The birth of ChatGPT has detonated the world’s technology circles. As it has become something everyone can talk about on the street, the giant model has also inspired a key question in the business world: what is it for? How should my business use it?
Hu Shiwei, the co-founder of 4Paradigm, remembers that at the 4Paradigm SHIFT summit in 2022, more than 30 corporate chairmen were essentially discussing one thing-how to break the embarrassing situation of “unsatisfactory results of digital transformation” .
At the summit of an AI company, entrepreneurs came to listen to technology, but they were actually thinking about the simplest problems in business. This is the most real look in the business field. Hu Shiwei said that the mission of technology is to solve real problems, “Just like this year, everyone is talking about how powerful the big model is, and customers will blurt out when they hear it – since it costs money, how does this thing affect my survival and expansion? How useful is it?” And this, is precisely the real question.
Hu Shiwei told Geek Park that as a company positioned to use AI and other technologies to help customers with digital transformation, they have different thinking in the slightly frenzied large-scale model market-he chose to focus on “large-scale models” What kind of positive changes will it bring to the enterprise?”
If the large model can only bring some “short-chain” productivity tools such as Vincent Wen and Vincent diagrams, it is obviously not its greatest value. How to play a role in the “long chain” of business management may be the most noteworthy trend today.
Hu Shiwei just feels that the biggest breakthrough brought about by the paradigm revolution of large models is the possibility of a huge acceleration of the digital transformation of enterprises, and even the space for rapid improvement of the “scientificity” of enterprises.
The digital transformation of enterprises is an industry-oriented development trend. Centering on the North Star indicators of enterprises, from the four key steps of the enterprise’s “strategy” and “strategy”, to “execution” and “evaluation system”, comprehensive and profound transformation The goal of the transformation is to change the core competitiveness of the enterprise from quantitative to qualitative, and take the lead in becoming a “next-generation enterprise”.
Hu Shiwei said: “In the past, after anchoring the strategy, we used AI to help companies solve strategic problems, such as what tasks to perform and which way to perform tasks is better. However, the strategy has always been limited by “execution efficiency”, or even It is simply not implemented effectively within many companies.”
The emergence of a large model is precisely the key to solving the problem of “execution” among the four elements.Just last week, 4Paradigm exposed the large-scale model product “Shishuo”, and proposed AIGS (AI Gerenated Software) technical ideas, focusing on solving the problem of efficient implementation of corporate strategies through digitalization.“It will even bring about a qualitative change from “month-level” to “day-level” on the executive side of the enterprise.”Hu Shiwei pointed out.
At the same time, the reconstruction of enterprise software with AI has opened up a new change in the huge enterprise software market.
01 Use AIGS to liberate “silent data” and improve “thinking chain”
In Hu Shiwei’s view, after a large number of enterprises undergoing digital transformation have clarified their strategies, the difficulty in implementing them largely comes from the problem of “data usage efficiency”.
It is very realistic that today a large number of enterprises can effectively use less than 5% of their data. The reason is simple. Within the enterprise, there are three paths for employees to communicate with the enterprise: the first is to receive training, the second is to communicate with people, and the third is to use the system. But in the past, in most enterprises, the third system path accounted for less than 5%, because the business communication between employees and the dialogue between employees and customers almost all occurred in unstructured (data structure irregular or incomplete) and unsystematic scenarios.
“One trend is that the more excellent a company is in digital transformation, the higher the proportion of employees talking to the system.”Hu Shiwei pointed out that in industries such as finance, retail, transportation, and residential platforms, in some companies with a high degree of digitalization, the proportion of employees working through the system path has exceeded 50%.
“We used to say that we should not only look at transaction data and result data, but also look at process data. Because real high-quality data is generated by the interaction process between people, not “forms” or PPT. In the past, about equals form.”
For this kind of process data, in the past, Party A could only force excellent employees to fill in “how to do a good job with customers” in the CRM. Then let the poor performers go to CRM to do the same thing. But the core dilemma here has always been that the better the front-line employees, the less time they have to use CRM.
“How people do business and how people use software are two different things.” Strong employees and weak employees get things done in different ways, and they just stuff the results into the software. In other words, even though there was an AI strategy model in the past, from the perspective of execution, even if many employees resolutely refused to do it, the company had nothing to do.
In addition, what is the biggest problem that an enterprise IT or information department usually faces-“I have been asking you for several months, why haven’t you done it yet?”
Hu Shiwei said that the internal development cost of the enterprise is too high, and there are few excellent PMs and architects. Therefore, many Party A companies only put limited resources on the top-level strategy to develop key systems, and other low-value parts are handed over to enterprise software. supplier.
