Email marketing content generation method based on public cloud AI technology

Abstract

The present application relates to the field of artificial intelligence and digital marketing technology, in particular to an email marketing content generation method based on public cloud AI technology, comprising the following steps: S1: monitoring user behavior event flow through a public cloud event bus; S2: acquiring page interaction behavior data in real time and synchronously acquiring sudden environmental event data within a user geographic fence; S3: calculating behavior intensity value and environmental interference factor; S4: screening content atomic blocks matched in priority and injecting dynamic modifiers; S5: assembling the content atomic blocks injected with dynamic modifiers into an email body according to a preset template and performing sensitivity check through a regional compliance engine. The present application, by constructing a dynamic content generation mechanism based on behavior intensity and environmental interference factor and combining regional compliance check and multi-terminal output strategy, realizes comprehensive improvement of email marketing content in terms of personalization, situational adaptability and compliance.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence and digital marketing technologies, and in particular to a method for generating email marketing content based on public cloud AI technology. Background Art

[0002] Existing email marketing content generation generally relies on static content templates and a limited behavioral tag library, and is unable to dynamically adjust email content based on users' real-time behavior and the external environment they are in. As the market demand for personalized and context-aware marketing continues to increase, traditional solutions have obvious limitations in user interest modeling, environmental sensitivity response, and content compliance management, resulting in limited marketing reach, inaccurate content triggering, and difficulty in ensuring cross-regional content compliance. Especially in the context of the rapid development of cloud computing and AI technology, the real-time acquisition and comprehensive utilization of user behavior data and environmental event data has become an important foundation for driving the generation of a new generation of intelligent marketing content.

[0003] However, existing technologies often lack efficient event monitoring and multi-dimensional data integration capabilities based on public cloud platforms, and lack a complete closed-loop mechanism that can dynamically map user behavior intensity and environmental interference factors to the content generation process. Therefore, a method for generating email marketing content based on public cloud AI technology is urgently needed to address these issues. Summary of the Invention

[0004] Based on the above objectives, the present invention provides a method for generating email marketing content based on public cloud AI technology.

[0005] The email marketing content generation method based on public cloud AI technology includes the following steps:

[0006] S1: Monitor the user behavior event stream through the public cloud event bus and extract the user identifier and event type from the user behavior event stream;

[0007] S2: Using the user identifier as an index, it obtains page interaction behavior data in real time and simultaneously calls the external environment API to obtain emergency environmental event data within the user's geofence;

[0008] S3: Calculate the behavior intensity value based on the page interaction behavior data, and calculate the environmental interference factor based on the sudden environmental event data;

[0009] S4: Use the behavior strength value to filter priority-matched content atomic blocks in the preset content library, and inject dynamic modifiers into the content atomic blocks based on environmental interference factors;

[0010] S5: Assemble the content blocks injected with dynamic modifiers into the email body according to the preset template and perform sensitivity verification through the regional compliance engine;

[0011] S6: The email body that passes the compliance check is output as the final dynamic email content body.

[0012] Optionally, the S1 specifically includes:

[0013] S11: Configure user behavior event monitoring rules in the public cloud event bus, set the event source path, event format identifier, and trigger conditions to continuously receive user behavior event streams from the target business system;

[0014] S12: When the event triggering condition is met, the event bus receives the user behavior event and encapsulates the event into a structured data format. The structured data is preset with multiple fields, including user identifier, event type, event occurrence time, page identifier, terminal type, and session number;

[0015] S13: Based on the preset field names in the structured data format, extract the user identifier field and the event type field.

[0016] Optionally, the S2 specifically includes:

[0017] S21: Based on the user identifier extracted in S1, the user behavior database is searched for the session record bound to the user identifier, and the page interaction behavior data within the current session period is extracted in real time, including the page dwell time, mouse click location sequence, and scrolling behavior trajectory;

[0018] S22: Based on the most recent location record corresponding to the user identifier extracted in S1, the user's geo-fence is constructed, and an HTTP request is initiated to the API exposed by the external environment data provider using the geo-fence parameters as input. The request content includes the API key, timestamp, and fence boundary coordinates.

[0019] S23: Receive API response data, parse and obtain emergency environmental event data within the current fence range, including meteorological anomaly categories, local real-time news headlines and event publishing agency identifiers.

