Enterprise software's "major crossover"

The boundary of enterprise software is shifting from product categories to task chains. Behind customers' purchases of CRM, HCM, customer service systems, or BI, what they truly need is a business path that can integrate goals, execution, and feedback.

AI Agents will first transform lightweight scenarios. Department - level applications, temporary processes, lightweight CRM, project tracking, recruitment progress, and work - order transfer will all be diverted by low - code, AI tables, and customers' self - development capabilities.

Complex systems will not be easily replaced. Permissions, data, security, business semantics, and responsibility chains remain the deepest barriers for enterprise - level software. AI has lowered the threshold for software production, but it has not lowered the responsibility threshold for enterprise - level software.

Platform vendors will compete for entry points again, and vertical SaaS providers need to move towards the professional capability layer. Entry points may be taken away by platforms and Agents. Vertical vendors need to transform industry know - how into executable, verifiable, and accountable capabilities.

The key to truly changing the landscape lies not in functional overlap, but in the migration of core processes and budgets. To determine whether a vendor has truly crossed boundaries, we need to see if it has entered the core customers, core budgets, and core business processes of the other party.

Traces of the changes in the enterprise software industry can be seen in partnerships.

In the past, the CRM vendor "Fenxiang Xiaoke" and the customer service vendor "Tianrun Rongtong" were collaborators on the same customer chain.

When Fenxiang Xiaoke encountered customer service, call center, or contact center requirements at the customer site, it often transferred business opportunities to Tianrun Rongtong. CRM manages sales and customer operations, while customer service vendors manage service reception and contact centers, with clear boundaries and smooth cooperation.

Now, this boundary is starting to loosen.

Niutoushe has observed that Fenxiang Xiaoke is extending its product line to service cloud, work - orders, on - site services, intelligent customer service, and service Agents.

Similar actions are also taking place among more vendors.

Weimeng and Youzan are developing AI shopping guides and AI customer service. HR vendors are providing sales coaching and leadership training. Learning platforms are entering the fields of sales efficiency improvement and job empowerment. OA vendors are getting involved in contract, expense control, project, and asset management. Collaboration platforms are quickly building lightweight business systems using low - code, multi - dimensional tables, and AI assistants.

Everyone is extending to adjacent tasks around their customers. Moreover, the deeper change is that the task chains of enterprise customers are getting longer.

In the past, enterprises could purchase software by department. The sales department bought CRM, the customer service department bought call centers, the human resources department bought HCM, the training department bought learning platforms, and the finance department bought ERP or expense control systems. Each system handled a segment of the process, and data and actions flowed between departments. Although the efficiency was not high, it could still operate.

After the emergence of AI Agents, the software logic based on departmental division of labor began to loosen.

Customer operations are no longer confined to the sales system. It will cover lead identification, marketing reach, sales follow - up, quotation, contract, delivery, work - order, after - sales service, repurchase, and business analysis.

Employee development is no longer limited to the human resources system. It will cover recruitment screening, interview evaluation, on - boarding training, performance feedback, leadership training, job coaching, and career development.

Tasks will not stay within a single system. Can there be a more complete execution chain?

This is the root cause of the loosening of the enterprise software boundary.

Luo Xu, the founder and CEO of Fenxiang Xiaoke, believes that CRM cannot only focus on SFA (Sales Force Automation) or just be a sales cloud. Customers' marketing, sales, and service form a complete chain. It is difficult to provide end - to - end solutions by only handling one segment.

The same change is also happening in human resources software.

Beisen was originally understood as an integrated HCM SaaS vendor. Now, it is extending from HR business scenarios such as recruitment, personnel management, performance evaluation, and training to strong collaboration with business departments, serving AI interviewers, AI coaches, and AI leadership trainers for employee empowerment.

Enterprise software was previously divided by departments, but now it flows according to tasks. The loosening of the boundary is the prelude to a new round of competition.

Three forces are breaking the old boundaries

The boundary of enterprise software is being pushed open by three forces simultaneously.

Platform - type vendors are sinking from the main business operations to business details; vertical SaaS providers are extending from professional scenarios to adjacent tasks; low - code and AI tools are squeezing standard SaaS from lightweight requirements in the reverse direction.

Yonyou is a typical representative of platform - type expansion. It is no longer just a financial software or ERP vendor.

Public information shows that Yonyou BIP will release more than 40 intelligent agents covering ten fields such as finance, human resources, and supply chain in 2025, and emphasizes supporting the operation of intelligent agents with inference models and more than 4000 enterprise - level APIs. Yonyou BIP5 has also integrated AI into 14 types of end - to - end business processes, covering more than 2500 process scenarios.