“However, software manufacturers often fail to meet the needs, which makes Party A very dissatisfied, so they start recruiting people to work overtime every day.” This leads to a vicious circle and slows down progress.
So, how to promote employees to actively “execute” on the system, how to let excellent employees drive poor employees to improve their business capabilities, and how to move the execution of most of the company’s strategies (not only high-value strategies) to the system , is a common problem facing digital transformation enterprises.
“In the past, consulting companies went to interview outstanding people, and then generated a standard behavior manual, so that those who lagged behind could learn to do it. What is the standard behavior manual? It is the chain of thinking.”
Hu Shiwei believes that the ultimate goal of AIGS, in a sense, is to find the best presentation form of “human excellent thinking links” in the computer world:
Excellent employees give continuous feedback to AIGS (maybe in the form of a dialog box), giving the system the correct logic for solving problems. The latter guides underperforming employees, and even “asks the wrong question, and the system will correct the way he asks.”
This shows that the dialog box already knows what the usual logic for solving a problem is from good employees. “As a human being, when you receive a complex job, you will infer in your brain the subtasks to be executed step by step, and execute them step by step; but if you switch to AIGS and face the same job, if it “sees” enough People (accumulated data) can summarize this routine and form a chain of thinking.”
Chain of thought container
Software ultimately serves the efficiency management of enterprises and organizations. With the help of the chain of thought, AIGS means a possibility beyond filling out forms and beyond interface design. From “people adapting to software” to a coach who corresponds to a chain of thought for each employee.
The biggest change it brings is to magnify the scientific nature in an organization, but it is also in line with human nature. This kind of true “partner” that is meaningful to employees will make employees more willing to work on the system and put all Business processes are left on the system.
Hu Shiwei said that once the large model system of an enterprise is put into use, it will always be in a state of self-iteration: a large amount of active and effective data feedback will be deposited in the system in real time, which is equivalent to incorporating unstructured data into the overall digital governance system of the enterprise. There is huge room for improvement.
“95% of the data that has not been collected will definitely make the digitalization process of the enterprise really turn like a flywheel on the basis of a large amount of industry knowledge and internal feedback closed loop.”
02Behind “Change Communication” is “AI BP’s new productivity
According to Hu Shiwei, the co-founder of 4Paradigm, the presentation form of AIGS in the first stage is a “dialog box”. This dialog box is not a box for chatting, but a box for completing various tasks in the form of dialogue. , which is the enterprise AI assistant Copilot (co-pilot).
Behind the “dialogue box”, it not only means a change in the form of interaction, but also a complete change in the way employees communicate.
The logic of using AIGS to plan an enterprise’s internal information system is first to transform or create a new communication link between employees and the system through the “dialog box” to improve the efficiency of information and strategy transmission. This is equivalent to providing each employee with a BP (Business Partner) in all enterprise-level software systems to help employees call functions/data built into traditional software. When working, the “AI business partner” listens to the instructions of the employees through dialogue with the employees, bringing an order of magnitude improvement in execution efficiency.
But at a higher level, these “AI BPs” are actually serving corporate strategies through digital connections with strategies. Obviously, it is precisely because of the breakthrough in the ability of the large model that the company can assign a strong “AI BP” (AI business partner) role to each employee, which can not only help employees improve efficiency and solve problems, but also achieve cooperation with the company. Strategy and execution standards are highly aligned.
This kind of “luxury organizational structure” originally only available to mature large companies can become an important productivity for more companies to improve efficiency. After employees and organizations get this kind of blessing, the efficiency of the organization will be greatly improved.
Hu Shiwei is excited that for many years, 4Paradigm has been making AI decision-making models in software, and now the paradigm revolution of large models and the emergence of AIGS technical ideas just complement the ability to execute strategies. Excellent AIGS capabilities must be combined with excellent AI decision-making capabilities, and strictly follow the direction of the company’s core competitiveness, so that the company can better achieve its strategic goals.
“Corresponding to managers and employees in the enterprise, decision-making AIIt is the digitization of the role of managers, while generative AI is the digitization of the role of employees. “
Then in the field of enterprise software, what kind of chain reaction will be produced due to the emergence of AIGS and fundamental changes in communication methods?
Hu Shiwei believes that at the current stage, AIGS is to change the way employees communicate with the company, rather than changing the way employees complete tasks. The latter needs to rely on the functions and processes that have been accumulated for many years in their respective fields by industry software service providers. sexual advantage.
Therefore, the relationship between 4Paradigm AIGS and industry enterprise software is symbiosis and mutual assistance rather than exclusion.
At 4Paradigm’s large-scale model conference, founder Dai Wenyuan mentioned that the idea of AIGS stems from a relatively intuitive cognition of many people——domestic B-end software has goal-oriented functions, but it is often lackluster in terms of experience. .
