Editor’s note: This article comes from WeChat public account “ToB Industry Headline” (ID: wwwqifu), author Li Xiaosong, 36 氪 released with authorization.If an industry is suddenly hot, it will eventually come to peace, which is an inevitable rule for the development of everything.Even the RPA that everyone knows and knows in the ToB industry in 2019, it is difficult to escape this fate.But calm is not a bad thing. Only by returning to calm, can we calmly examine the future of RPA behind the impetuousness of capital operation.Thanks to the crazy surge in UiPath valuation across the ocean, throughout 2019, there are endless RPA manufacturers in China.This year, there were no less than 10 Chinese RPA vendors that received new financing, released new products, and started another round of transformation.On the bright side, the RPA explosion seems to be a capital carnival.But if you really go deep into the industry, we will find that RPA can land, indeed because of its real value.RPA is intelligent software and can also be understood as an automated robot.As long as the rules are designed in advance, RPA can simulate the tasks of “fixing rules, repeating, and adding value” such as copying, pasting, clicking, and entering manually.We have all learned the pain of Ctrl + C and Ctrl + V. If the repeated paste work process is simple and the rules are clear, RPA can completely replace this type of work manually.First, the ability to land is the root cause of the explosion of RPA. Although RPA is strongly related to artificial intelligence, RPA is not like most artificial intelligence “more concepts, less landing, blind as soon as it is applied.”Almost from the beginning, Chinese RPA manufacturers have figured out the landing scene, and the product is to solve the problem.Throughout this year’s well-known RPA manufacturers in China, whether it is Abbots Technology, Daguan Data, or Cyclone or Cloud Expansion Technology, these manufacturers have almost always settled on financial, fiscal, and taxation processes with “a lot of duplication” andIn the field of “clear rules”.Liu Tiefeng, CTO of Abbots Technology, said bluntly: “Everyone knows that AI will replace labor in the future, and the former AI is a bit invisible. RPA is the tentacle of AI landing, which can make AI closer to reality, and make intelligence and automation easier to understand.The reason is very simple. Take a bank with a large amount of documents and relatively uniform business processes as an example.If the AI can “photograph” the data on the document and fill the data to the designated location directly through the RPA process, it will undoubtedly greatly improve work efficiency.In the same way, the invoices and bills processed by the company’s finance also need to extract a large amount of data from the physical invoice (paper) and fill them into the system.If there is blessing from RPA, this kind of work will obviously be easier.In the past, RPA was not popular because the AI was not mature enough, the “photograph” was not accurate, and it could not “understand” the semantic logic of the text.With the maturity of AI technologies such as OCR (Optical Character Recognition) and NLP (Natural Language Processing), extracting content from solid text and pictures for classification and filling has almost become a landing technology that everyone can understand.It is under this background that RPA can become an important position for AI landing.Most of the Chinese players who entered the RPA this year have a strong artificial intelligence background.2. The explosion of RPA is inseparable from the blessing of “ToB industry headlines” by AI. A simple inventory of RPA vendors with a significant impact this year, we found that before expanding RPA business, Abbots is a company focused on using AI technology to improve financial dataAn artificial intelligence company that generates, extracts, analyzes, and presents; Daguan data does text understanding and text analysis; it also does voice assistants and semantic robots … These RPA manufacturers with self-developed AI capabilities are actually in line with traditional Chinese RPA manufacturers.Manufacturers such as Saiqi or global RPA giant UiPath are different.AI technologies of RPA vendors such as UiPath and Yisaiqi are not self-developed, but integrated.We ca n’t yet say which way is better, “AI + RPA” and “RPA + AI integration”. We can only say that the former relies on self-developed AI capabilities to enable RPA to better connect the company ’s original business and ensure service quality;The latter can provide customers with more AI options to meet the rich needs of customers, and the platform can also rely on light weight to achieve rapid expansion.At present, the RPA industry is just emerging and everything is being explored.No matter which method is adopted, we all know that the development of RPA is definitely inseparable from the blessing of AI capabilities.Therefore, RPA manufacturers with self-developed AI capabilities, such as Abbots Technology, have not fully imitated the UiPath and Yisaiqi models and chose external AI capabilities.They prefer to use the “self-developed AI + RPA” model to establish a solid barrier for themselves. They just use RPA as a business extension rather than define it as the core of the enterprise.”RPA is driven by demand, not speculation. People want to use AI to solve repetitive, inefficient, and simple tasks. This is the actual demand. But in our view, RPA is more like an expansion of AI capabilities.”Tentacles and bonds form the role of more business, but RPA is not the whole company.” Liu Tiefeng said.In his opinion, strong AI capabilities can mine more RPA scenarios. Without self-developed AI capabilities, RPA is easily limited by the level of third-party technology at the scene expansion level.”In the future, the RPA market’s fight will lower the service price of pure RPA scenarios. In the scenarios that require AI functions, pure RPA companies are likely to be greatly challenged by AI + RPA companies.” Why RPA has developedFor more than a decade or two, there has been no explosion, and no manufacturer can be remembered?Why did RPA get hot when AI companies entered after Abbots, Daguan Data, and Lai Ye?Obviously, the hotness of RPA is closely related to the overall improvement of artificial intelligence technology.Since technology has created a blast in the industry, who dares to say that AI technology is not the new core of the RPA industry?Third, RPA has just fallen into a white-bladed battle. In the past, we always felt that RPA was only fired this year and should be a blue ocean market that can be entered.They will definitely choose the most suitable one from multiple RPA vendors. “Therefore, we found from Abbots ‘product matrix diagram that Abbots’ core architecture is still built on unstructured