In-depth research: A global enterprise-level knowledge management AI core player

In an era where AI is reshaping enterprise efficiency, knowledge management is no longer just a “document archiving” but a core asset of organizational intelligence. Who is leading this change? Who is quietly defining the future of knowledge work? This article focuses on the global enterprise-level knowledge management AI track, deeply analyzes the product strategies, technical paths and implementation practices of several core players, and takes you to see the pattern and direction of this “cognitive infrastructure” battle.

Today’s protagonist is: enterprise-grade knowledge management AI.

There are many unicorns in the global enterprise knowledge management AI space, and Glean is the largest of them, and Writer is also a unicorn in this track.

There are many enterprise-level knowledge management enterprises in China, including AI upgrades from traditional knowledge management companies (such as Lanling, Daoke, etc.), AI-native intelligent knowledge management platforms (such as LMU.AI, Hongyi, Lenovo Fifilez, etc.), and supporting solutions launched by giant ecosystems (such as Tencent Lexiang).

What I want to share with you today is the track of this tens of billions market. I will sort it out according to the following logic, the content is relatively long, you can jump directly to the part you are interested in:

1. What is enterprise-grade knowledge management AI?
2. Global & China Enterprise Knowledge Management AI Market Size
3. Global & Chinese core player analysis
4. Market demand characteristics and opportunity points
5. Summary and enlightenment

01 What is enterprise-level knowledge management AI?

Imagine a new employee who can’t get started with the company’s vast document library and simply asks in natural language: “How do I handle a customer’s XX product return request?” ”

OneSmart assistantInstantly extract information from different documents such as contract templates, customer service records, logistics policies, etc., to generate complete how-to guides. This is the daily scenario of enterprise knowledge management AI.

After 10 years of interaction design, why did I transfer to product manager?
After the real job transfer, I found that many jobs were still beyond my imagination. The work of a product manager is indeed more complicated. Theoretically, the work of a product manager includes all aspects of the product, from market research, user research, data analysis…

View details >

Unlike traditional knowledge management systems, this type of AI is not onlyStorage tools, is even more availableComprehension abilityandMobilityAgent. The core capabilities of enterprise knowledge management AI are reflected in three levels:

1. Knowledge comprehension: Analyze unstructured data in contracts, emails, and even meeting minutes through natural language processing technology, and “read” the content like a human.
2. Knowledge connection: Establish an association network between knowledge points, such as automatically associating product fault descriptions with solutions, responsible person information, and historical cases to form a knowledge graph;
3. Knowledge application: Transform “sleeping” knowledge into instantly available decision support through intelligent Q&A, automatic document generation, knowledge recommendation, etc.

Compared with the traditional system that only supported keyword search in the early days, the new generation of AI knowledge management platform has itSelf-directed learning ability。

Taking a medical device company as an example, when the product is updated, the system automatically identifies the difference between the old version of the manual and the new version, reminds the update of the knowledge base, and synchronously notifies relevant personnel to achieve “self-maintenance and update” of knowledge.

02 Panorama of the global and Chinese markets

Global Market Size and Growth Rate: According to IDC’s latest Global AI Knowledge Base Market Report, the enterprise knowledge base market is expected to exceed $42 billion in 2025, with a compound annual growth rate of 37%. The main drivers come from:

• Technological reconstruction: large models promote innovation in knowledge production methods (such as increasing the efficiency of clinical diagnosis and treatment plan generation by 40%+);
• Security requirements: hybrid architecture (on-premises + cloud) to meet the data isolation needs of manufacturing and financial industries (ISO27001 certification is standard);
• Decision-making efficiency: the average reduction of the strategic decision-making cycle of enterprises is 62% (McKinsey research);

In terms of regional distribution, North America dominates: Accounting for more than 40% of the global share, leading in applications in the financial and medical fields (such as Salesforce Einstein AI integrated CRM knowledge base);Europe accelerates: GDPR compliance drives localization deployment, with year-on-year growth32%； Asia-Pacific exploded: Market size forecast in 2025$12 billion, the world’s fastest growth rate (CAGR65%China, Japan and South Korea are the core markets.

