In the era of AI, how can meteorological services evolve from "service" to "decision-making"?

On July 16th, Moji Weather held a brand renewal press conference with the theme of "Weather is Different Because of 'Foreseeing'" in Beijing. For most people, the most prominent impression of this press conference might be that Moji Weather, a weather service company that has been established for 16 years, has changed.

Most intuitively, the visual system seen everywhere at the press conference has undergone a significant transformation. At the scene, a letter "O" installation in the "sunrise orange" color stood out against the black background board. The black - orange color scheme is inspired by the changes in the sun and moon time. The orange color symbolizes the warmth of the sun, representing the beauty brought by intelligent meteorological decision - making. The new MOJI font also adopts a geometric form, appearing simple and expansive.

The brand logo is often the most concise distillation of business concepts. The change in the brand logo reflects the iteration of Moji Weather's business concept. By identifying the pain points in the industry, Moji Weather is promoting the industry's transformation from "weather service" to "decision - making empowerment" with the idea of "technology penetrating scenarios and data driving decisions".

As an important manifestation of the strategic upgrade, Moji Weather launched new products for both B - end and C - end users at the press conference.

For the B - end, Moji Weather launched a product of "foreseeable meteorological service" - the AeroMetis Aviation Meteorological SaaS service platform, which has a large language model built - in. It can automatically generate a full - link decision - making plan of "accurate forecast/warning - route change suggestion" based on meteorological data and industry characteristics, helping airlines intelligently avoid adverse weather areas.

For the C - end, the newly launched "AI Life Index" of Moji Weather can deeply integrate meteorological information with life scenarios, providing users with dynamic and intelligent life decision - making suggestions. For example, it can tell you, "Going out at 10 o'clock can avoid showers", allowing you to adjust your itinerary in time and serving as a helpful assistant.

From the brand logo, to the business concept, and then to the new products, Moji Weather is trying to initiate a revolution.

As Xu Xiaofeng, the president of the China Meteorological Service Association, stated at the press conference, "In the current context where AI is driving a historical transformation in the meteorological industry, technological breakthroughs are helping meteorological services achieve the goal of providing high - quality services to the whole society. The 'foreseeable meteorological service' model may promote the entire industry to upgrade from passive response to active planning."

Pain Points of Traditional Meteorological Services

"Weather forecasts are inaccurate", "After reading the weather forecast, I still have to think for a long time about what to wear when going out" ... Perhaps many people have deeply felt the numerous pain points in climate prediction in their daily lives.

Firstly, the problem lies in the traditional weather forecast model. Although it is the cornerstone of meteorological prediction, limited by computing power and the complexity of physical models, the output spatio - temporal resolution often fails to meet the requirements of refinement and scenario - based applications.

Sometimes, there may be significant local weather differences within different areas of a city. Traditional forecasts are difficult to be accurate at the "one - kilometer - one - minute" level, and even more difficult to capture the phenomenon of "minute - level weather changes". This has led to the "generalized" phenomenon of past weather forecasts.

Secondly, the issue is with the data processing mode of traditional weather forecasts. Meteorological data sources are often diverse, including satellite remote sensing, radar echoes, ground observation stations, high - altitude sounding, and user - contributed data. These data have different formats and large volumes. Traditional methods face huge technical and computational bottlenecks in integrating, cleaning, and effectively utilizing multi - source heterogeneous data, making it difficult to fully tap the potential value of the data.

This results in most traditional meteorological services remaining at the level of "informing the weather", lacking the ability to transform meteorological information into specific action suggestions. When users or enterprises need to interpret forecast data by themselves and make decisions based on their own situations, it not only increases the decision - making cost, but also causes huge economic losses due to the lag in time - sensitive business scenarios.

As all industries are rapidly being empowered and upgraded by AI, weather services, as a basic function, are in need of a self - revolution. The key question is, what is the direction?

"Foreseeing", as the theme of Moji Weather's press conference, might be a way to break the deadlock.

If you pay attention, you will find that in the AI era, users are putting forward further requirements for applications. Simple meteorological information prompt services can no longer satisfy the public. People hope that applications can play the role of a "butler", helping them extract effective suggestions in advance from complex meteorological information and avoid risks.

