Introduction Start the adventure of BI+AI, but return to the essence of BI
The essence of business intelligence is to empower the business. From the very beginning, business intelligence, which takes data analysis and business analysis as its main work form, is a business action, or more precisely, a business goal.
BI is a continuous process of building an organization to improve the efficiency of information flow, and the form of data can reduce the dimension of complex business logic, so that valuable insights can be more efficiently reached to the front line, so that the right decisions can emerge directly from the front line of the business. The capabilities of AI are not magical, and the adoption of NLP in BI can be traced back 10 years, but this leap forward in LLM is indeed promising to push BI through the last mile of mass adoption. Behind the evolution of technology is driven by demand, and it is foreseeable that the long-term needs of the business side to be empowered are approaching the ideal state in the next step of this technology wave.
hengshisense5.0 Innovation Overview
Business metrics are the key to AI
In our buildIn the process of major releases, the boundaries and limitations of AI capabilities are the key directions to explore. After the verification of several customer real scenarios, it can be objectively said that the direct connection of the large model to the data layer is incomplete, and the landing effect cannot meet the requirements of the business analysis scenario for the accuracy of Q&A, and there is a definite threshold here.
The ability of large models is like a strong heart engine and large artery vessels, which need to be matched with fine capillaries to accurately construct the logical context of business analysis. The know-how of domain modeling knowledge needs to be able to be implemented in a structured manner, and the most appropriate way here is to use an indicator system to build it.
The management of business metrics is itself a requirement of a modern BI platform, and this is the point of difference between reporting BI and analytic BI. In combination with large models, this layer becomes even more important. The precise calculation logic definition of each business indicator and the business connections contained between the indicators just make up for the knowledge gap of the large model in the face of data, guide the large model not to guess or bias, and finally achieve the intelligent question and answer function we expect.
AI Copilot makes BI truly business-oriented and empowering businesses
In the process of market maturity when BI is becoming more and more mainstream capabilities of enterprises, analysts still maintain a very professional attitude as actual users of BI tools, and the cost of skill point training is high, and the comprehensive requirements for personnel portraits are high, which greatly limits the popularization of BI.
In the rapid evolution of generational change, it is necessary to grasp the long-term constant demand scenario of the enterprise, and BI is fortunate to have such a scenario that has both certainty and long-term antifragility. The concern that AI will replace BI is a misunderstanding, what AI really replaces is a rigid, static, and backward way of report display, AI promotes BI to business, scenario-based, and real-time, and BI combined with AI is the gospel for business personnel, and the jump in interactive form greatly reduces the threshold for use, removing the biggest obstacle to the basic popularization of this technology.
Through the unique BI PaaS open architecture, it supports more diverse data collaboration scenarios
Hengshi has been committed to the open architecture design of BI PaaS, which is a deliberate boundary protection mechanism that distinguishes between the standard domain of data capabilities and the professional domain of industry analysis. The continuous addition of AI capabilities will further strengthen the depth of empowerment of platform products to partners.
From the perspective of best practices, usually in the second and third years, Hengshi's partners have already surpassed the height of visual kanban and began to build analytical productized scenario applications, and the efforts of using Hengshi to reduce the cost of production and research and repeated investment can obtain great benefits. As we predicted, in the future, every enterprise-level business SaaS or software product will have data analysis and BI attributesFurthermore, the future business software will have intelligent characteristics, from first-class to data-based to intelligent, which is an irreversible evolutionary trajectory. hengshi senseThe name starts from Day 1, which refers to the direction of data intelligence, helping each partner's business scenario to install the intelligent engine of Sense & Response, and the open architecture in the form of PaaS is in 5X's journey will continue to be enhanced, empowering the cooperative ecosystem to build more diverse intelligent scenarios, and such an ideal vision worth working on can be best described as:Business Intelligence.
hengshisense5.0 highlights at a glance
hengshi sensePublished,The version is a follow-up to 4A major version upgrade after version 0. In this version, Hengshi AI analysis assistant is introduced, new report application types are added, and a new insight analysis module is launched by integrating the indicator library. A lot of optimizations have been made in interaction, many highlight functions have been added in visualization, data management, application management, etc., and the names and layouts of functional modules have been sorted out to better clarify the role of each module in BI analysis. The product module is divided into four categories according to functions: analysis, indicators, data, and basic management, which more clearly shows the product architecture and the life cycle of data analysis.
Let's take a look at these new features:
Indicator platform. hengshi ai copilot
Reports app. Visual analytics.
Data management. Application management.
Indicator platform.
hengshi senseThe indicator platform quickly builds a business indicator system based on the metric capability, unifies the indicator caliber, and allows business users, data consumers, and decision-makers to use the same metric caliber. The metrics platform centralizes the management of indicators, and quickly forms a business indicator analysis dashboard through the combination of preset metric analysis templates and business scenarios.