China Market Characteristics and Forecast: The Chinese market size is expected to reach 202512 billion yuanincrease speed65%Significantly higher than the global level, the scale of the generative AI software market in 2025$3.54 billion(IDC), of which knowledge management tools account for more than 30%. The main drivers come from:

• Policy traction: The “intelligent transformation and digital transformation” policy promotes the procurement of domestic systems of state-owned enterprises (such as the localization rate of more than 60% in the government sector); The Data Security Law requires local storage of sensitive data (10+ years deployment cycle in the financial and medical industries);
• Enterprise cost reduction needs: manufacturing focuses on cost reduction (such as Sinoma International R&D cost reduction of 34%); improve government efficiency (the efficiency of official document review in Futian District, Shenzhen increased by 90%);

Due to the requirements of the Data Security Law, it is difficult for foreign knowledge management AI unicorns to enter China, and with the outbreak of domestic large models, domestic enterprises have huge room for development. Specifically: Chinese NLP optimization has become the key to competition; Multimodal AI accelerates penetration, with medical imaging and industrial quality inspection as the core scenarios.

03 Market pattern and player distribution

Global players

1. Glean: 2019, California, USA, valued at $7.2 billion

Core competencies：

1. enterprise-level AI search platform, integrating 100+ SaaS applications to build a knowledge graph;
2. Multimodal RAG technology supports natural language retrieval and permission management;
3. AI agents automate multi-step tasks (e.g., generate reports, schedule meetings);

Represent the customer: Deutsche Bank, Sony, Reddit, Deutsche Telekom;

Differentiation：

Semantic understanding replaces keyword search with real-time workflow automation;

Strict permission control to display only content that users have access to;

High-frequency use: 10 queries per user per day, DAU/MAU reaches 40% (far exceeding the industry average);

Usage scenarios：

Sales plan generation: New employees enter customer names, aggregate historical contracts, technical documents, and competitive product analysis within 1 minute, automatically generate customized plans, and recommend relevant experts.

2. Hebbia: 2020, Los Angeles, USA, valued at $2.8 billion

Core competencies：

1. The AI agent “Matrix” processes massive amounts of unstructured data (PDF, audio, video);
2. Supports billion-level document indexing and cross-file correlation analysis;

Represent the customer: U.S. Air Force, top hedge funds, large law firms;

Differentiation：

Long document processing ability: can analyze complex data such as documents listed on the stock exchange;

Crisis response: Silicon Valley Bank quickly mapped the risk exposure of regional banks during the crisis;

Usage scenarios：

Financial compliance analysis: Asset managers use Matrix to scan millions of regulatory documents and automatically generate risk reports, reducing time from weekly to hourly.

3. Alation: In 2012, California, USA, raised more than $300 million

Core competencies：

1. Data cataloging and metadata management, building enterprise data graphs;
2. The behavior analysis engine tracks the trajectory of data usage to improve data credibility.

Represent the customer: Pfizer, Cisco, Munich Re;

Differentiation：

Data lineage visualization: algorithms similar to Google PageRank evaluate the value of data;

Open interface: integrate mainstream tools such as Teradata and Tableau;

Usage scenarios：

Pharmaceutical R&D: Pfizer manages clinical trial data through Alation, allowing researchers to quickly locate historical experimental parameters and reduce repeat trials by 30%.

Chinese player

4. Lanling Software: In 2001, Shenzhen, in 2018, Alibaba DingTalk received hundreds of millions of yuan in strategic investment

Core competencies：

1. knowledge management platform aiKM + Alibaba Cloud’s “Tongyi Qianwen” large model;
2. contract risk control scanning, intelligent question and answer engine;

Represent the customer: CITIC Group, China Merchants, Xiaomi, OPPO;

Differentiation：

1. Localization adaptation: 60% market share in the government affairs field;
2. Deep cultivation of the manufacturing industry: shortening the fault query time for Cialis Automobile by 70%;

Usage scenarios：

Manufacturing knowledge base: Engineers take photos and upload equipment fault maps, and the system automatically matches maintenance manuals and pushes historical cases, increasing maintenance efficiency by 50%.

5. Tencent Lexiang: Incubated internally in 2008 and opened to the public in 2017

Core competencies：

1. One-stop corporate community (knowledge base, online classroom, Q&A community);
2. Become an enterprise WeChat to support party building, training and other scenarios

Represent the customer: Rainbow shopping mall, Yunnan Baiyao, IKEA China

Differentiation：

C-end experience porting: seamless access to the WeChat ecosystem;

Multi-industry template: 40% increase in training efficiency for retail employees;

Usage scenarios：

Retail training: Rainbow Shopping Mall pushed product knowledge videos to new employees through online classrooms, and the pass rate increased from 65% to 92%.

