IDC's report clearly explains the RPA+AI market with a scale of 27.8 billion

Latest update time：2020-05-31

Reads：

Guo Yipu from Aofei Temple
Quantum Bit Report | Public Account QbitAI

RPA+AI is becoming a visible trend in the AI field, and various manufacturers are moving in this direction.

Another name for it is IPA, Intelligent Process Automation.

Over the past year or so, this field has been mentioned intensively and received close attention. Even large, medium and small companies at different stages have joined the market, including manufacturers with natural language processing backgrounds and AI startups with certain experience.

RPA+AI is indeed largely based on natural language processing capabilities. Of course, other AI models such as computer vision are equally important to it. Relatively mature technology and broad application space make RPA a low-hanging fruit in the commercialization process of artificial intelligence.

Perhaps, RPA in its early stages is to natural language processing what face recognition for security is to computer vision, and it is one of the earliest opportunities for growth in this field.

RPA+AI, how to liberate productivity

Based on RPA (Robotic process automation), by connecting to AI capabilities such as OCR, image recognition, and natural language understanding, automated robot system operations can become smarter and complete more tasks.

For example, if you need to enter some paper information data into a database and organize, classify and save it, traditional RPA can automatically complete the work of clicking, setting and uploading on the graphical interface of the database background.

These tasks are 1+1=2, they are hard-coded and do not require advanced intelligence to do.

However, in traditional RPA systems, when the machine operates automatically, if you want to copy and paste the characters in the file into the electronic database, it cannot do it. OCR must be added to recognize the characters. If you need to label the file or write a summary, it cannot do it because this requires a certain level of natural language understanding ability.

In this way, the automatic work will be constantly interrupted and human wisdom will be needed to supplement it. However, the work of recognizing text and labeling is undoubtedly very simple, repetitive and low-level mental work for humans. It would be strange if a small assistant who does such simple and repetitive work with no room for advancement does not leave.

Therefore, RPA+AI can completely mechanize and automate these low-intelligence human jobs, bringing higher levels of efficiency improvement and human liberation.

RPA used to only be able to walk, but now it can run, jump over hurdles, high jump, cross rivers, climb mountains... From a toddler, it has transformed into an explorer who can climb mountains and cross rivers.

The key is that the programmer community loves open source, and various AI capabilities don’t even need to be developed by ourselves.

OCR character recognition, image recognition, email content summarization, multi-round conversations with robot assistants... If you want to find a way to implement these functions, you can directly find the pre-trained model with the best performance for the task on GitHub, make it into an API and connect it to the original RPA software, and then you can use it.

With RPA in the left hand and various mature AI models in the right hand, when put together, they become new products with new functions.

It seems like a technical capability that has been picked up for free. As long as an application scenario is found, the profit prospects are visible to the naked eye.

No wonder many manufacturers have shifted towards RPA+AI in the past one or two years.

So how big is the market opportunity for RPA+AI?

IDC released a report titled "China Intelligent Process Automation (IPA) Market Opportunity Analysis", detailing the current opportunities and prospects of this track.

RPA+AI is about to explode

There are some differences in the development space of RPA+AI in China and globally, and the domestic growth space is obviously larger.

According to market size data tracked by IDC, IDC predicts that by 2023, the global RPA market will reach US$3.9 billion (RMB 27.8 billion), with a compound annual growth rate of 36% from 2018 to 2023. The Chinese RPA market has strong growth momentum, with a market size of US$1.02 billion in 2023 and a compound annual growth rate of 64% from 2018 to 2023.

In countries such as India and South Korea, many companies have already begun to deploy RPA applications on a large scale; however, in China, most companies have not yet started to apply RPA, or are only in the planning and trial stages. In other words, there is still a lot of room for Chinese RPA+AI vendors to expand their business.

The application of RPA varies in different industries in the Asia-Pacific region. The manufacturing and retail industries are relatively fast, while the public utilities, media, and financial industries are still in the early stages of understanding and observing. But overall, only 30% of companies have adopted RPA, and there is still a broad market opportunity.

RPA has certain application opportunities in any tedious and repetitive work field.

For example, in financial scenarios, tax returns or deductions are relatively simple and repetitive, and many operations can be automated;

For example, scenarios in which companies interact with customers, handle business, process formalities, terminal retail, surveys on customer preferences, and data collection are all simple, repetitive, and highly automated areas.

Please note that these are just applications of pure RPA. If AI capabilities are added, it will be like a person who climbs mountains and can reach a wider space and see richer scenery than a toddler. The scope of automated work will be larger and the process will be longer, so there will naturally be broader application scenarios and market opportunities.

In addition, due to the large differences in working methods and processes among different industries and companies, the customization requirements of different customers in the RPA field are also very high. Except for relatively universal functions such as recognizing invoices and resumes, the functional requirements of each company are completely different.

Therefore, it is currently difficult for RPA to develop a set of software like SaaS products that can be reused with almost zero marginal cost.

Therefore, this is not a winner-takes-all world, but a world of specialized and customized customer service.

Furthermore, general RPA functions cannot meet the specific needs of different industries. Many industries such as manufacturing, finance, retail, and media have different needs based on industry characteristics. Therefore, just like technology service companies, this field can actually accommodate a large number of players.

This is actually obvious to everyone in the industry, and the composition of players in the RPA track fully illustrates this point.

Track players: both new and old, with complicated backgrounds

The IDC report lists 15 RPA+AI manufacturers from home and abroad.

Comparing the business and development status of these manufacturers, we can find that:

There are many transformation manufacturers, all hoping to reap the benefits of RPA+AI

For example, YisaiQi, which was originally a user behavior and big data analysis software, was listed on the New Third Board in 2016, and entered the RPA+AI track at the end of 2018, and it has only been more than a year to date.

For example, Daguan Data first used NLP technology to provide industry-specific text mining, knowledge graphs, personalized recommendations, and vertical search solutions. In July 2019, it launched the intelligent text RPA product and entered the RPA field.

There is also Laiye Technology, whose early name was "Assistant Laiye", which provides AI assistant services. In June 2019, it merged with RPA company UiBot and entered the RPA+AI field.

Even Alibaba Cloud has launched its own RPA product.

New companies are relatively early stage and have not entered the C round.

In addition, except for early-stage companies, these startups joining the RPA track are basically in a relatively early stage, and there are no players entering the C round.

These companies, with different backgrounds and different generations but all focusing on technology, have entered the RPA track at similar times and made it a core product or even a main product. This is enough to prove that RPA+AI itself has strong technical implementation potential and monetization capabilities.

Seize the opportunity

So, how can we seize opportunities and improve competitiveness in the RPA field? IDC also gave some suggestions in the report, including:

Provide support to industry users in RPA skills, business process identification and use case expansion;

Expand the ecosystem through the establishment of a partner network, etc.

Supports cloud-native architecture, flexible deployment, and improved portability.

If the RPA industry becomes bigger, it will not just be a matter of the RPA industry.

Because if the vast majority of simple, repetitive tasks no longer require human intelligence, then the way humans work will obviously change. For individuals, thinking and communication will become more important than some basic skills, and the way companies train talent will also change in a more professional and sophisticated direction.