On the B - end enterprise side, the "change" in meteorological services is even more urgent. Enterprises need weather service providers to transform predictable meteorological risks into specific and executable business plans. After investigations, Moji Weather found that B - end enterprises are currently facing three core challenges: the disconnection between meteorological data and business decisions, the lag in disaster response efficiency, and the high cost of defense.

There is a shocking set of data. According to the report of the Ministry of Emergency Management, in 2024, natural disasters in China caused direct economic losses of 401.1 billion yuan, which is almost the annual revenue scale of China's two major home appliance giants, Haier and Midea. Among them, meteorological disasters accounted for 70% of the economic losses caused by natural disasters, directly affecting more than 20 key industries such as agriculture, energy, and transportation.

After identifying the pain points of traditional climate services and understanding the way to break the deadlock, how can AI technology be specifically implemented to effectively improve the new experience of climate services?

Moji Weather's Transformation Practice in the AI Era

To solve the problems of prediction accuracy and refinement in meteorological services, Moji Weather is optimizing the numerical forecast model through AI. It learns complex patterns from a large amount of multi - source meteorological data (satellite, radar, ground observation, historical data, etc.) to conduct more accurate modeling and prediction.

It is reported that currently, the "AI Life Index" and "Heavy Rain Spot Report" of Moji Weather can already output intelligent decision - making plans based on different user scenarios, which is beyond the reach of traditional climate monitoring methods. Among them, the "Spot Report" focuses on the full - cycle management of precipitation within 48 hours, providing hourly forecasts, high - risk warnings, and scenario - based safety plans for VIP members. In addition, Moji Weather is also using machine learning of historical errors to intelligently correct the numerical forecast results, further improving the forecast accuracy.

Improving accuracy is only the basic ability of a weather service provider, and Moji Weather is exploring deeper. Currently, Moji Weather's meteorological service model has changed from the previous "information notification" to "decision - making empowerment".

For example, in the B - end aviation field, AI can automatically generate a specific plan such as "Delaying the take - off by 30 minutes can avoid thunderstorms" according to the weather conditions. On the C - end, valuable personalized suggestions provided by the AI index, such as "The running index at 16:00 today is 92. The wind speed is level 3, which is suitable for long - distance running. It is recommended to carry quick - drying clothes and replenish electrolytes", are also shortening the user's decision - making path.

These interesting transformation practices essentially rely on the solid technological barriers built by Moji Weather in recent years, laying a good foundation in the model layer, algorithms, and data processing.

Firstly, the core technological foundation of Moji Weather is to build a multi - source grid - level meteorological data fusion and deep - learning model. In this model, with the support of deep - learning algorithms, a large amount of meteorological data is efficiently fused, cleaned, analyzed, and modeled, enabling refined meteorological prediction with a resolution of 1 kilometer × 1 minute or even higher. Moji Weather has cumulatively integrated multi - source data of "sky - earth - people". In addition to professional meteorological data, it also includes nearly 1 billion user - contributed weather records, providing high - density samples for model training and ensuring the authority, credibility, spatio - temporal sufficiency, and scenario adaptability of the data.

Moreover, Moji Weather innovatively applies the large language model (LLM) to the "decision - making empowerment" stage of meteorological services. Specifically, they use "weather vectors + sports scenarios + user profiles" to build a Prompt project, enabling the application to dynamically generate personalized and executable "one - sentence prompts + action suggestions".

Taking Moji Weather's "AeroMetis Aviation SaaS Platform" as an example, the large language model automatically generates a full - link decision - making plan of "accurate forecast - route change suggestion".

Before building a good model foundation, the core algorithms are also a solid foundation for forming technological barriers.

For this reason, Moji Weather has proposed an AI intelligent error correction and self - evolution optimization algorithm, which can intelligently correct the systematic errors of the numerical forecast model and optimize the model according to the characteristics of the "local micro - climate" in different regions.

The model, algorithms, and data are the core technological competitiveness of Moji Weather. In fact, no matter how Moji Weather "changes", it cannot be separated from the good foundation laid by the technological base.

How AI Opens up the Trillion - Dollar Meteorological Service Market

After the AI technology foundation is established, Moji Weather's business applications have been fully rolled out.

