"Before the end of the year, there will be many more process mining companies."
Wen | Wang and Tong and Wu Sijin
Edited by | Shi Yaqiong
Cover source | Visual China
RPA Track, what's new?
Three years ago, the answer was AI.
Now, the answer to this is "process mining."
Process mining is not new overseas. In 2011, Celonis, a process mining company, appeared overseas, with the largest market share and a valuation of up to $11 billion. In October 2020, Celonis acquired Integromat, a well-known RPA vendor in Europe.
Coincidentally, in October 2019, Uipath, the world's largest RPA vendor, acquired Dutch process mining supplier ProcessGold and launched Process Mining based on it. After the product was launched, Uipath's influence in the field of process mining has increased rapidly, and it has become the third largest manufacturer in the market share.
RPA x process mining has become a popular collocation for a while. This is enough to attract domestic believers to follow suit.
Changes have begun to occur as early as 2019, Rongzhi, Hongji, Jiuke, Real Intelligence, Yisaiqi, Laiye, these RPA vendors have launched process mining business:
Rong Zhi has been planning this since 2019,
Hongji has served nearly twenty customers,
Jiuke Information released the process mining platform in 2021,
Real Intelligence is developing its own process mining module,
On April 22, Yisaiqi announced its iS-RPM robot process mining products.
Lai also announced its Process Mining Explorer product on April 26
…
36Kr is collated according to public information
According to a recent Gartner research report, the global process mining software market size of $550 million (excluding consulting revenue) in 2020 can reach $1 billion in 2022 and is expected to reach $2.25 billion by 2025, with a compound annual growth rate of 32.6% from 2021 to 2025, with Greater China growing at a growth rate of 37.2%.
"Before the end of the year, there will be many more process mining companies." One RPA practitioner said so.
So, what is process mining? What can process mining bring to the RPA track? What are the advantages and difficulties of RPA vendors doing process mining? In this article, 36Kr tries to answer the above questions by interviewing a number of domestic head RPA vendors and related investors, as well as analyzing relevant public information.
01 Process Mining: A Physical Examination Center in the Digital Age
Process mining, what is sacred?
In simple terms, if the digital process of the enterprise is compared to seeing a doctor, then the process mining is the physical examination center, which can help the enterprise sort out the process and find out where the problem is; RPA is like a prescription, which can help the enterprise solve the pain of some links.
Process mining is not a new thing, overseas manufacturers have already entered the track and achieved good results, such as the above-mentioned Celonis, which is currently valued at $11 billion. However, in China, process mining is only known to a few people.
Process Mining, from the perspective of academia, is a discipline that spans multidisciplinary fields such as data mining, machine learning, process modeling and analysis - through the targeted collection and cleaning of the information logs stored in various IT systems of the enterprise, visually restoring the actual business flow of the enterprise, helping the enterprise to find vulnerabilities, defects and bottlenecks from the process variants that deviate from the standard process and continuously monitoring, and then guiding the enterprise to find the direction of improvement.
From the definition, it can be seen that process mining is not suitable for smaller-scale, simple-process businesses and enterprises, after all, the simpler the process, the less likely it is to produce variants; the larger the enterprise and the more business collaboration departments, the more complex the process, and the more corresponding process variants.
Source: Celonis process analysis interface
For example, in the following figure, the actual process and the standard process are completely overlapping, and many variations will appear. These variants reflect unreasonable processes that can lead to long time margins, lower efficiency, and even order cancellations.
Image source: Hyhy12580's CSDN blog
Taking the more complex and necessary B2B procurement process in the enterprise scenario as an example, in general, its standard process includes four links: application (demand), procurement (inquiry, business negotiation), purchase (order, copy), and acceptance, but in the actual scenario, the real process variations are as high as hundreds, such as the purchase order approval is not passed, the invoice is not submitted within the specified time limit, the preset special approval process is taken, the temporary price adjustment caused by market fluctuations, and the sudden delivery time change.
