This submit was cowritten by Mulay Ahmed, Assistant Director of Engineering, and Ruby Donald, Assistant Director of Engineering at Principal Monetary Group. The content material and opinions on this submit are these of the third-party creator and AWS shouldn’t be accountable for the content material or accuracy of this submit.
Principal Monetary Group® is an built-in international monetary providers firm with specialised options serving to individuals, companies, and establishments attain their long-term monetary objectives and entry better monetary safety.
With US contact facilities that deal with thousands and thousands of buyer calls yearly, Principal® needed to additional modernize their buyer name expertise. With a strong AWS Cloud infrastructure already in place, they chose a cloud-first method to create a extra personalised and seamless expertise for his or her clients that might:
- Perceive buyer intents by means of pure language (vs. contact tone experiences)
- Help clients with self-service choices the place doable
- Precisely route buyer calls based mostly on enterprise guidelines
- Help engagement middle brokers with contextual knowledge
Initially, Principal developed a voice Digital Assistant (VA) utilizing an Amazon Lex bot to acknowledge buyer intents. The VA can carry out self-service transactions or route clients to particular name middle queues within the Genesys Cloud contact middle platform, based mostly on buyer intents and enterprise guidelines.
As clients work together with the VA, it’s important to constantly monitor its well being and efficiency. This permits Principal to establish alternatives for fine-tuning, which may improve the VA’s capability to know buyer intents. Consequently, it will cut back fallback intent charges, enhance purposeful intent achievement charges, and result in higher buyer experiences.
On this submit, we discover how Principal used this chance to construct an built-in voice VA reporting and analytics resolution utilizing an Amazon QuickSight dashboard.
Amazon Lex is a service for constructing conversational interfaces utilizing voice and textual content. It gives high-quality speech recognition and language understanding capabilities, enabling the addition of refined, pure language chatbots to new and present purposes.
Genesys Cloud, an omni-channel orchestration and buyer relationship platform, gives a contact middle platform in a public cloud mannequin that permits fast and easy integration of AWS Contact Middle Intelligence (AWS CCI). As a part of AWS CCI, Genesys Cloud integrates with Amazon Lex, which allows self-service, clever routing, and knowledge assortment capabilities.
QuickSight is a unified enterprise intelligence (BI) service that makes it easy inside a corporation to construct visualizations, carry out advert hoc evaluation, and shortly get enterprise insights from their knowledge.
Answer overview
Principal required a reporting and analytics resolution that might monitor VA efficiency based mostly on buyer interactions at scale, enabling Principal to enhance the Amazon Lex bot efficiency.
Reporting necessities included buyer and VA interplay and Amazon Lex bot efficiency (goal metrics and intent achievement) analytics to establish and implement tuning and coaching alternatives.
The answer used a QuickSight dashboard that derives these insights from the next buyer interplay knowledge used to measure VA efficiency:
- Genesys Cloud knowledge reminiscent of queues and knowledge actions
- Enterprise-specific knowledge reminiscent of product and name middle operations knowledge
- Enterprise API-specific knowledge and metrics reminiscent of API response codes
The next diagram exhibits the answer structure utilizing Genesys, Amazon Lex, and QuickSight.
The answer workflow includes the next steps:
- Customers name in and work together with Genesys Cloud.
- Genesys Cloud calls an AWS Lambda routing perform. This perform will return a response to Genesys Cloud with the mandatory knowledge, to route the client name. To generate a response, the perform fetches routing knowledge from an Amazon DynamoDB desk, and requests an Amazon Lex V2 bot to offer a solution on the person intent.
- The Amazon Lex V2 bot processes the client intent and calls a Lambda achievement perform to meet the intent.
- The achievement perform executes customized logic (routing and session variables logic) and calls obligatory APIs to fetch the info required to meet the intent.
- The APIs course of and return the info requested (reminiscent of knowledge to carry out a self-service transaction).
- The Amazon Lex V2 bot’s dialog logs are despatched to Amazon CloudWatch (these logs shall be used for enterprise analytics, operational monitoring, and alerts).
- Genesys Cloud calls a 3rd Lambda perform to ship buyer interplay reviews. The Genesys report perform pushes these reviews to an Amazon Easy Storage Service (Amazon S3) bucket (these reviews shall be used for enterprise analytics).