[0020] Optionally, the geo-fence range of the user is constructed including:

[0021] S221: Obtain the most recent location record corresponding to the user identifier, including latitude, longitude, and record timestamp;

[0022] S222: Setting a fence radius value for spatial perception and constructing a circular fence area with the latitude and longitude as the geographic center. The fence boundary is determined by calculating the spherical distance between the center point and multiple candidate boundary points, and filtering out the coordinates of all boundary points whose distance does not exceed the set fence radius.

[0023] S223: Combine all boundary point coordinates that meet the spatial distance requirement into a fence boundary coordinate set, and encapsulate the fence boundary coordinate set and the corresponding timestamp into fence parameters.

[0024] Optionally, calculating the behavior intensity value based on the page interaction behavior data includes:

[0025] S31: After receiving the page dwell time, the mouse click position sequence, and the scrolling behavior trajectory, the three behavior data are respectively normalized to map each behavior indicator to a unified dimension interval of 0 to 1;

[0026] S32: Setting fixed weight coefficients for page dwell time, number of mouse clicks, and scroll displacement length, and performing a weighted summation of the three normalized behavior indicators based on the weight coefficients to obtain a comprehensive behavior score for the user's current session;

[0027] S33: Matching the behavior comprehensive score with a pre-set multi-level threshold interval, and outputting the value of the corresponding interval as the behavior intensity value. A higher numerical interval of the behavior intensity value represents a higher user interest level.

[0028] Optionally, the calculating of the environmental interference factor based on the environmental emergency event data includes:

[0029] S34: After receiving the environmental emergency event data, classify and process it according to the event source, wherein the source includes a meteorological data source and a local news data source;

[0030] S35: Convert each type of emergency event into a corresponding initial interference value according to a preset impact level mapping rule;

[0031] S36: Weighted integration of the interference values ​​of all emergencies occurring within the same time window to form a cumulative environmental interference score, where the weight coefficient of each event type is set based on its average historical impact;

[0032] S37: Match the accumulated environmental interference score with a preset multi-level threshold interval, and output a corresponding level value as the environmental interference factor.

[0033] Optionally, the S4 specifically includes:

[0034] S41: Matching the behavior strength value with the priority interval corresponding to each content atomic block in the content library, and filtering out a set of content atomic blocks whose behavior strength values ​​fall within the interval range;

[0035] S42: For each filtered content atomic block, select a corresponding modification strategy parameter group according to the current environmental interference factor level;

[0036] S43: injecting the modification strategy parameter group into the editing placeholder position of the target content atomic block, and forming a dynamic modifier combination with environmental adaptability through text replacement to generate a processed content atomic block.

[0037] Optionally, the S42 specifically includes:

[0038] S421: Using the level value of the current environmental interference factor as an index, a parameter group of the corresponding level is retrieved from a preset modification strategy parameter table, specifically including the original tone adjustment factor, the urgency label template, and the emotion guiding phrase index;

[0039] S422: Modify the strategy parameter table to pre-define a total of 5 levels of strategy groups, each level corresponding to a level of environmental interference factors;

[0040] S423: Adjust the tone adjustment factor according to the behavior intensity value to generate a final tone coefficient. The formula is: ,in, is the final tone coefficient; is the original tone adjustment factor; is the behavior intensity value, ranging from 1 to 5; is the tone adjustment weight coefficient, which is a constant.

[0041] Optionally, the S5 specifically includes:

[0042] S51: Based on the email template structure bound to the user's business scenario, determine the insertion position information of the email body, and fill the content atomic blocks injected with dynamic modifiers into the corresponding placeholder positions in the order of the beginning of the body, the middle narrative paragraph, and the closing call to action paragraph, to generate a structured email body draft;

[0043] S52: Load the corresponding compliance rule set based on the regulatory standards applicable to the country or region where the user is located, and perform a sensitivity review on the draft email body.