Process variations show Source: Celonis
At present, several typical process mining application scenarios in overseas markets include payable/receivable management, order management, procurement management, etc., specific to industry applications, such as loan processing, claim management and insurance application in the financial and insurance industries, etc., and customer complaint handling processes in the telecommunications industry, which are all scenarios where process mining can help enterprises improve execution efficiency and compliance and shorten process time.
What is the relationship between RPA technology and process mining?
According to Gartner, the top five factors driving the process mining market are digital transformation, artificial intelligence, task automation, hyper-automation, and operational resilience, and RPA falls under the category of task automation.
Overseas, after process mining finds problems, problems other than management problems can be solved through process automation; one of the specific solutions is to automatically generate automation scripts based on past log data, and what people need to deal with is to choose which solution to use to automate.
One of the automation scripts is task automation tools, including RPA, which can effectively improve the rate of process automation. For example, Celonis can automatically trigger the procurement process according to the algorithm after detecting that a part of the enterprise is about to run out, so as to achieve automatic replenishment.
In summary, through process mining to identify automation nodes, RPA enterprises can find more use scenarios that can be landed in Party A's enterprise, reduce the cost of manual communication in the early stage, and at the same time reach a closed loop of a set of solutions. According to the Process Mining Industry Scan released in 2021, 78% of process mining customers say that process mining is the key to achieving RPA. Process mining customers also agree with this closed loop.
However, the maturity of domestic process mining and RPA is not as good as overseas, and the corporate management concept is quite different, and whether such a combination can achieve overseas equivalent effects at the current stage is still in the preliminary verification period.
At present, the RPA vendors that choose to do process mining in the market have differences in their solution ideas in similarities. For example, Hongji and Lai also regard process mining as one of the products of the overall solution; while Rongzhi prefers process mining as the second growth curve; Real Intelligence takes process mining as one of the modules of RPA products... The same thing is that everyone uses process mining to serve the RPA business itself.
36Kr found that the current domestic process mining companies are divided into several categories, in addition to RPA vendors, there are also from IT operation and maintenance, upstream from log analysis, as well as direct process mining.
Data source: Sky Eye Check, 36Kr collation
02 RPA company expands upstream
From 2019 to 2020, China has entered an RPA entrepreneurship and investment boom, according to Tianyancha data, in two years, there have been more than 20 RPA track financing events:
Source: Sky Eye Check
RPA is very user-friendly to Xiaobai users, mainly aiming at ease of use, by simulating and enhancing the interaction process between humans and computers, automatically performing highly repetitive daily tasks. Widely applicable to finance, government affairs, e-commerce and retail, logistics and other industries, but often solve the problems are very subdivided, such as information entry, statistics, generated data, automatic download, report summary, auxiliary review, email, etc., especially a relatively simple, single interactive application, can not be copied on a large scale within the same customer.
This determines that at this stage, RPA vendors mainly serve large customers and more scenarios, resulting in the entire delivery process relying on "people", requiring a lot of manual pre-sales consultation and communication before landing, requiring a lot of labor to deliver when landing, and still needing a lot of manual operation and maintenance after landing.
With the exception of a handful of RPA vendors that focus on one segment, most RPA vendors have chosen the general path. An investor told 36Kr that most of the manufacturers that do general RPA on the market have revenue of 100 million to 300 million, which also includes integration, implementation, and service revenue.
The immediate consequence of this is that people will be short-armed in various fields and there will be scuffles. Not only that, because the differences between different fields and different subdivisions are large, it is difficult for products to have enough industry knowhow accumulation, which in turn affects product capabilities.
As a result, the RPA track has been highly competitive at the commercial level. One of the most direct manifestations is low-price competition, fighting a price war. A practical case is that a domestic listed company plans to purchase RPA tools, in the process of bidding, the two overseas RPA head manufacturers, the first and second quotations are in the order of hundreds of thousands, the difference is only tens of thousands; and the domestic RPA head manufacturer's quotation is only 1/27 of the average price of the first two, or even lower than the aforementioned difference, more volume, the quotation also includes 6 months of labor and travel.
