What is the threshold and future of RPA?
Enterprises are welcoming a more “automated” future, and RPA is opening a new era of smart office. On September 10th, the “WISEx Enterprise Intelligence Industry Summit” was held in Shanghai, inviting major capitalists and representative companies in the domestic RPA industry to focus on how RPA disrupts corporate processes and promotes new changes in enterprise intelligence.
RPA is a hot topic in this year’s venture capital circle. Previously introduced in the New Wind Report, the concept of RPA (Robotic Process Automation) was launched abroad in early 2000. The product form is The purpose of automation software is to replace duplication and intensive labor. The work scene is based on strong rules, and humans need to jump between different systems to obtain and input information.
In the morning of the round table summit, the guests shared their role in the AI in the RPA. Views on the three topics of meaning and significance, the development stage of the domestic RPA market, and the challenges of the Chinese RPA market. As AI plays an increasingly important role in RPA, the domestic RPA market is gradually ushered in a booming stage, and more and more companies are entering this field.
However, what is the market threshold for entering RPA, what is the future development trend of RPA, and where is the future development boundary of RPA? In the afternoon round table summit of the summit, Chen Yuyang, Director of Innovation Consulting Department and Chen Shengde, Partner of Ernst & Young Consulting Services, Wang Guanchun, Chairman and CEO of Lai Technology, Tang Qisong, founder and CEO of Yisaiqi, and founder of Yunsheng Technology and CE Liu Chungang Fang Wenyu, general manager of Xingye Digital Innovation Incubation Center, discussed these issues together.
The following is a roundtable record, edited by editor:
Enter the market threshold for RPA
Chen Yuyang: Our Kr8 Innovation Consulting Platform helps large companies discover new technologies and new business models in China.To promote open communication between traditional industries and entrepreneurial ecology, and jointly bring innovation to the ground. In the process of communicating with many customers, we also heard that customers are talking about RPA, how can we help our business through these new technologies? Today, I heard a lot of interesting directions. The big business needs to consider RPA, the combination of RPA and AI, the construction of RPA’s own ecology, including the application of RPA in the financial direction. We all think it is a good direction for the future. . I got the issue of the organizer, it is still a bit sharp, and it is a common problem in the industry. I want to listen to what everyone is saying.
The first question, some of the keywords mentioned in the RPA are not particularly new. They have been mentioned in the past, including the management of business processes and the outsourcing of business processes. Some manufacturers with operating systems, such as Microsoft and Siri with operating system resources, have already participated in this work, and some of the world’s top 500 companies are developing their own RPA. What is the threshold for players to enter the market now in the ecology of RPA?
Chen Shengde: In Ernst & Young, one way we think about the problem is to ask questions. For example, intelligent automation is not necessarily implemented by RPA, but also by system development or other technical means. Two years ago, I talked to a CFO of a central enterprise about RPA. He would say, “Our company is optimizing our company’s ERP, so I don’t need RPA.” I asked him “Is it possible to achieve a seamless connection after optimization for ERP? Is it not necessary for people to do this?…”
In fact, from the point of view of the function realization of RPA itself, the biggest selling point or starting point can still replace the work of people, whether it is from the perspective of cost efficiency or liberation of people, it is to avoid people to go. do. Especially the younger generation of colleagues in the enterprise, if he is allowed to do some repetitive work, the value is not too high and too boring, he will directly “go”, so the work of solving repetitiveness is the real starting point.
As for the threshold, from my point of view, the enterprise traditional system has been solidified, such as ERP for so many years, RPA because there is no cut-in, so the original system should not be changed, nor It is too possible to replace it with RPA. Another problem is how AI does better. This is a deep question. If you emphasize more about what AI capabilities RPA can call, then this requires a platform to build.
Simple answer your question, the threshold is not really high. From a market perspective, the future may face fierce competition. However, it must emerge in terms of scale and depth. The real big waves are sanding, leaving behind the company that really performs well.
Wang Guanchun: Many concepts or technologies in RPA are not new, it is very interestingThe phenomenon. From a technical point of view, the development of technology is often cyclical. For example, the concept of deep learning, which was particularly popular in the past two years, is behind neural network technology. Neural network technology existed twenty or thirty years ago, even after a period of fire. But now it has changed a vest and reappeared, but it does add a lot of other aspects, combined with more computer resources, it has become very useful compared to before.