[0044] Optionally, the S6 specifically includes:

[0045] S61: After confirming that the email body has passed the regional compliance check in S5, the corresponding output format template is selected based on the user terminal type, including HTML format, rich text format, and plain text format, which are respectively suitable for web pages, clients, and lightweight terminal environments;

[0046] S62: Filling the structured email body content into the fixed layout area of ​​the selected format template, and completing tag encapsulation, character escape, and semantic structure closure according to the target output format specification to form a renderable email content file;

[0047] S63: performing content summary generation and digital signature processing on the email content file to generate a final dynamic email content body.

[0048] Beneficial effects of the present invention:

[0049] The present invention realizes the real-time capture of user behavior events by configuring an event monitoring mechanism in a public cloud environment, and combines user page interaction data with sudden environmental event data within a geographic fence to construct behavior intensity values ​​and environmental interference factors, effectively improving the timeliness and personalization of email content generation; by mapping behavior intensity values ​​to content priority intervals, accurately screening content atomic blocks, and injecting dynamic modifiers according to environmental interference factors, the content has context-awareness capabilities, thereby enhancing the ability to respond to user interests and environmental changes.

[0050] The present invention uses a templated structure to dynamically assemble content atomic blocks, and combines it with a regional compliance engine to perform multi-regional regulatory verification to ensure that the content of the email body meets the dissemination standards of the target area; at the same time, it supports terminal output in multiple formats and content integrity signatures, improving cross-platform adaptability and transmission security. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0052] Figure 1 This is a schematic diagram of a method for generating email marketing content according to an embodiment of the present invention;

[0053] Figure 2 Schematic diagram of the process of screening content atomic blocks and injecting dynamic modifiers in an embodiment of the present invention. DETAILED DESCRIPTION

[0054] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments. It is also noted that, to provide a more detailed description, the following embodiments are best and preferred embodiments, and those skilled in the art may employ alternative methods for implementing certain known technologies. Furthermore, the accompanying drawings are intended only to provide a more detailed description of the embodiments and are not intended to limit the present invention.

[0055] like Figure 1-Figure 2 As shown in FIG, the method for generating email marketing content based on public cloud AI technology includes the following steps:

[0056] S1: Monitor the user behavior event stream through the public cloud event bus and extract the user identifier and event type from the user behavior event stream;

[0057] S1 specifically includes:

[0058] S11: Configure user behavior event monitoring rules in the public cloud event bus, set the event source path, event format identifier, and trigger conditions to continuously receive user behavior event streams from the target business system;

[0059] S12: When the event triggering condition is met, the event bus receives the user behavior event and encapsulates the event into a structured data format. The structured data is preset with multiple fields, including user identifier, event type, event occurrence time, page identifier, terminal type, and session number;

[0060] S13: Based on the preset field names in the structured data format, extract the user identifier field and the event type field, where the user identifier is used to uniquely identify the user identity, and the event type is used to identify the category of the behavioral event that occurred; the above steps can achieve stable acquisition and element extraction of user behavioral events by setting monitoring rules in the public cloud event bus and parsing structured event data, providing a reliable starting basis and input entry for subsequent behavioral data collection.

[0061] S2: Using the user identifier as an index, it obtains page interaction behavior data in real time and simultaneously calls the external environment API to obtain emergency environmental event data within the user's geofence;

[0062] S2 specifically includes:

[0063] S21: Based on the user identifier extracted in S1, the user behavior database is searched for the session record bound to the user identifier, and the page interaction behavior data within the current session period is extracted in real time, including the page dwell time, mouse click location sequence, and scrolling behavior trajectory;

[0064] S22: Based on the most recent location record corresponding to the user identifier extracted in S1, the user's geo-fence is constructed, and an HTTP request is initiated to the API exposed by the external environment data provider using the geo-fence parameters as input. The request content includes the API key, timestamp, and fence boundary coordinates.

[0065] S23: Receive API response data, parse and obtain emergency environmental event data within the current fence range, including meteorological anomaly categories, local real-time news headlines and event release agency identifiers; the above steps obtain highly timely page interaction behavior data by using user identifiers as retrieval indexes, and obtain external environmental information in combination with the API call mechanism built based on fence parameters. This can simultaneously build a dynamic data foundation of user behavior and environmental interference, providing a two-dimensional input source for subsequent feature calculations.

