Distributed Rule-Based Sequencing Engine for Customer Interaction Analytics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing customer engagement analytics platforms face limitations in sequencing customer interactions across multiple channels and scaling with large data sets, making it difficult to track key performance indicators (KPIs) and accommodate new interaction types or issues.
Innovation Solution
A distributed computing system processes interaction data across multiple nodes using a map-reduce model, evaluating defined rules to generate sequences based on time frames, communication channels, and contact reasons, allowing for easy configuration and processing of large data sets without code changes, and enabling cross-channel analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a distributed computing system with map-reduce model is used to process interaction data, then the processing efficiency and scalability for large data sets is improved, but the system complexity increases
Solution Approach 1:
The system divides the interaction data processing into separate mapping and reducing phases across multiple computing nodes. Each node independently processes portions of the data in the mapping phase, then results are aggregated and processed in the reducing phase, enabling parallel processing of large data sets while maintaining manageable node-level complexity
Solution Approach 2:
The computing nodes are designed to perform multiple functions: data reception, mapping operations, reducing operations, and sequence generation. This multi-functionality allows the same infrastructure to handle various interaction types and sequencing requirements without adding specialized hardware for each function
2Device complexity
If sequencing rules are limited to specific interaction channels, then the rule complexity is reduced, but the adaptability to new interaction types and channels is worsened
Solution Approach 1:
The sequencing engine is designed with a universal rule evaluation framework that can handle multiple interaction types (email, phone, SMS, web, IVR) through a common processing architecture. New interaction types can be accommodated by adding appropriate handling logic without redesigning the entire sequencing system
Solution Approach 2:
The system allows dynamic addition and modification of sequencing rules for new interaction channels and types. The rule set is not fixed but can be updated to accommodate emerging communication formats and interaction patterns as businesses expand their customer engagement channels
3Device complexity
If traditional sequencing methods are used without distributed computing, then the system simplicity is maintained, but the ability to process large amounts of interaction data is limited
Solution Approach 1:
The system transitions from single-node sequential processing to multi-node parallel processing by adding the spatial dimension of distributed computing. Interaction data is distributed across multiple computing nodes that process simultaneously, dramatically increasing throughput while maintaining the logical simplicity of the sequencing rules through the map-reduce abstraction
4Device complexity
If sequencing is performed without a distributed architecture, then the implementation simplicity is maintained, but the scalability to accommodate growing data volumes is reduced
Solution Approach 1:
The distributed architecture segments the processing workload across multiple independent nodes, allowing the system to scale horizontally by simply adding more nodes. Each node maintains a simplified implementation while the collective system achieves the scalability needed for growing data volumes
Solution Approach 2:
The system is designed to handle more data volume and interaction types than any single node could process alone. By distributing the workload, the system can accommodate excessive growth in interaction data without requiring complex coordination or redesign of individual node functionality
Data Source
AI summary
A computing device including a processor and memory may receive input from a user to define one or more sequence types, each sequence type based on a time frame. A distributed computing system may receive interaction data on a plurality of computing nodes and describing an interaction event between a customer and an agent. The one or more computing nodes may associate the received interaction event with the customer's identifier in a customer history table, wherein the customer history table stores historical interaction events for the customer. Further, the computing nodes may assign the received interaction event to a sequence, based on the defined one or more sequence types.