RPA is not really a new technology. For UI automation, just after the arrival of the technology founder has done button wizards, this product has served hundreds of millions of users before 2018, there are hundreds of developers community. Automated scripts made with button wizards are more used in games and other scenarios. In the past ten years, 20% of users have used the button wizard to do office scenes, but they have not met today’s security requirements. What may be done is to meet the needs of simpler punching automation. In all enterprises, the informationization process is very fast. It is very valuable to use the information system to prepare the basic environment and replace the repeated and boring work done by people with RPA.
So does this product need to be innovative? UiBot and button wizards, which are also made by technology, are not completely unified in the overall product architecture. They are completely different from the perspective of security or integration of various new AI capabilities. This software complexity can be very high, but the adaptability can be very good. Excel is such an example. It is very simple for any company or individual to make a table processor, but the investment in Excel spreadsheets is very huge, because although 99% of its functions are not used, they really want to use them. It can be very important to use it when you use it.
Microsoft rarely does other SaaS tools because they think Excel is the biggest SaaS. As a product of human resources, or a product of CRM, for SMEs, excel can do so many functions. Since it can be a platform product, there is no need to do vertical SaaS. Why is the platform product so popular, not into the vertical field, because of its product features, developers have done a lot of innovation, built this barrier. This is the same today, the RPA tool is currently a product used by developers, enterprise IT staff or good business people, If the ability to combine artificial intelligence with RPA is good enough, like a robot assistant The product may be used by everyone on your computer or on your own mobile device.
Tang Qisong: I am often asked this question, where is the value? Where is the threshold? After RPA entered the Chinese market, it suddenly exploded in the past two years. Who came to China in the past 20 years? From Kingdee and UF for twenty years. Such a productWhy did it suddenly fire up? This technology is not new, including gartner’s four low-code, and we’re all doing it. There is a process for the derivation of a technology. There is no one to seal the throat, only a little bit of accumulation. The problems involved here are particularly numerous and complicated. Excel has a lot of features, we have to look at it rationally.
The first step is in the case of China’s Internet and global upswing. First of all, we look at the products. At that time, I wanted to use SAP, no cloud, and now everyone can experience it. Why open the cloud for free, because we have a lot of eco-collaborators, including partners who have never used our console, everyone in the market is trying, every robot consulting company to help us. The server is not too expensive, but he does not know how to use it. I don’t know how to understand it. This is a guiding process.
New is good, old and old. We do framework design, develop languages, and then what do we use in the era of big data? It should be py (English). In any aspect, in a developing market like China, as a manufacturer in this industry, should we have a little thought of ourselves? This threshold is a way to go from the beginning of the product to your ecology and to the future AI.
I have been in the OCR field for 20 years. Now OCR brings us the most useful use in the RPA field is the verification code. I have not seen anything else. At present, the current development is different from that of 20 years ago. Now it is on the same starting line. How to get out of China, we can look at the global changes. The above is some of my feelings.
Liu Chungang: Many people are concerned that RPA technical barriers are not high, or whether there are difficulties. In the country, there are some technical barriers that some people think are not high, and some people even think that the barriers to cloud computing technology are not high. From the perspective of making a product, it is necessary to be awe-inspiring and to be awe-inspiring in every function of the product. If the company employee tells him that the technical difficulty is too low, I say that you are too low on your own product. If he says that the technical difficulty is not high, I will say that he is too low on his own.
Our product name is Tianshi. First, I hope that Tianya RPA can bring robots to our customers, bring robots that can replace repetitive work, and do things like the craftsmen sent by heaven. The second is to hope that each product team member has an ingenuity, to do what he is doing, to maintain awe of the products he has done, and to make every detail to the extreme.
RPA has a lot of technical parameters. Whether it is different product compatibility, execution stability, space recognition, it is not difficult to make a prototype from these requirements. I believe these friends Have done it. But it is technically difficult to adapt to a different industry internal information system and various network systems, so that any technical product can be used by business personnel and those who do not know how to code. It is difficult and has barriers.
Fang Wenhao: We are a financial technology company, so when we do integration services, there will be different voices within the company. We can not do it ourselves. If the threshold is not high, why integrate, we can start from the original ecology, starting from 0, as long as there are more than a dozen people involved, from core software to delivery services are all provided by ourselves, our answer is No.
There are two main considerations. One aspect is the threshold. Whether it is the example of excel by Wang, or the ingenuity of Liu, for the software company, the originality And the principle of first nature is very important. What are you good at, for example, you are starting from the button wizard to other related products, this is the natural transition, that is, it is a very persistent, very focused area. But for other software companies, it is not possible to do everything.