Strengthen the management capability of the indicator platform and open up the indicator system between the platform and tenants, aiming to improve the efficiency of the use of indicator data, empower businesses to carry out insight analysis, and accelerate the conversion of data to business value.
Enhance business metrics management
hengshi senseThe indicator system consists of atomic indicators and operational indicators. Atomic indicators are formulaically defined to form a unified calculation caliber for the same business within the team. On the basis of atomic indicators, business indicators add business information such as business scenario constraints, analysis dimensions, and timelines. Each business metric needs to be defined with corresponding business information. The concept of granularity is introduced to manage the relevant dimensions, timelines, and constraints set when defining business indicators in a unified manner. When creating business indicators, you can directly refer to the granularity of the corresponding business scenarios, reducing the setting of dimensions and other information, making the creation of business indicators more convenient and fast. Correspondingly, granularity can also be associated with metrics, and the metrics can be associated with the granularity on the granularity page, and the dimension list, timeline, and constraints in the associated metrics can be maintained synchronously and uniformly, and the definition details of the granularity can be automatically reflected in the relevant metrics after modification.
Synchronization enhances metric permission management to make metrics visible on a packet-by-packet basis. After enabled, the business metrics of the packet are visible to the outside world and can be explored and analyzed. The metric channel between the platform and the tenant is opened, and the tenant can use the platform indicator for exploration and analysis, expanding the usage scenarios of the metric.
Improve the ability to analyze business indicators
Business metrics are explored and analyzed in insights. Optimize insights and improve your metrics analysis capabilities. The dashboard adds a batch operation function to quickly load multiple business metrics in the dashboard at once. During batch operations, you can filter metrics by group, and automatically filter irrelevant metrics.
A common filter is added to the dashboard, which can filter and filter the indicators in the dashboard in a unified manner. The system has a built-in common time filter, and also supports user-defined dimension filters, so that the dashboard can display business indicator data information under different conditions. Users can customize the display of year-on-year information according to the current time dimension of the chart, and the timeline defined by the indicator will be automatically linked with the public time filter of the dashboard, so that you can quickly obtain the changes of core indicators in different time periods.
hengshi ai copilot
hengshi sense 5.0Introducing AI Copilot. AI Copilot quickly provides intuitive visualization of business scenarios and corresponding analysis insights through dialogue, helping users quickly interpret data, assist users in drawing required business conclusions, and improve the efficiency of overall business data decision-making. AI Copilot autonomously recommends relevant business scenario analysis questions for business data and guides users to ask questions, further lowering the threshold for data analysis and enabling junior analysts to explore and gain insights in real time. AI Copilot's business data analysis capabilities and intelligent data query experience can meet the needs of users at different levels.
Inhengshi senseAI Copilot implements the following capabilities:
Conversational interaction: Search and analyze business information in the form of natural language questions, and further generate visualizations related to business scenarios and business data summary and analysis reports.
Multi-round Q&A: AI Copilot supports multi-round Q&A, automatically analyzes the business situation based on the indicator data and business scenarios mentioned in the Q&A, and provides action optimization suggestions. Users can iterate over the conversation to get more accurate metrics to query targets.
Intelligent recommendation: Intelligently recommends business-related questions based on data content to guide users to ask questions more effectively, so that AI Copilot can accurately obtain user intent and give analysis results of related services.
Indicator platform + large model analysis method: The professional knowledge that is not understood by the large model is precipitated through the indicator platform to form a complete AI capability of general knowledge of the industry + proprietary knowledge of the enterprise.
Reports app.
hengshi sense 4.5Launched the complex report function, integrated Excel rich and powerful report editing functions, to achieve complex layout, complex calculation of Chinese reports, BI+ reports seamless cooperation, to meet the needs of most users. Launched the Reports app to further lower the threshold for using reports. Through the highly compatible Excel-compatible operation interface, the report application allows users to directly operate reports in a WYSIWYG way, which is more suitable for the usage habits of report users.
The report has a simple structure and consists of two parts: a conditional filter area and a data region. The Condition Filter area allows you to use filter controls and parameter controls to control the data display conditions. Data regions present business data. The report application supports importing Excel templates and quickly developing them on the templates. A report app can consist of one or more reports.
The report application adds functions such as *** recognition
The system supports QRCODE **PDF417 **GS1 128 barcode, EAN8 barcode, EAN13 barcode, Code93 barcode, CODABAR barcode.
Support ** identification, and the report content can be displayed in ** form.
Add a pagination switch, when the report content is large, you can turn on the switch to display the report in pagination.
Visual analytics.