Vertically, Moji Weather announced at this press conference that it has launched an internationalization strategy. It is reported that in the future, Moji Weather will integrate global resources, build a global meteorological service system, and provide intelligent life decision - making suggestions and targeted solutions for users in different countries and industries. Currently, Moji Weather has completed multi - language adaptation in more than 30 countries and regions, including Chinese, English, French, Spanish, and Portuguese, and has entered the capability verification stage.

Horizontally, the upgraded Moji Weather has new involvements in both B - end and C - end businesses, constantly inspiring new imagination.

On the C - end, AI is promoting the transformation of user services from "general forecasts" to "personalized decisions". According to internal data, since the launch of the "AI Life Index" function, the average daily usage frequency of users has increased by 35%. The click - through rate of scenario - based functions such as the "Running Index" and "Cycling Index" accounts for more than 60%. In the future, Moji Weather plans to further explore the C - end business potential through member segmentation.

On the B - end, Moji Weather has been deeply involved for many years. Since its establishment in 2016, it has been widely involved in more than 20 industries, providing professional meteorological solutions for more than 200 enterprise customers.

The new "foreseeable meteorological service" model will undoubtedly make commercial cooperation deeper. AI - enhanced meteorological data can be more deeply integrated into the core business processes of enterprises, especially in fields such as aviation, rail transit, energy dispatching, agriculture, and urban management. By exploring the deep relationship between meteorological data and industry scenarios and providing customized solutions, new value space can be created.

A business case can prove the deepening value of Moji Weather on the B - end. Moji Weather once developed a special thunderstorm forecast system for the special scenario of high - altitude airports, helping airlines accurately evaluate flight operation conditions. Practice has shown that this system reduces the return and alternate landing rate of flights at high - altitude airports by 5% - 10%, delivering millions of passengers to their destinations on time every year.

Jin Li, the CEO of Moji Weather, said that ToB and ToC are the two key aspects of Moji Weather's business strategy, which complement each other and jointly form the complete ecosystem of Moji Weather's foreseeable meteorological service.

Jin Li, the CEO of Moji Weather, delivered a keynote speech

In his view, C - end users, as the cornerstone of the brand, provide a large user base and high - frequency usage scenarios, accumulating a large amount of meteorological data and insights into user behavior. These valuable data will later feed back into the optimization of Moji Weather's AI algorithms and models, making the meteorological prediction ability more accurate.

The investment in the B - end more reflects Moji Weather's understanding and value of the industrial chain. "Moji Weather will more deeply understand the business processes and pain points of various industries, deeply integrate meteorological data with industry scenarios, and provide customized and executable decision - making suggestions. It not only helps enterprises improve operational efficiency and reduce risks but also hopes to create greater economic and social value," Jin Li said.

Conclusion

Actually, the "change" of Moji Weather is not unexpected but a natural choice for the enterprise to adapt to the trend in the AI era. Obviously, in the era of information overload, both individuals and enterprises really need not just simple data, but "solutions" that have been deeply processed and can directly guide actions.

Riding on the wave of AI, the model upgrade of Moji Weather is opening up new market opportunities. Data shows that currently, the scale of China's meteorological economy has exceeded 200 billion yuan. According to the prediction of the China Meteorological Service Association, by 2025, this figure will jump to 300 billion yuan, and the entire industry is running at a high - speed annual compound growth rate of 11.6%.

The signal conveyed by these figures is that meteorological services are upgrading from the traditional "barometer" to a new - quality productivity engine driving industrial transformation.

Moji Weather has also figured out how to "conquer the market" with the new model. They have two answers:

Firstly, the low - altitude economy. As drone delivery and air taxis gradually become a reality, meteorological support is the lifeline of low - altitude safety. In the future, Moji Weather will provide targeted data such as thunderstorm and wind shear warnings to assist business decision - making in the low - altitude economy industry.

Secondly, the new energy revolution. In the future, Moji Weather will comprehensively consider data such as wind speed and irradiance that can guide power generation dispatching, help improve energy utilization efficiency, and use data translation ability to transform meteorological language into industry productivity.

The "old" industry of weather services is showing new vitality in the AI era.