Hongji believes that the development of the RPA track so far, will go through three stages, 1.0 stage is to prove the value of RPA, 2.0 stage to compete for cost performance, that is, single point product capacity and price, 3.0 stage is the overall automation closed-loop solution, that is, the overall digital productivity formed by digital employees and digital business.
Now, the RPA track is undergoing a transition period from 2.0 to 3.0, and it needs to compete in terms of single-point product strength and overall product strength.
The industry's view is that there are several ways out of the transition period from 2.0 to 3.0: the first way out is to improve technical barriers, such as improving AI capabilities; the second way is to expand more product modules; the third way is to provide integrated solutions to serve more large customers, so as to expand the customer base; the fourth way is to find more business scenarios where RPA can land.
RPA products are the downstream of process mining, the two customer groups partially overlap, and the products can also form a closed loop, which is more easily accepted by Party A's large customers.
The upstream process mining can not only be a new business module, but also help RPA companies find more scenarios that can be landed by combing the process, and process mining can also become a way to obtain customers for RPA products.
Rong Zhi gave an example to 36Kr: "Government customers because the approval of various departments involves different systems, so the core need is to simplify the process and improve efficiency without changing the system, and the process mining + RPA can solve this problem very well, after the process variant is found, through clear rules, with RPA robots instead of people for 7 x 24 hours of automatic approval, effectively solved the repeated approval, approval time is too long, etc., and finally most of the process is shortened from the original one month to one week, and the original week is shortened to one day." ”
Of course, there are several special manufacturers in the track at present, they cut from a single e-commerce or financial track, the industry knowhow deeper and thicker, so there is no plan to open "process mining" for the time being.
In the future, more manufacturers may join the ranks of process mining, but when to act is related to the established goals of manufacturers. "It's not easy to do two things at the same time," one RPA vendor lamented, "especially when new businesses are developed that require a lot of manpower and tens of millions of financial support." ”
So, will the process mining that RPA vendors have explored be a good recipe for customer acquisition?
03 RPA does process mining and acquires new customers
If the process mining is in the physical examination center, RPA is a prescription drug, then in the middle, there is still a missing link, that is, the doctor who prescribes the drug.
Without a doctor, it is decided that the digital closed loop of serving large customers cannot be closed.
Generally speaking, the role of "doctor" is taken by consulting firms.
These large customers of RPA and process mining are often already on the information and intelligent systems. This means that they often already have some IT consulting firms as technology service providers. RPA companies and process mining companies go directly to grab the market, which is difficult.
In addition, consulting companies through in-depth interviews with employees at all levels of the enterprise to find problems and formulate solutions, but often after several rounds of research, the core problems of the enterprise are not completely solved by tools, and some management problems under the lights need to be objectively expressed in the role of consulting companies. The value of process mining is highlighted here, it can be a real-time, objective and comprehensive response to the real business flow of the enterprise at present and the analysis of the problems layer by layer, which can help the enterprise sort out the needs and find the scenarios that require RPA.
For consulting firms, process mining is a productivity tool, and RPA is a value-added product that can generate premium revenue for them.
In addition, the consulting company is also a user of process mining, which can develop audit tools based on process mining, and quickly form audit report drafts by monitoring the full amount of customer data in real time.
This determines that RPA vendors need to cooperate with consulting companies to do process mining.
Of course, after the enterprise decision-makers find management problems through process mining, they can also choose to solve themselves, and the logic is that changing the process and changing the system will be a big injury to the enterprise itself, and it is not suitable for making too many adjustments, but can be fine-tuned by using RPA tools to achieve the goal of improvement. For example, government affairs enterprises, Chinese-funded enterprises, secret-related enterprises and foreign-funded enterprises in China.
Therefore, some manufacturers have chosen another path - not to bind with consulting companies, but to cooperate directly with customers, or to use their own channels as the mainstay, supplemented by cooperation channels. For example, Hongji is more inclined to dig up products in the secondary sales process of the old customer group, so its own channels and external channels account for half of each. There are also manufacturers that choose to deeply bind with IT system integrators, such as Rongzhi Information.
So, interdisciplinary process mining, is it difficult for RPA companies to do it?
