What kind of sparks can the collision of ChatGPT and RPA produce?

Currently, AIGC (artificial intelligence content automatic generation technology) is developing faster and faster. Especially in 2022, its development speed is astonishing and AI painting and AI writing once became the focus of people's discussion. Recently, another technology belonging to the field of AIGC has "occupied" our field of vision.

This is a conversational product based on the new LLM technology, ChatGPT, launched by the American artificial intelligence research laboratory OpenAI more than two months ago. It is like a spring thunder that rises from the ground and quickly becomes popular at home and abroad. It reached 100 million monthly active users in January this year, and the growth rate of users is unprecedented.

What is ChatGPT? what can be done?

ChatGPT, in simple terms, is an artificial intelligence that can "talk and talk" and can be used as an automatic text-generation tool. This is similar to artificial intelligence assistants such as Siri, Xiaoai, and Xiaodu who can talk to people, but they are even better.

The GPT-3.5 model in ChatGPT learns a large amount of natural language data and can understand the syntax and semantics of the language. When a user enters a question, ChatGPT converts the question into data that a computer can understand and processes it using the GPT-3.5 model. The GPT-3.5 model will generate a set of answer candidates based on the language data and select the best answer from it. Finally, ChatGPT will return the generated answers to the user in the form of natural language.

Therefore, what surprises and satisfies ChatGPT most users is its excellent intent recognition and natural language understanding capabilities, as well as the resulting practical problem-solving capabilities.

It can learn the knowledge of the entire Internet, has a massive information knowledge base, knows astronomy, and geography, and can conduct conversations by learning and understanding human language. It can also interact according to the context of the chat and assist humans to complete a series of tasks. :

For example, students can use it to write homework, workers can use it to write weekly newspapers and emails, programmers can use it to write codes and check program errors, and media people can use it to write scripts... These daily practical problems, "IQ Online" ChatGPT can solve it.

Then, behind the popularity of ChatGPT, we also began to think about a question: the similarities between AIGC and RPA are automation, AICG is the automation of content, and RPA is the automation of business processes. Can the two be combined? What kind of sparks will they collide with?

When RPA and ChatGPT meet, what kind of sparks can they collide with?

RPA (Robotic Process Automation) is equivalent to a "digital employee". That is, let the robot execute the customized process in advance, to replace the human to perform a large number of repetitive work procedures.

At present, major RPA manufacturers are committed to lowering the threshold of RPA use, especially last year, smart RPA has indeed achieved a qualitative leap in ease of use from "drag and drop" to "point selection", and the emergence of ChatGPT, Bring us more possibilities: Can it be used to lower the threshold for using RPA again and make RPA business automation more efficient? (Supplementary Note: Real Intelligence has also done some early exploration and attempts on the combination of ChatGPT and RPA)

So here, let's take RPA, which is intelligent "what you see is what you get", as an example. Let's take a look at the difference between RPA combined with ChatGPT and not combined with ChatGPT:

Without ChatGPT, only RPA, users need to generate an automated process (threshold) by artificially "drag and drop" or "click to select" according to their own needs, and then RPA digital employees execute the established process and monitor it, and finally complete the closed loop of business process automation.

The combination of ChatGPT and RPA, users only need to tell ChatGPT their needs in colloquial expressions, and then ChatGPT intelligently understands user needs through natural language processing technology and converts them into expressions in the RPA field while realizing practical implementation After the completion of the plan (that is, the generation of automated processes), RPA digital employees execute the established plan, monitor, judge, and make decisions, provide users with automated work services, and finally complete the closed loop of business process automation. Radically increase efficiency, reduce operational risk, and improve response times and customer experience.

For example, if a user wants to use RPA to crawl image data on a web page, they can ask chatGPT:

"Using RPA to grab pictures in web pages, how to use RPA components?"

ChatGPT uses parameters learned based on training data and learning strategies, as well as information about RPA components in user questions, to answer the RPA process required to use RPA to capture web page images.

The answer may include multiple steps, each of which provides an RPA component and a colloquial parameter setting method, and these steps finally form a complete RPA process for capturing images on web pages.

Summary: ChatGPT intelligently understands user needs and generates RPA processes, and gives feedback to users in colloquial expressions, lowering the threshold for users to understand and use RPA.

In addition, since ChatGPT can be used as an automatic text generation tool, scenarios involving automatic text generation, creation, and automatic question-and-answer can be realized by ChatGPT, and it can cooperate with RPA to complete business process automation more efficiently and improve user experience.

Therefore, I think that ChatGPT or ChatGPT-like will become the best partner of RPA shortly. What do you think?