[0066] Constructing the user's geo-fence range includes:

[0067] S221: Obtain the most recent location record corresponding to the user identifier, including latitude, longitude, and record timestamp;

[0068] S222: Setting a fence radius value for spatial perception and constructing a circular fence area with the latitude and longitude as the geographic center. The fence boundary is determined by calculating the spherical distance between the center point and multiple candidate boundary points, and filtering out the coordinates of all boundary points whose distance does not exceed the set fence radius.

[0069] By preset fence radius , with position coordinates As the center, a circular geo-fence is constructed, and the fence boundary points meet the following conditions:

[0070] ,in, Represents the spherical distance between two points on the Earth's surface, calculated using the Haversine formula, specifically:

[0071] , where the parameters are defined as follows, is the spherical distance between the user position and the candidate boundary point; is the average radius of the Earth, set to rice; The latitude of the user's location; Longitude of the user's location; is the latitude of the candidate fence boundary point; is the longitude of the candidate fence boundary point; is the preset geo-fence radius;

[0072] S223: Combine all boundary point coordinates that meet the spatial distance requirements into a fence boundary coordinate set, and encapsulate the fence boundary coordinate set and the corresponding timestamp as fence parameters as spatial retrieval input for subsequent environmental event data API calls.

[0073] S3: Calculate the behavior intensity value based on the page interaction behavior data, and calculate the environmental interference factor based on the sudden environmental event data;

[0074] The behavior intensity value calculated based on page interaction behavior data includes:

[0075] S31: After receiving the page dwell time, mouse click position sequence, and scrolling behavior trajectory, each of the three behavior data items is subjected to interval normalization processing, and each behavior indicator is mapped to a unified dimension interval from 0 to 1. The normalization processing is performed according to the following formula: ,in, is the original behavior data value, 、 The preset minimum and maximum values ​​for the corresponding behavioral indicators; is the normalized behavioral index value;

[0076] S32: Set fixed weight coefficients for page dwell time, number of mouse clicks, and scroll displacement length, and perform weighted summation of the three normalized behavior indicators based on the weight coefficients to obtain a comprehensive behavior score for the user's current session. The weighted calculation formula is: ,in, A comprehensive score for behavior; are the weight coefficients of the corresponding indicators, satisfying ; are the three behavioral indicators after normalization;

[0077] S33: Matching the behavior comprehensive score with a pre-set multi-level threshold interval, and outputting the value of the corresponding interval as the behavior intensity value. A higher numerical interval of the behavior intensity value represents a higher user interest level.

[0078] Table 1 Multi-level threshold interval mapping table

[0079]

[0080] By mapping the comprehensive behavioral score into discrete behavioral intensity level values, the floating range of behavioral characteristics can be effectively compressed, and the matching stability between behavioral characteristics and content priorities can be enhanced, thereby improving the accuracy and efficiency of personalized content generation.

[0081] Environmental interference factors calculated based on sudden environmental event data include:

[0082] S34: After receiving the environmental emergency event data, classify and process it according to the event source, which includes a meteorological data source and a local news data source;

[0083] S35: Based on the preset impact level mapping rules, each type of emergency event is converted into a corresponding initial interference value. Meteorological events include heavy rainfall, storms, and extreme high temperatures, while local news events include accidents, road closures, and emergency management. Each type of event is associated with a unique interference level.

[0084] S36: Weighted integration of the interference values ​​of all emergencies occurring within the same time window is performed to form a cumulative environmental interference score, where the weight coefficient of each event type is set based on its average historical impact. The cumulative environmental interference score is calculated as follows: ,in, score points for environmental disturbances; is the total number of emergencies in the current time window; For the The weight coefficient of each event type satisfies ; For the The initial interference value of each event is an integer level value;

[0085] S37: Match the accumulated environmental interference score with the preset multi-level threshold interval, and output the corresponding level value as the environmental interference factor. The level value is a discrete integer. The higher the value, the stronger the interference of the external environment on the user's attention.

[0086] Based on the cumulative environmental interference score The value of falls into the interval, and the table output corresponds to the environmental interference factor , where the level intervals and output values ​​are mapped according to Table 2 below:

[0087]

[0088] By converting heterogeneous environmental events into standard interference levels and implementing weighted integration and interval mapping, it is possible to achieve quantitative control of the intensity of environmental impacts, effectively supporting the dynamic adaptability and targeting of subsequent content modification strategies.

