Article | TechTrends

"Super AI will arrive in 'thousands of days.'"

Recently, Sam Altman, CEO of OpenAI, made a rare long post on social media, predicting this. Previously, many experts predicted that super AI would arrive within five years, and Altman's prediction may make this timeline even more optimistic.

Super AI, also known as Artificial Super Intelligence (ASI), refers to AI with intelligence far surpassing that of humans. Altman believes that "it can truly learn the distribution patterns of any data" and "generate the underlying rules of any data distribution."

"Data" has become a keyword in the human vision of ASI. On the path to ASI, more and more tools or applications highly related to data are achieving rapid development. Among them is Business Intelligence (BI), which deeply mines data value within enterprises and supports data analysis. Its progress resonates with the evolution of AI's ability to handle data.

In China, according to the Forward looking Industry Research Institute , it is estimated that by 2029, the market size of China's business intelligence software will exceed $3 billion, with a Compound Annual Growth Rate (CAGR) of around 20% in the next five years.

However, while the value of BI is widely recognized, the long-standing challenge of "product strength" cannot be ignored. For example, the high level of professionalism leads to insufficient coverage within enterprises (only specific groups have the ability to use it), and BI's output capabilities are limited, including a lack of diversity in report formats and insufficient depth of data analysis.

At this point, AI, which is on its way to ASI and experiencing rapid development, has found its niche.

From ANI (Artificial Narrow Intelligence, including various assistants) to AGI (Artificial General Intelligence, currently being explored) and ultimately to ASI, AI's understanding and value mining of data is a deepening process. It gradually realizes scenario-driven value across various industries, ultimately leading to Altman's prophecy. Intelligent BI, a combination of AI represented by large models and BI, has become a typical representative and annotation of this trend.

With conversational BI (ChatBI) as its primary manifestation, intelligent BI is becoming the pursuit of many vendors. For example, Smartbi released its new generation of intelligent BI based on AI Agent in early August, Smartbi AIChat Baize, a convenient and efficient tool that enables data analysis through ChatGPT-like conversational interactions.

With the support of large models, BI's "product strength" is ushering in new breakthroughs.

Intelligent BI Strengthens BI's "Product Strength"

From a product perspective, the "upgraded" value brought by the combination of large models and BI is manifested in several aspects.

1. Decreasing Usage Thresholds

Intelligent BI targets business personnel through conversational BI, allowing business needs to be directly satisfied and achieved without the need for translation by data analysts or IT engineers. Taking Smartbi Baize's product experience as an example, simple instructions can generate the desired bar chart:

Without dragging and dropping components, creating dashboards, or reports, it is more intuitive, flexible, and efficient.

2. Enriching Presentation Forms

Intelligent BI can automatically generate more types of charts and reports, helping users better understand data.

For example, Smartbi Baize can invoke all 30+ graphical components in Pyecharts, such as bar charts and pie charts, using natural language, and can also adjust titles, legends, etc.

With intelligent BI, users can explore and develop more presentation forms, making business support and data mining more effective.

3. Strengthening Customer Value in Data Analysis

If interactive simplicity and diverse output forms are merely the "exterior" of data analysis products, data accuracy is the core value that products can bring to customers.

Currently, vendors such as Microsoft and Smartbi have developed intelligent BI that adds one or more data analysis capabilities to the existing dashboards and reports of traditional BI, such as time calculations, attribution analysis, data prediction, and data interpretation.

Taking attribution analysis as an example, in Smartbi Baize, simply asking about the reason for an abnormal contract amount in a particular month will provide possible explanations:

In practical applications, when a securities company found abnormal indicators on its dashboard during a business analysis meeting, it could directly inquire with Smartbi Baize to further analyze and investigate the abnormal indicators.

This demonstrates that intelligent BI is evolving BI from simple descriptive and limited diagnostic analysis to in-depth diagnostic analysis combined with internal and external knowledge, and even directly providing "prescriptions" for business personnel and managers.

Of course, in addition to these three points, with increasing concern about data security among enterprise customers, intelligent BI also excels in access control. For example, Smartbi Baize controls data access by interfacing with a unified data model, satisfying financial-grade data access control. With support for private deployment of large models, enterprises no longer have concerns about data security.

Behind "Product Strength" Lie Multiple Thresholds for AI+BI

Strengthening product strength, large models are bringing new growth curves to the BI industry and may even become the primary growth driver in the future market. However, AI+BI is not something that "anyone can follow the trend of"; there are objective thresholds.

In summary, this results in one outcome: the already intense competition in the BI industry will become even more intense, making it difficult to maintain a stable landscape. Some vendors are leveraging the opportunity of AI+BI to achieve overtaking, which not only provides domestic vendors with an opportunity to reverse the dominance of established international vendors but also allows some vendors to occupy a more favorable position in the industry based on the unique advantages of intelligent BI.

The former scenario involves BI vendors that have developed for many years abroad gradually being surpassed by domestic vendors. According to the "China Business Intelligence and Analytics Software Market Tracker Report, 2023H2," the current top 10 BI vendors are FineReport, Microsoft, Baidu, SAP, Smartbi, Yonghong, Inspur, Salesforce, IBM, and Yixin Huachen. This indicates that the market has shifted from being dominated by foreign vendors alone to a competitive landscape involving both domestic and foreign vendors. Notably, among the top 10 vendors, only Microsoft and Smartbi have currently launched direct intelligent BI products. Furthermore, against the backdrop of the information technology application innovation ( Xinchuang ) industry, domestic vendors have even more prominent advantages, with various factors combining to significantly impact the industry landscape.