Our company’s position is to provide financial technology services to financial institutions, but what is the purpose of its services? In fact, it is to solve the pain point of the customer, so it is not to solve his pain point by doing his software from beginning to end. It is the use of the industry to specialize in the use of professional software, plus its own professional services. Solve his pain points, This is the second aspect we consider.
The future of RPA
Chen Yuyang: The next question, if we look ahead, from the future development trend, we will think that RPA this product includes the most relevant services in the value chain. Where is the node? Is it the development of the process engine, or the specific technology of AI, what do we think?
Chen Shengde: What is the future of RPA? I personally think that each person is equipped with a robot assistant, which is a preliminary stage, or a stage that meets the small needs of small and medium-sized enterprises. I firmly believe that the future is deeper in the large to B business. In the example I just mentioned, the future robot is not just a helper, but a concept of replacing people. Telling many people that this is not a layoff, but a situation in which the robot can operate itself.
In the previous management changes, in order to improve efficiency and share, it is necessary to divest and concentrate the organization, and then completely unify and standardize the process. Another exampleTo do ERP implementation, ERP is not a product function, but a BPR (process reengineering) problem, but a problem of enterprise flow optimization. The Chinese industry is very wide and the market is very big. I think that in the future, it is very complicated and deep to do a thing. I think the business logic behind the future development of RPA is also like this.
At the same time, scenery is a very critical thing. What kind of problem do you help the customer solve in what kind of scene, and what kind of object is targeted. In the face of large enterprises, how to truly solve the problem of operating the entire robot team under the complex structure of the enterprise. What’s more important in the future is that differentiation is not in the product itself, but in the entire ecology, including the ability to integrate resources, as well as everyone’s cooperation and understanding of the business.
You just mentioned that RPA is expensive, but why many customers are willing to choose us. Some big customers are rich and lazy, but the deeper reason is “I hope that you come in not to tell us what your products are like, how do I use them, but to understand what our company and our industry are like.” You tell me what I should do.” I hope that our robots will do this step forward, which is a very important part of it. Of course, the product is a foundation, at the same time, it is necessary to deepen the depth of the industry profession and deepen the overall operation of the ecology.
Wang Guanchun: Listening to Mr. Chen’s introduction greatly enhanced our confidence in making robot platform products. In the future of the technology, each company will have a robotic team and a robotic iron army. This is something we believe can be achieved, but it is still in the early stages of society. Just mentioned by Tang, most companies can use one or two machines. However, every department has robots, and many departments have robots. Ernst & Young has already painted such a future for us. This is what we believe can be realized. s future. This morning, I talked about the narrow RPA and the generalized RPA. Today, the RPA for UI automation is a relatively simple robot assistant, which is a stupid apprentice.
The generalized RPA or RPA of Lai Technology is constantly evolving. I think it is a new type of human resources. is a robotic employee who can play a role in the department and be creative. Things, robots solve simple and repetitive tasks. When we also plan our own product line, we need robots that can follow the process RPA robots, and the front-end business needs to communicate with external customers. This is the robot with more NLP and voice dialogue technology today. If the enterprise post is an operator, it needs to do interaction and needs to do business. It needs two kinds of business combination to complete. From the alternative work 10%, 50% to 80%, it is a process of gradually replacing, and finally we see It is the stage of social evolution of robot human resources.
Microsoft is the world’s greatest software company, the first software is programming software, and today’s RPA product – UiBot is also a RPA robot programming software. With the increasing ecology of personal computers, there have been many applications when personal computers have come to all enterprises, such as Excel, Word, PPT, etc. These applications are killer-level computer software, so there will be killer-level robot applications in the future. Robot applications need to be combined with data and applications. To achieve such a goal, there are still many elements that are not ready. Whether the market needs to be better educated, or need more RPA talent, RPA technology, AI technology. The market will push products forward and develop.
Tang Qisong: Yisaiqi is a company that does To B. We have been privatized in the user field. Why do you do sharing today and let everyone use it for free? Users should be hierarchical, and every company should determine who is your user and who is the one that makes the company profitable.