[0089] S4: Use the behavior strength value to filter priority-matched content atomic blocks in the preset content library, and inject dynamic modifiers into the content atomic blocks based on environmental interference factors;

[0090] S4 specifically includes:

[0091] S41: Matching the behavior strength value with the priority interval corresponding to each content atomic block in the content library, and screening out a set of content atomic blocks whose behavior strength values ​​fall within the interval range. The priority interval is a preset behavior response level. A higher value indicates a stronger information density and marketing guidance degree of the content atomic block. The priority interval corresponds to the behavior response level one by one: behavior strength value 1 corresponds to priority interval P1, behavior strength value 2 corresponds to priority interval P2, behavior strength value 3 corresponds to priority interval P3, behavior strength value 4 corresponds to priority interval P4, and behavior strength value 5 corresponds to priority interval P5. Increasing priority interval values ​​indicate increasing information density and marketing guidance strength of the content atomic block. Screen out a set of content atomic blocks whose behavior strength values ​​fall within the corresponding priority interval.

[0092] S42: For each filtered content atomic block, select a corresponding modification strategy parameter group according to the current environmental interference factor level;

[0093] S43: Inject the modification strategy parameter group into the editing placeholder position of the target content atomic block, and form a dynamic modifier combination with environmental adaptability through text replacement to generate a processed content atomic block; the above steps drive content granularity screening through behavioral intensity values, and inject semantic modification strategies in combination with environmental interference factors, which can ensure personalization accuracy while improving the content's dissemination adaptability and response efficiency in complex external situations.

[0094] S42 specifically includes:

[0095] S421: Using the level value of the current environmental interference factor as an index, a parameter group of the corresponding level is retrieved from a preset modification strategy parameter table, specifically including the original tone adjustment factor, the urgency label template, and the emotion guiding phrase index;

[0096] S422: The modification strategy parameter table predefines a total of five levels of strategy groups, each level corresponding to a level of environmental interference factor, wherein a higher level value indicates a stronger environmental interference, a higher corresponding tone adjustment factor value, and a more stimulating urgency label template and emotion-guiding phrase;

[0097] Table 3 Modification strategy parameters

[0098]

[0099] In Table 3 above, the environmental interference factor level is obtained by the environmental interference factor calculation process in step S3, and the value range is 1–5; the original tone adjustment factor Used to control the intensity of the emotional tone in the modifier, with higher values ​​indicating stronger tone. The Urgency Label Template Number specifies the style number of the label template (e.g., "Limited Time," "Act Now," etc.) inserted into the content. The Emotional Introducing Phrase Index Number selects a phrase template that matches the context (e.g., emphasizing anxiety, tension, anticipation, etc.).

[0100] S423: Adjust the tone adjustment factor according to the behavior intensity value to generate a final tone coefficient. The formula is: ,in, is the final tone coefficient, which is used to control the tone strength of the text; is the original tone adjustment factor; is the behavior intensity value, ranging from 1 to 5; is the tone adjustment weight coefficient, which is a constant and is used to balance the amplitude of behavior-driven tone influence. The above steps establish a direct mapping relationship between the level of environmental interference factors and the multi-level modification strategy parameter group, and introduce a joint regulation mechanism of behavior intensity on tone parameters. This can dynamically control the text tone and label emotional style, achieve dual adaptive modification of the external environment and user behavior, and thus improve the contextual responsiveness and conversion efficiency of marketing content to target users.

[0101] S5: Assemble the content blocks injected with dynamic modifiers into the email body according to the preset template and perform sensitivity verification through the regional compliance engine;

[0102] S5 specifically includes:

[0103] S51: Based on the email template structure bound to the user's business scenario, determine the insertion position information of the email body, and fill the content atomic blocks injected with dynamic modifiers into the corresponding placeholder positions in the order of the beginning of the body, the middle narrative paragraph, and the closing call to action paragraph, to generate a structured email body draft;

[0104] S52: Load the corresponding compliance rule set based on the regulatory standards applicable to the user's country or region, and perform a sensitivity review on the draft email body;

[0105] Specific reviews include:

[0106] S521: Compare keywords, tags, modifiers, and other fields in the draft text one by one to detect whether they contain sensitive content marked as restricted use or prohibited dissemination in the loaded rule set;

[0107] S522: Replace, block, or make modification suggestions for the parts determined to contain sensitive content until the verification results meet the marketing communication compliance standards of the current region.