The latter scenario is also evident in this IDC report, where the growth rates and rankings of vendors do not match, and the market is far from the Matthew Effect. For example, Smartbi, ranked fifth overall and second among Chinese BI vendors, leads the industry with a 45.7% growth rate, far exceeding other peers and more than 12 times the market average. This strong performance is closely tied to the development of intelligent BI.

Analyzing this changing trend in the industry landscape, several thresholds for successful intelligent BI gradually emerge. Since it's AI+BI, "AI," "BI," and the "+" are all indispensable.

1. "AI" - Large Models as an Advanced Step in AI's Application to BI

Intelligent BI must first be built on long-term "intelligent" integration.

Smartbi, which has experienced rapid growth, began its forward-looking layout of AI early on. In 2019, it was the first in the industry to integrate AI and data analysis and implemented conversational analysis capabilities based on self-developed Natural Language Analysis (NLA). Subsequently, it was also the first to release the ABI platform that integrates AI into data analysis, leading to the 2023 launch of a conversational analysis large model version combining large model technology and BI products, ultimately facilitating the emergence of Baize.

Behind this lies the fact that intelligent BI's utilization of large models should be a realization of an "opportunity" for AI+BI. Approaches that directly seek to "add" large models to BI products without a foundation may be difficult to implement, as this is not a trend but rather a prepared battle.

This is evidenced by Smartbi's relatively prominent technological advantage in conversational analysis compared to the industry. For example, Baize is functionally similar to the Advanced Data Analysis capability offered on ChatGPT-o4 - users specify CSV or Excel files, ChatGPT understands the question, and generates Python code to solve the user's problem (Figure: Advanced Data Analysis example, checking weather data, analyzing data, and creating charts).

The computational power stratification required behind this becomes Smartbi's unique AI technological advantage in intelligent BI, ultimately enabling complex data analysis logic rather than simply querying numbers. Obviously, this can only stem from long-term AI layout .

2. "BI" - The Core of AI+BI Lies in BI Capability Accumulation

From the natural language interaction in text boxes, it is evident that AI+BI's enhancement of product strength essentially addresses the "last mile" of BI. To some extent, AI+BI is AI for BI, making BI capability accumulation the core and crux.

Breaking it down, BI accumulation essentially involves data models and indicator models, which are precisely the advantages accumulated by rapidly growing Smartbi in its underlying models. For example, in data models, features such as the star schema, multiple fact tables, and OLAP analysis capabilities can directly compete with Microsoft Power BI's Dax. In indicator models, stronger encapsulation capabilities for data models provide multi-business perspective management capabilities that better integrate with scenarios.

In reality, this also demonstrates that in the AI+BI combination, BI goes down to address the integration with big data platforms, data lakes, and data hubs at the underlying data architecture level, extending the basic data architecture. Meanwhile, AI goes up to solve intelligent analysis applications and business execution issues, forming a complete system.

Figuratively speaking, this is akin to e-commerce plus logistics. Users simply place orders and wait for delivery, but behind the scenes lies a complex e-commerce operation system and a nationwide or global logistics system, the invisible yet core components.

3. "+" - Scenario Practice as the "Wall of Lamentation" for Latecomers

The BI vendor with the leading growth rate is domestic runner-up Smartbi. In the BI market, another noteworthy aspect is the "2023 China Banking IT Solutions Market Share Analysis Report" released by CCID Consulting, a subsidiary of the Ministry of Industry and Information Technology's China Center for Information Industry Development (CCID). In the 2023 banking business intelligence software product market, Smartbi also ranks first in market share.

Industry know-how is a crucial aspect of how AI+BI is "+". Only through in-depth practice oriented towards industries and scenarios can intelligent BI be done well.

In specific industries, without practical experience, latecomers struggle to gain a competitive edge, which applies to both new BI players and established BI vendors facing certain industries where they lack practical understanding. The BI market share ranking in the financial industry is precisely a product of this phenomenon, with the market ultimately trusting and choosing only BI vendors who "know the industry."

Looking back, if industry knowledge is energy storage, then large AI models become the fuse for value explosion. As evident in the Smartbi case, its ability to stir up the market landscape and achieve significant growth is directly related to its 13 years of experience in BI data analysis and accumulation of over 5,000 customer experiences. Beyond finance, Smartbi has also completed knowledge reserves in various industries such as government, manufacturing, healthcare, and education. According to official information, with intelligent BI, it has successfully signed contracts with a leading securities firm and a large integrator, and is continuing to follow up with over 100 large customers. This directly demonstrates the immense value of industry know-how to BI vendors.

It is thus evident that to excel in intelligent BI, "AI," "BI," and the "+" are all indispensable. They are inherent "barriers" of intelligent BI since its inception, and vendors with comprehensive endowments are leveraging them to seize new market opportunities.

Conclusion

Starting in 2023, AI and large models are transforming the era's landscape across various industries. Question-and-answer queries and intelligent assistants (like Copilot) have significantly improved product usability and reduced the operational complexity of professional business analysis.

Intelligent BI is precisely amidst this wave. It not only helps users extract value from data and realize enterprises' digital operations but also aids business personnel in optimizing processes and providing better service experiences for customers. Intelligent BI directly enhances enterprises' competitiveness and efficiency across various dimensions, with even more profound changes and value for enterprises.

With a comprehensive push from "AI," "BI," and the "+" and the efforts of vendors like Smartbi, BI is continuously integrating with large models. The treasure trove of data is further mined by intelligent BI, and enterprise business decision-making is advancing towards higher levels of intelligence and automation. Meanwhile, intelligent BI is becoming a winning strategy for BI enterprises like Smartbi to comprehensively enhance their competitiveness, serving as a weapon for individual enterprises and even the entire industry to achieve a second growth curve.