04 Make up for the technical threshold and solve the resource problem
If "borrowing a ship to go to sea" is the main way for RPA vendors to do process mining, then compared with process mining companies, problems such as technical thresholds and insufficient resources are problems that RPA vendors need to solve.
As mentioned above, there are also many companies specializing in process mining in China, such as Wangfanxin, which has three rounds of financing in one year, and Xuanxing Technology, which will complete the third round of financing in one year.
Some people in the process mining industry believe that RPA has nothing to do with process mining. A process mining manufacturer told 36Kr that the reason why it is simply from the perspective of process mining, rather than the domestic hot "RPA + process mining", is because he has been engaged in process mining for 6 years, and there is not much cooperation with RPA vendors, the entrepreneur believes: "The problem of the process is actually a problem of management, and the problem of management is not something that RPA can solve." On the other hand, task mining is more closely related to RPA, and the results of task mining can be directly output as RPA's project files rather than process mining. ”
Competing with companies that specialize in process mining, RPA companies also need to solve some challenges. RPA is market- and customer-oriented, not technical, and therefore does not require particularly high technical barriers. Process mining is not the case, and its technical implementation is more difficult.
As a comprehensive discipline spanning data mining, machine learning, process modeling and analysis, process mining has two core capabilities, one is BI big data processing capabilities, the other is process mining methodology; having a database specifically for process mining is a major threshold for process mining enterprises, and its direct appearance is the maximum amount of data that enterprises can carry, such as Celonis is 25 billion, and Uipath is 100 million; in addition, how to analyze the processed data with algorithms for various displays. It is not biblibility that is tested but the ability to understand the industry and the customer's business.
This means that RPA vendors need to increase their investment in this regard.
"RPA vendors are not going to invest in this." One practitioner scoffed at the RPA "transformation." "The biggest threshold for process mining is the amount of data, how to ensure that hundreds of millions or even tens of billions of data can be quickly and stably analyzed after being poured into the platform, behind which it requires a lot of manpower and time to develop."
At present, the investment of RPA vendors is mainly divided into two types of solutions.
One is self-research, an RPA manufacturer self-research for two years, invested in a team of thirty or forty people, across the two R & D centers in China and the United States, and finally let its process mining products can be commercialized, if you count process mining + task mining, has been delivered to twenty customers. But this manufacturer also believes that many RPA vendors now say that they are doing process mining, but it is only "performance art" and "the next story for investors", and RPA companies that truly embrace super automation are preparing real solutions to the problems found in process mining.
Other RPA vendors choose to take "money for time", such as OEM overseas process mining vendor cooperation. An RPA vendor announced its own process mining products this year through cooperation with a mature process mining vendor in Germany.
Whether to go "fast" or self-research depends on the maturity and stability of the RPA vendor's main business, because the second growth curve needs to be built on having sufficient human and financial resources. According to public information, the Celonis R & D team has 267 people; a domestic practitioner said, "80 people R & D team dedicated to the development of 2-3 years to achieve Celonis 2017 product level, the annual development cost alone will exceed 20 million yuan." The old business needs manpower maintenance and continuous iteration; the new business line needs to be developed by laying people, if the financing capacity of the enterprise is insufficient, consuming too much existing income, it is easy to be "volumeized" out.
Even if it can be invested, if the RPA vendor's product is not mature enough, and the customer industry and business know how is not enough, when serving customers, it will expose problems and lead to customer churn.
In summary, RPA vendors do process mining, which is also a Hard model and faces many challenges.
So, are there any other options besides Hard mode?
Overseas, Celonis acquired RPA company Integromat in 2020 and launched MES (Execution Management System), which means that process mining and RPA vendors are beginning to embark on the same path.
However, due to various reasons such as the degree of digitalization of domestic enterprises and the ideas of enterprise management, domestic process mining manufacturers are in the initial stage and may become one of the indispensable partners of RPA enterprises in the next 5-8 years.
Digital Endgame: Hyper-Automation
Regarding the prospect of RPA process mining, there is currently controversy in the industry.