Behind these numbers is Yonyou's intention to re - compete for the main entry point of enterprise operations.

In the past, ERP was too heavy and had a poor user experience, so many vertical SaaS providers were able to enter from single - point scenarios such as sales, human resources, procurement, expense control, and projects.

Now, if AI Agents can turn complex systems into natural language entry points and connect scenarios such as finance, human resources, supply chain, procurement, manufacturing, and marketing, the heaviness of the ERP platform may instead be transformed into an advantage. It has processes, data, the main business operations, and the entry point for customer budgets.

Kingdee is also taking a similar path. Kingdee Cangqiong AI is an AI management assistant for enterprise users, covering scenarios such as finance, human resources, procurement, and development, and providing different assistant capabilities for all employees, professional employees, and managers.

The pressure from platform - type vendors on vertical SaaS comes from the unified entry point.

Once a platform can cover finance, human resources, procurement, supply chain, management decision - making, and process approval, the boundaries of single - point systems will be compressed. Customers will also consider whether they still need to purchase a separate system for each department when most general needs can be solved within one platform.

Collaboration vendors take another approach.

DingTalk and Feishu are sinking from the office entry point to business systems.

DingTalk AI Table is positioned as an application building platform in the AI era, suitable for scenarios such as customer management, project tracking, and inventory statistics;

Feishu Multi - dimensional Table also emphasizes the ability to build tables and business systems, and highlights process, collaboration, data visualization, and AI capabilities.

These products may not replace in - depth CRM, HCM, ERP, or BI in large enterprises, but they will capture a large number of department - level, project - based, temporary, and small - and - medium - sized enterprise scenarios.

Many requirements that previously required the purchase of standard SaaS can now be built using tables, processes, automation, and AI assistants. For traditional software vendors, this part of the market may not be the most important source of income, but it is the foundation of customer perception and lightweight entry points.

OA vendors are also expanding along the process base.

Vendors such as Fanwei, Zhiyuan, and Lanling are extending from OA, processes, and knowledge management to fields such as contracts, legal affairs, expense control, projects, and assets.

Their advantages lie in processes, approvals, organizational permissions, electronic signatures, documents, and compliance. AI provides new interaction entry points for these process - based capabilities and makes the past horizontal expansion of OA vendors visible again.

The second force comes from vertical SaaS.

Xiaoshouyi's product line has expanded from traditional CRM to service cloud, partner cloud, intelligent analysis cloud, and Data Cloud.

This also shows that CRM vendors are moving from sales management to a more complete customer operation chain.

Fenxiang Xiaoke's expansion is carried out along the customer operation chain. Luo Xu emphasized in an interview with Niutoushe that Fenxiang Xiaoke defined itself as an integrated CRM for marketing, sales, and service as early as its transformation in 2016. Customer operations from Marketing to Sales to Service are a complete closed - loop.

Now, Fenxiang Xiaoke is turning this chain into an Agent - based system. Luo Xu summarized the future direction of SaaS as AaaS, that is, Agent as a Service.

Software should not only be designed for human use but also for Agents. Agents need to have autonomy, tool - calling ability, context - awareness, open combination, and cross - step collaboration capabilities.

Under this framework, Fenxiang Xiaoke focuses on building two levels of capabilities:

One is the platform, which provides inference capabilities, tool - calling, business memory, context memory, multi - dimensional connection framework, and Skill building capabilities;

The other is to build standard Agents around the three main scenarios of marketing, sales, and service, such as intelligence Agents, on - site Agents, quotation Agents, order Agents, and service Agents. Customers, implementation teams, and ecological partners can also build personalized Agents or Skills on the platform.

In a specific scenario, when a salesperson is preparing to visit an old customer, the Agent can first retrieve the customer's historical orders, contract status, service work - orders, recent complaints, payment status, and expiration time, and then generate the key points for the visit.

After the visit, the system automatically generates a summary, identifies the next actions, creates a service work - order, a quotation task, or a repurchase reminder if necessary, and pushes key information to the sales supervisor.

In this process, CRM is no longer just about recording customers and business opportunities. It begins to participate in customer operation actions.

This also explains why CRM vendors are entering the service cloud, work - orders, intelligent customer service, and on - site services. They do not think they are leaving their main business but are completing the customer operation chain.

CRM vendors have access to customer, business opportunity, contract, order, and payment data. Once they extend to after - sales and service, they will incorporate customer problem handling, engineer operations, after - sales delivery, and repurchase judgment into the same operation system.

Beisen's spill - over direction is from HR management to employee capacity building and business on - site.