Charts support multi-level linkage
When a dashboard analyzes a chart data, it will filter other chart data in conjunction to analyze the impact on business data under a certain condition. Previously, dashboards only supported first-level linkage filtering, that is, only specific conditions on one chart were allowed to be used to filter all charts in the dashboard at the same time. This is optimized to support multi-level linkage filtering of charts, and all charts in the dashboard can be filtered at the same time by using conditions in multiple charts to analyze business scenario data under multiple conditions. When you filter by linkage, you can view the linked charts on the chart that triggers the linkage, and you can view the linkage filter conditions on the linked chart or on the linkage filter panel. When you cancel a linkage, you can cancel it one by one on the linkage filter panel.
Filter enhancements
Filters are one of the commonly used functions of dashboards, which display data of different business scenarios through different filter conditions. The filter has been optimized in many aspects, and the settings such as operators and default value custom expressions have been added to expand the scope of data filtering and improve the filtering efficiency.
Filter operators can realize a multi-faceted combination of filtering options and adapt to more data filtering scenarios. Added five operators: Do not contain, Arbitrarily equal (or), Not equal (and), All match (and), and Arbitrary match (or), and the original include operator is used by default.
The default value of the filter is dynamically selected, and the custom filter function is added, and the user can write the default value through expression, making the default value setting more flexible.
The date range filter optimizes the date range selection method, supports different shortcuts at different time granularities, and can quickly select a date range through shortcuts.
Charts support expressions to set data
hengshi senseWhen dragging data into a chart for display, a variety of calculation methods are provided, including sum, average, maximum, minimum, counting, deduplication, duplication, retention, and active, but it still cannot meet the data processing needs of all usersAdded expression editing function to dimension and measure fields. When the built-in calculation method of the system cannot meet the user's requirements, you can use expressions and functions to customize the data processing method to make data calculation more flexible and convenient.
Optimize the subtotal
*The subtotal is the statistics of the metric information according to a certain dimension or several dimensions, and the calculation method of the subtotal can be set independently, and is not associated with the measurement calculation method. When there are multiple measures, some of the subtotals of the measures are not meaningful and do not want to participate in the calculation. If the subtotal of the categorical fields in the student's course statistics is not of practical significance, they do not want to participate in the statistical calculation. Optimized the subtotal, you can select the measurement field for subtotaling to participate in the subtotal for multiple measures, and leave the field content blank for the field content that does not need subtotal.
Optimized the display of mobile filters
When there are multiple global filters in the dashboard, the mobile terminal will display all these filters at the top of the interface, and you need to operate the slider to find the filters when using them. Due to the small slider on the mobile terminal, it is easy to misoperate, which affects the user's mobile experience. Optimized for this situation, when there are multiple global filters, only two global filters are shown at the top, and the others are collapsed. Click the expand button to set it when using, making it more convenient to operate the global filter on the mobile terminal.
Data management.
Tenant inheritance uses platform-side data
hengshi senseTenant scenarios for one platform and multiple enterprise users are supported. The platform can share data with tenants, and tenants can view and use data for analysis and exploration, but they cannot process the data, such as creating new fields and new indicators, which greatly hinders the tenant's data exploration. Breaking this data barrier, the data inheritance function allows tenants to perform in-depth data processing on existing data by adding new fields, adding new indicators, and expanding association models when using the data shared by the platform, so as to meet the flexible and changeable needs of tenants under self-service analysis.
Data model enhancements
In the data model, the dataset aggregated by the primary table cannot appear in the model, and there may be circular references, but in some cases, the aggregated dataset can be filtered and viewed in detail on the primary table. This model restriction is relaxed, users can drag the dataset aggregated by the main table into the model for data association, and a detection mechanism is added to the system to send a prompt message to the user when there is a circular reference.
Parameter optimization
hengshi senseParameters must be set at the time of creation, which makes the parameters must be used within the allowable value range, while in the business analysis scenario, the parameters need to have no default values at the beginning, and the user can select the parameter values according to the situation. Optimize parameters for the preceding business analysis scenarios. Parameters can be created without a default value, and parameter controls can be set without a default value when they are used.
Application management.
Added message push mode
hengshi senseIt supports sending the business data related information of the dashboard to relevant personnel by email, but for users who use office software such as WeCom and Feishu, it is more convenient to use software such as WeCom to send messages than email push, and users can read the dashboard information in a more timely mannerWeCom and Feishu were added to the dashboard information push method, and the subscription function was renamed. Similarly, the notification methods of WeCom and Feishu have also been added to the indicator early warning.
Wrapping up
The geese are in the sky, and the scales swim in the pool.
hengshi sense 5.0It has opened a new era of BI business intelligence, with an all-round upgrade of the capabilities of the indicator platform, the auxiliary analysis of the AI assistant, and the convenient use of the report application, each of which is worth your experience and trial
The Ming Dynasty changed the new law, and the plum willow waited for the spring.
The next generation of BI is herehengshi sense 5.1Follow the trend, continue to explore and innovate, will surely create more wonderful new features, welcome you to continue to pay attention!