We understand this situation in three ways. First, for the company that makes the software, we locate the product company, and the first company we took out was the product. How is the product done? Let everyone experience, evaluate, and even open the product. How do we integrate and call each other, first of all, from the perspective of products, our first level – mutual integration must be done. Just after Mr. Chen talked about ecology, large enterprises from privatization to sharing, can these experiences be used by these small enterprises? Since it is a public cloud SaaS, there is no autonomy, because the SaaS upgrade is to concentrate on each other’s advantages to take a commonality. If you want privacy, personalization can only be privatized. The second is ecology. How to position your ecology depends on which ones are your real users. Third, AI is the future direction of development, but if AI really helps us write code, can we achieve liberation in many aspects. This is my point of view, from three perspectives, product, ecology, AI.
Liu Chungang: Many of the technical principles in RPA technology have a long history, but as a complete product, there have been few years in the world. We firmly believe that RPA technology will continue to evolve and evolve as a product. RPA can be applied more and more within the enterprise in the future. At present, RPA technology is more to solve some repetitive work, but in the future, whether it is in the banking industry or more, it helps us not only do more than just other industries. It is a repetitive work. In the future, RPA understands the internal processes of the enterprise and understands the internal data of the enterprise. It can replace humans or assist human employees to do some analysis and even make some decisions.
From the perspective of cloud expansion to analyze the direction of future RPA, the future is divided into three levels, one is the implementation of RPA robots, the current can help humans, help employees to implement the process according to fixed rules; the second is the perception robot, RPA It is given to traditional AI algorithms, such as NLP can extract information in text, OCR recognizes documents, recognizes image information, and allows RPA robots to take out the executed data in unstructured data. The ability to sense data. The third is cognitive RPA. RPA is now based on rules in the execution process, is the information of the execution of the input, it does not understand what it is doing. In the future, RPA can understand this process. It can be recorded in SIP and can be recorded in ROC. In the process of entering employee information, instead of entering one parameter, people understand the meaning of each parameter. RPA robots become more and more intelligent like this, and finally can go through these three stages and bring to each enterprise. Greater value.
Fang Wenhao:From the perspective of financial institutions, the development of RPA will be more and more extensive. With the in-depth application, it is a great challenge to the internal management of RPA’s own users. One is the management of technical architecture. As mentioned by Mr. Tang, it is no longer a small assistant, but a console, and a robot management robot. Thousands of employees have their own assistants. What do thousands of virtual employees do, who knows when they are free, when they are busy, and whether they control it? This is a topic of technical architecture.
The second is process management, especially for financial institutions. For example, the issue of publishing is very important for the bank. The R&D team and the production and operation team have to complete a very standardized release process in order to release and iterate the version. This situation has been broken in the RPA, this is a very important case encountered in the bank implementation process. We want to do things for users to cultivate. When we do RPA management norms and management systems, when we regard robots as virtual employees, it is the integration of R&D + production operations. How to manage it, we need to do the process system. Progress.
RPA’s future development boundary
Moderator: I did a little homework on the RPA market. The differences in the caliber of many reports vary greatly, especially the size of the market, including reports on Ernst & Young. I believe that each report Will have their own basis. As a practitioner in the RPA field, how do we look at the boundaries of RPA? What is the future position of the RPA market in the entire enterprise service ecosystem? What do we think?
Fang Wenhao: The border is difficult to generalize. When should I use RPA inside the company and when I should not use RPA, there is no standard answer. You can use some matrix to analyze, for example, it is a dimension for income increase, efficiency, and cost savings, and a dimension for technology implementation. At the same time, in addition to these direct effects, it also has indirect effects, external customer satisfaction, internal employee happiness index, are the factors that balance this process or not. In fact, I don’t have a standard answer.
Liu Chungang: When you pay attention to the market size of a product, you will often count some data, how much your product is sold, and how much it sells among different customers. From our perspective, the market size of a technology and a product depends on how big the problem is solved for the customer. The bigger the problem you can solve, the more problems that more customers can solve in more industries, the more the market size Big.
The results of different statistical caliber statistics RPA are very different. If you count the RPA products sold by RPA vendors around the world, it is not that big. If you consider how many RPA products can solve problems in the future and reduce the amount of repetitive labor costs, this is very huge. This is my understanding.
Tang Qisong: I am here to talk about my feelings. In this market size, each company has its own position. There were not many companies doing behavioral analysis software throughout the world in 2015, only three in the United States, a total of thirty people. We are now more than 30 people specializing in behavior analysis, doing security analysis to the left, having behavior, going to the right to do efficiency analysis, and also behavior, going up to the RPA, from the assistant to the cluster. How is the market size, how can you do it, and ultimately what your product can offer to your users. Like us for so many years, a lot of income in the industry comes from users upday, behavior analysis to solve A problem or B problem or C problem.