Combining AI technology and more than 20 years of human resources experience, Beisen has expanded from an integrated HR SaaS to form a product matrix of AI Family, with a series of AI Agents such as AI interviewers, AI leadership trainers, AI coaches, AI recruitment assistants, AI performance assistants, AI learning assistants, and AI employee assistants.

Based on the research of "positions" and "people" in People Science and with the help of a consulting and service team of more than 200 psychology experts, Beisen helps customers implement more accurate AI Agents.

Ji Weiguo mentioned in an interview that Beisen previously had AI capabilities in products such as recruitment, performance, learning, and human resources shared centers.

Now, Beisen has created three relatively independent large Agents, namely the AI interviewer, the AI leadership trainer, and the AI coach.

The AI coach is the most obvious product of Beisen's spill - over to business departments.

Ji Weiguo gave several specific scenarios.

In the sales coaching of a leading jewelry brand, a couple goes to buy jewelry. The girl likes it very much, but the boy thinks it is expensive. The shopping guide needs to practice how to handle price objections and move the transaction forward;

A Western fast - food chain giant launches new products every month, and employees need to practice how to recommend new products to customers;

In the pharmaceutical industry, pharmaceutical representatives need to practice all aspects of drug conference access, practice department conference speeches, and how to present drug PPT solutions.

Beisen's AI spill - over strategy is based on strategic opportunity selection and the barriers of industry implicit knowledge.

In terms of commercial results, Beisen's AI products are leading the entire SaaS industry in commercialization progress. The company's overall AI revenue is expected to grow at a rate of more than 3 times. Just the AI interviewer alone has been implemented in nearly 1000 customers.

Hongquan uses AI coaching for product support staff training, allowing employees to practice how to introduce products to CSMs and salespeople in the system first, reducing the pressure on managers for repeated acceptance.

These scenarios originally did not belong to the core scope of traditional HR software. They occur in stores, hospitals, customer sites, and sales support processes.

But Beisen can enter these scenarios not because of its sales management ability, but because of in - depth research on the sales position model and the decomposition of human ability requirements into trainable, feedback - able, and reviewable behavior structures. Human capacity building begins to overlap with the business growth system.

Xuanxing is also expanding along its main channel.

Xuanxing (formerly Yunxuetang) started with enterprise learning. Now, it has shifted its strategic focus to enterprise intelligent productivity, and its product matrix covers directions such as AI base, talent management, sales intelligence, and learning and development.

This change has begun to be reflected in commercial results. In 2025, Xuanxing's NRR of subscribed customers increased to 101.4%; the MRR of AI - related products also increased from 300,000 yuan in the previous year to 1.1 million yuan.

This shows that Xuanxing's expansion is not simply moving from a learning platform to a sales tool, but extending to job capabilities and enterprise productivity along the main channel of enterprise learning and talent development.

The third force comes from low - code, AI Agents, and customers' self - development capabilities.

Liu Chengzhong, the founder of Hengshi, told Niutoushe that some enterprises have used AI to self - develop lightweight CRM to replace some functions of mature SaaS.

In his view, this will become a trend. Many enterprises only use a small number of core functions of SaaS products on a high - frequency basis. After AI lowers the development threshold, more and more lightweight requirements will turn to self - development.

This impact may not first manifest as the loss of large customers. It is more likely to first appear in department - level requirements, lightweight processes, and new project pilots.

In the past, enterprises would purchase SaaS for a few tables, a process, or a lightweight customer management module; now, Feishu Multi - dimensional Table, DingTalk AI Table, and low - code platforms can quickly undertake a considerable part of these requirements.

Liu Chengzhong's judgment on ChatBI also reminds us not to overestimate the significance of functional overlap. In his view, ChatBI is more like a general - purpose ability in software. Many SaaS products can add it, but this does not mean that they have truly entered the BI market.

Whether the industry landscape is changed depends on whether the vendor has entered the core customers, core budgets, and core business processes of the other party.

This also means that not all functional overlaps represent boundary wars. AI Q&A, intelligent summaries, and automatic report generation will gradually become standard features of software, but standard features do not equal landscape reconstruction.

Under the superposition of the three forces, enterprise software is forming a new competitive map, that is, platform vendors are covering more scenarios downwards, vertical SaaS is extending outwards along the task chain, and low - code and AI tools are continuously diverting lightweight requirements from the bottom.

Software is no longer just about selling systems; customers want task closed - loops

The root cause of the loosening of the enterprise software boundary is not just that vendors want to sell a few more modules.

The deeper change is that the problems that enterprise customers need to solve are difficult to be completely covered by a single system.

In the past, software was divided by departments: CRM managed sales, customer service systems managed services, HR systems managed employees, and