[0108] S6: Output the email body that has passed the compliance check as the final dynamic email content;

[0109] S6 specifically includes:

[0110] S61: After confirming that the email body has passed the regional compliance check in S5, the corresponding output format template is selected based on the user terminal type, including HTML format, rich text format, and plain text format, which are respectively suitable for web pages, clients, and lightweight terminal environments;

[0111] S62: Filling the structured email body content into the fixed layout area of ​​the selected format template, and completing tag encapsulation, character escape, and semantic structure closure according to the target output format specification to form a renderable email content file;

[0112] S63: Perform content summary generation and digital signature processing on the email content file to generate the final dynamic email content body. The summary is used for subsequent content version identification, and the signature is used for integrity verification and anti-tampering control of the email content during transmission. The above steps, through the combination of terminal adaptation, format encapsulation and digital signature mechanism, convert the compliant email body into an email content body that meets the publishing conditions, which can ensure the compatibility, security and traceability of dynamic content in different application scenarios, and improve the reliability and professionalism of the overall marketing delivery.

[0113] The present invention encompasses any alternatives, modifications, equivalents, and solutions that fall within the spirit and scope of the present invention. To provide a thorough understanding of the present invention, specific details are described in detail below in connection with the preferred embodiments of the present invention, but those skilled in the art will be able to fully understand the present invention without these detailed descriptions. Furthermore, to avoid unnecessary confusion regarding the essence of the present invention, well-known methods, processes, procedures, components, and circuits have not been described in detail.

[0114] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for generating email marketing content based on public cloud AI technology, characterized in that: The following steps are involved: S1: Monitor the user behavior event stream through the public cloud event bus and extract the user identifier and event type from the user behavior event stream; S2: Using the user identifier as an index, it obtains page interaction behavior data in real time and simultaneously calls the external environment API to obtain emergency environmental event data within the user's geofence; The S2 specifically includes: S21: Based on the user identifier extracted in S1, the user behavior database is searched for the session record bound to the user identifier, and the page interaction behavior data within the current session period is extracted in real time, including the page dwell time, mouse click location sequence, and scrolling behavior trajectory; S22: Based on the most recent location record corresponding to the user identifier extracted in S1, the user's geo-fence is constructed, and an HTTP request is initiated to the API exposed by the external environment data provider using the geo-fence parameters as input. The request content includes the API key, timestamp, and fence boundary coordinates. S23: Receive API response data, parse and obtain emergency environmental event data within the current fence range, including meteorological anomaly categories, local real-time news headlines, and event publishing agency identifiers; S3: Calculate the behavior intensity value based on the page interaction behavior data, and calculate the environmental interference factor based on the sudden environmental event data; Calculating the behavior intensity value based on the page interaction behavior data includes: S31: After receiving the page dwell time, the mouse click position sequence, and the scrolling behavior trajectory, the three behavior data are respectively normalized to map each behavior indicator to a unified dimension interval of 0 to 1; S32: Setting fixed weight coefficients for page dwell time, number of mouse clicks, and scroll displacement length, and performing a weighted summation of the three normalized behavior indicators based on the weight coefficients to obtain a comprehensive behavior score for the user's current session; S33: matching the behavior comprehensive score with a pre-set multi-level threshold interval, and outputting the value of the corresponding interval as the behavior intensity value. A higher numerical interval of the behavior intensity value represents a higher user interest level. S4: Use the behavior strength value to filter priority-matched content atomic blocks in the preset content library, and inject dynamic modifiers into the content atomic blocks based on environmental interference factors; S5: Assemble the content blocks injected with dynamic modifiers into the email body according to the preset template and perform sensitivity verification through the regional compliance engine; S6: The email body that passes the compliance check is output as the final dynamic email content body.