Some practitioners said that so many RPA vendors may not be rational in the process of entering the market: "The momentum of this great leap forward is very similar to the early stage of RPA in that year, and whether it is RPA or not is desperate to enter the game, hanging sheep's head and selling dog meat to tell stories, only by really patiently doing it to know that there is actually a relatively high threshold." ”
Some practitioners believe that if RPA vendors start to do process mining, which is a difficult business, it may accelerate the arrival of the industry reshuffle period.
According to the feedback of RPA companies, the domestic market is still in a very early stage, and the awareness of process mining is generally low, still in the education period; from the perspective of capital, there has not yet been a large-scale boom; from the perspective of products, the current RPA manufacturers are not fully mature; another reality is that on the customer side, there is a situation where foreign capital cannot be changed, Chinese enterprises cannot change, and because there are not many successful cases, corporate customers are still in a wait-and-see state.
From the perspective of overseas process mining tracks, it is possible to cut from RPA to process mining. According to Gartner's statistics last November, celonis, which currently has a global market share of process mining, is the largest, myInvenio, which was acquired by IBM, and Uipath, an RPA vendor, third.
Why was Uipath able to catch up to third place? Uipath's process mining technology was acquired from independent process mining vendor Process Gold. Automation Anywhere (AA), one of the world's top three RPAs, also acquired FortressIQ in January to prepare for its IPO.
According to the 2021 Everest Group PEAK Matrix Process Mining Technology Vendor Report, Uipath provides an end-to-end automation platform that processes through all steps of an enterprise automation project. Combined with RPA, process mining can help companies redesign processes to improve efficiency and introduce more targeted automated processes to increase productivity and maintain resources.
Source: 2021 Everest Group PEAK Matrix Process Mining Technology Vendor Report
In fact, the entire "discovery" section of Uipath's end-to-end automation architecture is designed to better implement the design and development of the all-round mining process.
In other words, by acquiring a very strong process mining technology, Uipath's process mining capability has found more landing scenarios for its RPA, and the RPA capability can better undertake the steps after process mining, forming a complementary positive cycle.
But one thing that cannot be ignored is that there are enough mature consulting markets overseas. The early intervention of the consulting company "can greatly help the process mining enterprises to reduce the difficulty of entry, so as to achieve the results of high productization, low delivery and high human efficiency." When an investor mentioned the differences at home and abroad, he said, otherwise, at the moment when the domestic process mining has just started, every time an industry scene is landed, it is necessary to deliver the project system and customize the development, which is "too heavy."
And this role of "doctor" is exactly what is missing in China.
At the same time, overseas new business, more accustomed to acquisition. Domestic one is the lack of acquisition environment, and the other is that there is no target for process mining in China. At present, there are only a few companies that do process mining alone, and they have also received the acquisition intention of RPA manufacturers, but they have refused, after all, the domestic process mining market is in the ascendant, and the future space of manufacturers is much higher than the acquisition price.
Process mining, task mining, RPA, or dialogue AI are all part of the implementation of hyper-automation. Players at home and abroad ultimately want to provide customers with an overall super-automated solution. This solution can find problems, find countermeasures, solve problems, and form a real closed loop.
Source: Gartner
The reason why the concept of hyper-automation will be proposed and sought after at home and abroad is that the global digital transformation has entered a new stage. A single RPA can only realize the automation transformation of the enterprise, and the overall digital needs of the new era of the enterprise cannot be met; a single process mining can only find problems, and the final solution is not digital if it still relies on people.
At present, the first batch of domestic enterprises trying to digitize have entered a bottleneck period - with the continuous deepening of enterprise informatization, the process of enterprises is becoming more and more complex, for bosses and managers, if you want to further understand the status quo of the enterprise's process, process mining is indeed a tool that can improve operation management and efficiency, so the trend is very obvious.
"As digitalization accelerates and IT infrastructure matures, the next 5-8 years are likely to usher in a wave of rapid growth." A partner of an investment institution said of the development potential, "Because in the face of cumbersome process links, a concise system environment, efficient processes and fast response speed are very much needed for large enterprises."
(End of this article)
For more analysis and reflection on the process mining itself, 36Kr will be presented to the reader next month.