents, chat histories, and other knowledge assets exceeding 100 billion items.
Thus, the goal of enterprise knowledge empowerment is to make organizational knowledge dynamic and flowing, serving as the basis for decisions made by both people and AI.
Next, let’s discuss process-related issues.
AI must penetrate specific business processes and genuinely assist employees in their work.
At Nanjing Iron & Steel’s heat treatment workshop, a veteran technician is set to retire this December. Every morning, he must manually verify processing specifications for over a hundred steel plates on his computer. Since steel plates vary in thickness and require different treatments, any change in the production plan means starting over. He said that after riding home each day, his eyes are blurry, and he still worries about misreading data and causing accidents before falling asleep. Later, a young worker in the workshop—though unfamiliar with coding—used Doubao Work to iterate 140 versions over seven days, creating an Agent colleague for the veteran. After the Agent went live, the process was cut in half. The technician uploads the shift details, and the Agent outputs the required specifications with one click. Verification accuracy improved from 85% to 98%. Query time dropped from three hours to under one minute, saving over 700 hours annually.
The veteran uses this Agent daily to verify and correct inputs, gradually embedding decades of experience into it. As mentioned, although he will retire at year-end, his expertise has already been embedded into the production process via the Agent, preserving craftsmanship at the role level. Moreover, this is not an isolated case at Nangang—over 1,500 Agents are now operating within business processes.
Hence, empowering enterprise processes means integrating experience into workflows, ensuring smooth handoffs and stable operation of Agents.
Finally, let’s address talent.
Our talent empowerment doesn’t merely involve letting employees casually install AI software to improve PPT aesthetics or write weekly reports faster—such time savings rarely translate into enterprise compounding benefits.
We partnered with Perennial Beauty, a domestic cosmetics company, to develop a solid AI talent empowerment mechanism: executives enter classrooms—managers bring real business problems to learn AI hands-on; employees join competitions—over 1,500 participated in the "Feishu AI Pioneer Competition," using competition as training to solve numerous business challenges; technical and business teams pair up—one side builds AI Agents, the other validates effectiveness; we offer real lab challenges and direct access passes to students, encouraging young people to make their "first appearance."
The result: AI has transformed job responsibilities and reshaped talent capabilities.
So how should enterprise talent standards and job responsibilities adapt to the AI era?
This is part of Feishu’s service offering. We dive deep into our clients’ frontline operations, first identifying key positions and breaking down their manuals, responsibilities, and performance requirements for detailed review. Then, we accompany employees throughout a full workday, tracking their activities from morning to night. Through this, the true nature of the role becomes fully visible.
On average, analyzed roles involve over 100 distinct tasks. Take the position shown on screen—every task can be supported by AI. Ultimately, we gradually restructure: delegate transactional, inefficient tasks to AI, shifting human focus toward judgment, collaboration, and value creation. Many executives joke that this planning is clearer than their own undercover visits.
Therefore, empowering enterprise talent means redefining role requirements so everyone can thrive in the AI era.
By now, the four challenges discussed are clear: management leads, knowledge lays the foundation, processes build the structure, and talent sustains capability. Only when these four elements integrate can enterprises truly harness AI effectively.
Speaking of which, recently a term frequently associated with AI is FDE. Finding high-quality FDEs on the market is costly and difficult. But looking back, Feishu’s team—deeply embedded on the frontlines, long-term stationed, co-building AI value with enterprises—is precisely the FDE that best understands Chinese enterprises!
Finally, let’s talk about benefits
Now, onto the benefits. The first is sharing the accumulated experience and methodologies from this coaching team over the years.
Reply “service” in our official account to receive it.
This Feishu AI Service Map includes practical strategies across the four domains—management, knowledge, processes, and talent—as well as Agent scenarios already implemented across industries, plus ready-to-use Skill toolkits.
The second benefit is Doubao Work membership. If you haven’t tried it yet, download the desktop version now to claim one month of free membership.
The third benefit: we’ve prepared 100 AI Boost Gift Packages, each valued at RMB 99,000, awarded to the top 100 enterprises advancing to the finals of the ten industry tracks in the "2026 AI Pioneer National Competition." We look forward to seeing you on the national stage this winter, bringing innovations capable of transforming your industries.
Doubao Work gives Agents practical capabilities
Feishu enables people and Agents to collaborate seamlessly
Feishu Services help enterprises continuously achieve great things
Benefits may be rare, but great service endures. That concludes today’s discussion on services. Thank you all!