2. The method for generating email marketing content based on public cloud AI technology according to claim 1, characterized in that: Said S1 specifically includes: S11: Configure user behavior event monitoring rules in the public cloud event bus, set the event source path, event format identifier, and trigger conditions to continuously receive user behavior event streams from the target business system; S12: When the event triggering condition is met, the event bus receives the user behavior event and encapsulates the event into a structured data format. The structured data is preset with multiple fields, including user identifier, event type, event occurrence time, page identifier, terminal type, and session number; S13: Based on the preset field names in the structured data format, extract the user identifier field and the event type field.

3. The method for generating email marketing content based on public cloud AI technology according to claim 1, characterized in that: The geo-fence range where the user is located includes: S221: Obtain the most recent location record corresponding to the user identifier, including latitude, longitude, and record timestamp; S222: Setting a fence radius value for spatial perception and constructing a circular fence area with the latitude and longitude as the geographic center. The fence boundary is determined by calculating the spherical distance between the center point and multiple candidate boundary points, and filtering out the coordinates of all boundary points whose distance does not exceed the set fence radius. S223: Combine all boundary point coordinates that meet the spatial distance requirement into a fence boundary coordinate set, and encapsulate the fence boundary coordinate set and the corresponding timestamp into fence parameters.

4. The method for generating email marketing content based on public cloud AI technology according to claim 1, characterized in that: Calculating the environmental interference factor based on the environmental emergency event data includes: S34: After receiving the environmental emergency event data, classify and process it according to the event source, wherein the source includes a meteorological data source and a local news data source; S35: Convert each type of emergency event into a corresponding initial interference value according to a preset impact level mapping rule; S36: Weighted integration of the interference values ​​of all emergencies occurring within the same time window to form a cumulative environmental interference score, where the weight coefficient of each event type is set based on its average historical impact; S37: Match the accumulated environmental interference score with a preset multi-level threshold interval, and output a corresponding level value as the environmental interference factor.

5. The method for generating email marketing content based on public cloud AI technology according to claim 1, characterized in that: The S4 specifically includes: S41: Matching the behavior strength value with the priority interval corresponding to each content atomic block in the content library, and filtering out a set of content atomic blocks whose behavior strength values ​​fall within the interval range; S42: For each filtered content atomic block, select a corresponding modification strategy parameter group according to the current environmental interference factor level; S43: injecting the modification strategy parameter group into the editing placeholder position of the target content atomic block, and forming a dynamic modifier combination with environmental adaptability through text replacement to generate a processed content atomic block.

6. The method for generating email marketing content based on public cloud AI technology according to claim 5, characterized in that: The S42 specifically includes: S421: Using the level value of the current environmental interference factor as an index, a parameter group of the corresponding level is retrieved from a preset modification strategy parameter table, specifically including the original tone adjustment factor, the urgency label template, and the emotion guiding phrase index; S422: Modify the strategy parameter table to pre-define a total of 5 levels of strategy groups, each level corresponding to a level of environmental interference factors; S423: Adjust the tone adjustment factor according to the behavior intensity value to generate a final tone coefficient. The formula is: ,in, is the final tone coefficient; is the original tone adjustment factor; is the behavior intensity value, ranging from 1 to 5; is the tone adjustment weight coefficient, which is a constant.

7. The method for generating email marketing content based on public cloud AI technology according to claim 1, characterized in that: The S5 specifically includes: S51: Based on the email template structure bound to the user's business scenario, determine the insertion position information of the email body, and fill the content atomic blocks injected with dynamic modifiers into the corresponding placeholder positions in the order of the beginning of the body, the middle narrative paragraph, and the closing call to action paragraph, to generate a structured email body draft; S52: Load the corresponding compliance rule set based on the regulatory standards applicable to the country or region where the user is located, and perform a sensitivity review on the draft email body.

8. The method for generating email marketing content based on public cloud AI technology according to claim 1, characterized in that: The S6 specifically includes: S61: After confirming that the email body has passed the regional compliance check in S5, the corresponding output format template is selected based on the user terminal type, including HTML format, rich text format, and plain text format, which are respectively suitable for web pages, clients, and lightweight terminal environments; S62: Filling the structured email body content into the fixed layout area of ​​the selected format template, and completing tag encapsulation, character escape, and semantic structure closure according to the target output format specification to form a renderable email content file; S63: performing content summary generation and digital signature processing on the email content file to generate a final dynamic email content body.

Images