data extraction and financial industries with completely independent intellectual property rights.NLP, and knowledge graphs.The main business is still products and integrated solutions for the financial industry, including intelligent investment research, intelligent risk control, intelligent post-investment, and intelligent operations.Abbots Technology only uses RPA as a catchment for various businesses in series, improving the overall efficiency of the solution internally and better serving customers outside.In this way, one can ensure that the company does not miss the RPA vent; the other is that the enterprise’s AI capabilities can be continuously updated.”Self-developed AI + RPA ‘may at this stage better connect the company’s business and bring real efficiency improvements to customers.” Liu Tiefeng said.Although the RPA industry has existed for nearly two decades, the concept of “AI + RPA” has been hot for almost one year.The industry’s exploration of the RPA landing field and landing method is still continuing. Everyone agrees that the RPA industry still has great potential to be tapped. Numerous investors also regard RPA as the next greenhouse to breed unicorns.According to research from TransparencyMarket Research, by 2024, the global RPA market will reach $ 5 billion.Zhu Xiaohu, general manager of Jinshajiang Venture Capital, also said: “In five years, half of China’s jobs will be replaced by RPA, and 80% of customer service personnel will be replaced by intelligent customer service. In the next few years, tens of billions of dollars in this market can be expected.”Many parties have different views and understandings of RPA, but it cannot be denied that RPA is indeed a practical application that can be implemented.As Zhu Xiaohu described: “We found that RPA and intelligent customer service are the most willing to pay for Chinese business owners, because the return on investment is the easiest to calculate, and robots can easily calculate for several employees.” As long as the company tries it lightly,Can you know how much RPA has saved the company’s manpower, material resources and value.Therefore, in the case of countless investors and industry professionals, RPA’s exploration into unknown areas is extremely important.4. Discussion on the scenarios where RPA fails to expand. As mentioned earlier, RPA manufacturers currently focus on the financial, fiscal and tax fields. These mature scenarios are easy to saturate, and sooner or later manufacturers will fall into low-price competition.So what do RPA vendors do?Many manufacturers may not have thought about this issue, but Liu Tiefeng provided a few ideas for the “ToB industry headlines”.”In the past, RPAs mainly fell into industries with a relatively high level of informationization, but I think that the connection between software and software will only become less and less. Abbots believes that in industries with a low level of informationization, there may be new opportunities. Especially in the field of hardware connection software and physical connection software, RPA has the possibility to explore forward. “Here, he gave an example of the transformation of an old factory.The core machine tool of an old factory has a built-in computer that cannot be replaced. The data in the computer is very important, but the computer cannot be upgraded because the system is too old, so the data cannot be synchronized to the new system.The factory couldn’t solve this problem after thinking about many ways. It had to manually copy the data from the old computer and enter it into the new system, which was extremely inefficient.After taking over this project, Abbots solved the problem in a very “stupid” but extremely effective way.This method is to continuously capture the data on the computer screen by taking pictures, and then use the AI recognition capabilities such as OCR and NLP to directly extract the data from the pictures and input them into the new system.Although this is just an example, it can’t help but cause us to think more.Now that the market with strong demand for RPA is already in the throes, then is there a need for transformation in industries that have a low level of informationization and have not yet been deeply touched by RPA?Is this need not there?Or is it because the technology is immature and the service is not in place before, and there is still no way to transform it?Moreover, the RPA industry has always had a debate over whether to take a standardized route or a customized route.In fact, this kind of controversy does not only exist in the RPA industry. The entire Chinese ToB market is discussing this unsolved problem.DCM Zeng Zhenyu said in an interview with IT Orange: “RPA charging by robot is a natural advantage, and may break the barriers of traditional SaaS or enterprise services in China’s revenue ceiling.” No more scenarios have been explored, and no more ideal has been found.Before the rapid expansion approach, Liu Tiefeng said that the SaaS model is difficult.The ratio of “80% service + 20% technology” may be the solution that Abbots prefers.”The competition in the ToB industry is the competition of the comprehensive strength of the enterprise. RPA is based on the overall service capabilities of the enterprise. Although technology is the key, it can be based on technology. How to ensure the continued stability of the product nodes and how to truly penetrate the customer’s understanding of the requirements are the most important.The key thing. ”The reason why 20% is focused on technology is that the current RPA manufacturers’ technology levels are relatively close. Without further breakthroughs in the RPA scene, the competition has not yet reached the stage of comparing technologies.On the contrary, China’s ToB market is currently in a period of gradual improvement, and there are many diversified demands from customers, and they are not cold about standardized products.80% of the service focus is related to the general environment of the Chinese service market. After all, the focus of enterprise service is on service.Of course, these are just the doubts raised by the “ToB industry headlines”. How to solve them still needs to be discussed by manufacturers, users, investors and experts in the RPA industry.However, it is certain that the “AI replaces people to work” is an irreversible trend. It is for this reason that RPA landing and explosion is the right thing to do.Looking at the 36th “E-Club Creators Club” recruiting of the first batch of CEOs, relying on the media insights, influence and industrial resource advantages of 36 ，, connecting the value of venture capital circle resources and contacts to accelerate the empowerment of entrepreneurship.Insight into the core needs and carefully polish the five rights and interests to help you on the road to entrepreneurship!.