6. Lenovo Filez AI: 2006, internal product line

Core competencies：

1. “File + content + knowledge” full-chain management; Enterprise network disk and online document collaboration;

Represent the customer: Government, finance, education leading customers;

Differentiation：

Hybrid cloud deployment: Supports flexible switching between localization and cloud.

First market share: 2020 IDC report ranks first in the market share of Chinese enterprise network disks;

Usage scenarios：

Engineering collaboration: The architectural team edits CAD drawings across regions, automatically synchronizes versions, and shortens the project delivery time by 25%.

7. AnyShare: Shanghai, 2011

Core competencies：

1. Unstructured data center supports unified management of multiple document domains. OCR recognition and automatic content classification;

Represent the customer: Finance and smart city projects (such as a provincial government cloud);

Differentiation：

Content data lake architecture: cataloging and labeling of massive unstructured data;

Industry compliance: meet the requirements of Classified Protection 2.0 and GDPR;

Usage scenarios：

Government document management: Provincial archives automatically identify millions of paper files through AnyShare, establish electronic indexes, and reduce the file search time from 3 days to 10 minutes.

From the core players in China and the United States, it can be seen that global enterprise knowledge management AI presents two major paths:

• American companies (such as Glean, Hebbia) are known for their cross-system aggregation + complex analysis, and are good at unstructured data value release (Hebbia) and human-computer collaboration closed loop (Glean), focusing on the financial and technology industries.
• Chinese companies (such as Lanling and Tencent Lexiang) are deeply involved in industry scenarios + localization adaptation, focusing on government affairs and manufacturing.

Although the implementation paths of enterprise knowledge management AI in China and the United States are different, the ultimate goal is the same: to build a set of enterprise knowledge closed loops that can “find→ understand, understand, → apply”.

04 Market demand characteristics and opportunities

European and American market characteristics:

1. High enterprise maturity: About 42% of large enterprises have implemented AI, half of which are part of a search/content management system;
2. Focus on semantic retrieval and content summary: natural language query, knowledge graph, and question and answer systems are widely used;
3. Compliance and trust first: security, permission control, and audit logs are standard requirements;

European and American market opportunities:

1. Customized services for large enterprises: deep cultivation of vertical industries can create high-value products;
2. Evolution to Agent: Retrieval + Analysis + Execution is the future direction;
3. Cross-system integration barriers: The ability to unify multiple SaaS applications will become a long-term barrier;

Chinese Market Features:

1. Industry pain point orientation: manufacturing (such as Cialis Automobile) focuses on equipment failure knowledge base, and Lanling software shortens maintenance query time by 70%; government affairs (such as Futian District, Shenzhen) require the efficiency of official document review to increase by 90%;
2. Government leadership and policy tilt: The data security law mandates localized deployment, and the proportion of state-owned enterprises purchasing domestic systems exceeds 60%;

Opportunities in the Chinese market:

1. Industry scenario-driven: there is an urgent need for knowledge integration in manufacturing, finance, government affairs and other fields;
2. Policy support dividends: government subsidies and standard formulation provide soil for industry applications;
3. Cost and speed advantages: Localized deployment, AI model computing power cost-effective advantages are outstanding;
4. Talent dividends appear: the number of AI practitioners in China is as high as 2.2 million, and education and training are advancing rapidly;

Table: Comparison of the core differences in knowledge management AI market demand between China and the United States

Common Opportunity Points:

1. The Agent era is coming: the transition from “information retrieval” to “knowledge organization and drive” mode;
2. Strong vertical industry opportunities: Professional scenarios such as finance, manufacturing, and government affairs have a strong demand for fine knowledge management;
3. Growth space for small and medium-sized enterprises: lightweight and low-threshold products can accelerate popularization;

05 Summary and enlightenment

The enterprise-level knowledge management AI track is in a critical period of explosive growth and technology iteration, as can be seen from a number of unicorn companies in the United States:

• The knowledge management AI track is expected to build high-growth, scalable AI products in China, and should focus on the needs of large enterprises in the early stage, while laying out the future medium/small and medium-sized enterprise market.
• The evolution of AI from information retrieval to task execution is an inevitable path, and entrepreneurs should pay attention to “action” capabilities, multi-system collaboration and contextual closed-loop design.
• AI products need to be designed with enterprise-level architectures that support private deployment and multi-tenant governance to win the trust of large customers.
• First, deeply cultivate a certain vertical track, build professional models and indicators, deeply integrate with industry processes, and then expand horizontally.

Above, I wish you all a happy day.