In the market, security services, a company invested 10 million, by the end of the month, the boss said that nothing happened, where did the money go? The same investment of 10 million, this year there are several failures, the boss will ask, the investment of tens of millions will still have problems. In the end, the vote should not be cast. In fact, this is a problem without solution. Security is a strategic thing that the state needs to do, and the privacy of the user must be protected. Now the enterprise’s assets are data, unlike previous PCs and different servers, so the investment in this area is worthwhile. 30% of enterprise vendors will not be priced uniformly. In fact, the ROI brought by each enterprise using RPA is different. China alsoIt has a lot of characteristics, people are intensive, we have a very large space, after three years of analysis and prediction, the predictions for China are not necessarily correct, this is my point of view.
Wang Guanchun: Mr. Tang has been doing this for many years, so his predictions are more accurate than many organizations. In the short term, the RPA project is more focused on the ability to implant a robotic solution into a company through RPA today, because it also believes that every company will use robotic employees. The easiest and fastest way is to be able to put a safe robot into the enterprise. Today, we see that RPA is faster than the dialogue robot. If the company tastes the sweetness, it will buy a second robot, or add value to the robot. At this time, it will enter a high-value scene. Companies like automation and intelligent decision making today can capture such a benefit, which is a short-term thing to see.
In the long run, RPA or robots are still solving human resources. The biggest trend in the past ten or twenty years is that companies have built their own computer rooms in the local area. Today is the time to go. Ali, Amazon, and Microsoft buy cloud computing. In the form of renting, large companies will not buy fixed assets to build computer rooms. The future of human resources is the same. In the future, in the process of expansion, what we see today is flexible employment. In the future, it is very likely to rent robots. It is possible to rent a robot in the cloud. Or the requirements for privacy are relatively high, or possessiveness is relatively strong. It is also possible to put robots locally.
This thing is like today’s enterprises put the computer room on the cloud. In the future, more companies will use the robot solution to solve the need for temporary expansion. Think about it, the budget for buying a computer room is just the company’s IT budget, but the budget of human resources, any company, whether it is doing technology or business-related business is a big budget. In future service-oriented enterprises or technology companies, human resources account for 80% or more of the cost. If a robotic solution can create value here, companies are willing to invest, which is much larger than a company’s IT budget.
Chen Shengde: The budget that Mr. Tang has just said is not accurate. In fact, basically the budget is always inaccurate. If you look back at the communication industry 20 years ago to make predictions, who can think that mobile communication can develop to today’s scale, has it begun to mature or even decline? The trend is more important than the forecast number. Now everyone is still in a good market development atmosphere, including everyone’s understanding and understanding of RPA, large-scale enterprises from the working level to the management level, and the digital company’s digital transformation. The guidance and influence of domestic enterprise demand.
Either RPA or IPA, there are two letters that haven’t changed. People themselves are very lazy, this is the driving force for human progress., I am replacing 20% of the work or 50% of the work. From the perspective of consultants, sometimes the process of the company itself is of low value or has no value. The original practice is problematic. It used to be the process of process optimization and automation, but now I want to start doing things in a smarter way, which is the work of RPA.
Automation is still a relatively heavy piece in a large enterprise system. We can’t overthrow the system, and all work makes RPA a replacement. The core idea of RPA is virtual labor. What do I let him do, what can he do best? This is also the purpose of doing RPA. For example, recruiting, financial recruiting and human resources recruiting are not universal, but RPA is universal. There are two levels in it:
One is the complementary and alternative relationship between alternative process automation and system implementation. There are customers who have done this work in the IT system. After that, this project is too heavy, too much money, and there is no scheduling. Simply do it with RPA first, after one year, a system upgrade. Consider system automation improvements. This is a question that the company as a whole considers.
The second is the AI problem, RPA is calling this ability. AI is too wise, because companies have many types of modeling, because the decision-making methods are different, and the decision-making stage is different, so there are many imaginable spaces in it.
From the market figures, it is nothing more than the number of the demand side or the provider. The demand side is how much the customer has spent on what, the provider is how many manufacturers, it depends on how everyone defines. As long as we expand around these two parts, I believe this market will get better and better.
Chen Yuyang: Today, due to time, I will discuss it here. I am very grateful to everyone for their opinions and very beneficial. Thank you!