In his view, if you want to implement a function, you have to click the system menu three times to complete it, and the experience is not as good as Chat (dialogue form). And if you can install the heart of generative AI in the software – you say “I want to be reimbursed”, take the ticket to the system, it asks: “Who did you eat with?”, you answer: “Zhang San”, it says: “I accept”. Finish. This experience is obviously much better than the current reimbursement system.
Whether it is internal OA, ERP, CRM, BI, or external business management systems, there are too many software that can be transformed. For example, a catering store wants to check the sales of a certain SKU on the day; the workshop director of a factory wants to know whether there is any illegal operation on the assembly line-in the past, they had to manually find the interface, click this, and pull that. But now, just ask in the dialog box: Has the factory had any violations in the past day?
At the press conference of 4Paradigm, an aviation manufacturing company asked questions by voice, “Help me find similar parts”, “Give me the assembly plan of these two parts”. In , all similar 3D digital model parts are found, and various digital model assembly schemes are given.
The relevant person in charge of the company explained to the media that their core intellectual asset is “3D digital modeling”. As long as the 3D digital model exists, even if a disaster occurs and the factory disappears or the employees are dismissed, it can be rebuilt within three months. “A small step for AIGS is already a big step for our industrial design software,” said the relevant person in charge.
3D model query and assembly
Hu Shiwei believes that technical ideas and new paradigms such as AIGS will definitely enhance the competitiveness of enterprise software suppliers.
“We will work with enterprise software partners in the future. In the past, many software companies did CRM, supply chain, and BI systems, and our relationship with us was relatively indirect. But today, in the AIGS field, companies looking for people to do systems also have our business.”
In other words, in the past, the difference between traditional software suppliers and 4Paradigm was “manpower-intensive” and “technology-intensive”; then in the current era of large-scale models, everyone will become a computing power-intensive enterprise.
03 from AI Paradigm Revolution to Organizational Paradigm Revolution
When AIGC becomes a global hotspot, the AIGS concept, which focuses more on getting through the digital strategy and execution of enterprises, is an important thought entry point for large models to enter the 2B field.
Looking at it from a more grand perspective, human business organizations have actually been using technological tools to continuously carry out “paradigm” evolution, from the agricultural age to the industrial age, and then to the digital age. The emergence of enterprise management software had a major impact on enterprise organization and management thinking, and today’s breakthrough in large-scale model technology will mean that the flexibility and capabilities of software will be further enhanced. And, it can deliver this capability to more business organizations more universally.
If Microsoft is taking the lead in infiltrating large-scale models into the personal office field in the world, then the fourth paradigm is to use large-scale models to penetrate into the efficiency field of enterprise software.In the past, enterprise software served as a carrier of abstract best practices, and abstraction meant rigidity and solidification. Now relying on the large model, there is an opportunity to create a software workflow with excellent experience, minimal interface, and self-iteration, and even create important productivity such as “AI BP” to enter the organization to help the enterprise’s strategy to be implemented most effectively.
The AIGS technical idea is to generate a high-level thinking chain through a large model, strengthen the collaboration between employees and the system, and strengthen the effectiveness of businesses. From the past strategy model, it can radiate and affect the middle layer——to be able to radiate and influence up to the strategic layer. (What the CEO has to do), down to the executive level and evaluation system, this is an exploration of a new paradigm of organizational management.
Its greatest significance may be that it opens up a new solution to the “eternal problem” of business organizations – the organization itself needs to be scientific, but in fact, the “people” cannot be completely removed from this organization, and there is only science at all. Sexual organizations also have no vitality.
The core of this brand-new software interaction and collaboration mechanism supported by the future large model is how to effectively cooperate with humanity and science in an organization, so that the two can shine together.This may be issues such as strategy and execution, stability and innovation, excellent employees and qualified employees, exceeding expectations and bottom line, etc., using a digital and AIConstructed new management ideas.
Under this idea, the scientific nature of an organization can be enhanced with lower cost, higher efficiency, more flexibility, and better fit for business. Then based on this, there may be more efficient commercial organizations appearing in batches.obviously，AI The paradigm revolution of 2019 is bringing about a paradigm revolution of enterprise software, and ultimately, it may be a paradigm revolution of organizations.
“In the final analysis, China’s To B industry, no matter what kind of growth model or business model it is going to, must essentially be oriented towards the improvement of core competitiveness and the creation of organizational value. AIGS can bring about an order of magnitude improvement in data acquisition efficiency and execution efficiency , Computing power elements and data elements will gain unprecedented opportunities for synergy. With the strong desire for digital transformation and high-quality development of enterprises, various enterprises in the ecological chain of To B software may also usher in a historic change worthy of attention .”