- An Amazon Information Firehose supply stream ships the dialog logs from CloudWatch to an S3 bucket.
- The Firehose supply stream transforms the logs in Parquet or CSV format utilizing a Lambda perform.
- An AWS Glue crawler scans the info in Amazon S3.
- The crawler creates or updates the AWS Glue Information Catalog with the schema info.
- We use Amazon Athena to question the datasets (buyer interplay reviews and dialog logs).
- QuickSight connects to Athena to question the info from Amazon S3 utilizing the Information Catalog.
Different design concerns
The next are different key design concerns to implement the VA resolution:
- Price optimization – The answer makes use of Amazon S3 Bucket Keys to optimize on prices:
- Encryption – The answer encrypts knowledge at relaxation with AWS KMS and in transit utilizing SSL/TLS.
- Genesys Cloud integration – The mixing between the Amazon Lex V2 bot and Genesys Cloud is completed utilizing AWS Identification and Entry Administration (IAM). For extra particulars, see Genesys Cloud.
- Logging and monitoring – The answer displays AWS sources with CloudWatch and makes use of alerts to obtain notification upon failure occasions.
- Least privilege entry – The answer makes use of IAM roles and insurance policies to grant the minimal obligatory permissions to makes use of and providers.
- Information privateness – The answer handles buyer delicate knowledge reminiscent of personally identifiable info (PII) in line with compliance and knowledge safety necessities. It implements knowledge masking when relevant and applicable.
- Safe APIs – APIs applied on this resolution are protected and designed in line with compliance and safety necessities.
- Information sorts – The answer defines knowledge sorts, reminiscent of time stamps, within the Information Catalog (and Athena) in an effort to refresh knowledge (SPICE knowledge) in QuickSight on a schedule.
- DevOps – The answer is model managed, and adjustments are deployed utilizing pipelines, to allow quicker launch cycles.
- Analytics on Amazon Lex – Analytics on Amazon Lex empowers groups with data-driven insights to enhance the efficiency of their bots. The overview dashboard gives a single snapshot of key metrics reminiscent of the full variety of conversations and intent recognition charges. Principal doesn’t use this functionality as a result of following causes:
- The dashboard can’t combine with exterior knowledge:
- Genesys Cloud knowledge (reminiscent of queues and knowledge actions)
- Enterprise-specific knowledge (reminiscent of product and name middle operations knowledge)
- Enterprise API-specific knowledge and metrics (reminiscent of response codes)
- The dashboard can’t combine with exterior knowledge:
- The dashboard can’t be personalized so as to add extra views and knowledge.
Pattern dashboard
With this reporting and analytics resolution, Principal can consolidate knowledge from a number of sources and visualize the efficiency of the VA to establish areas of alternatives for enchancment. The next screenshot exhibits an instance of their QuickSight dashboard for illustrative functions.
Conclusion
On this submit, we offered how Principal created a report and analytics resolution for his or her VA resolution utilizing Genesys Cloud and Amazon Lex, together with QuickSight to offer buyer interplay insights.
The VA resolution allowed Principal to take care of its present contact middle resolution with Genesys Cloud and obtain higher buyer experiences. It affords different advantages reminiscent of the power for a buyer to obtain help on some inquiries with out requiring an agent on the decision (self-service). It additionally gives clever routing capabilities, resulting in lowered name time and elevated agent productiveness.
With the implementation of this resolution, Principal can monitor and derive insights from its VA resolution and fine-tune accordingly its efficiency.
In its 2025 roadmap, Principal will proceed to strengthen the inspiration of the answer described on this submit. In a second submit, Principal will current how they automate the deployment and testing of recent Amazon Lex bot variations.
AWS and Amazon aren’t associates of any firm of the Principal Monetary Group®. This communication is meant to be academic in nature and isn’t meant to be taken as a suggestion.
Insurance coverage merchandise issued by Principal Nationwide Life Insurance coverage Co (besides in NY) and Principal Life Insurance coverage Firm®. Plan administrative providers supplied by Principal Life. Principal Funds, Inc. is distributed by Principal Funds Distributor, Inc. Securities supplied by means of Principal Securities, Inc., member SIPC and/or unbiased dealer/sellers. Referenced firms are members of the Principal Monetary Group®, Des Moines, IA 50392. ©2025 Principal Monetary Providers, Inc. 4373397-042025
In regards to the Authors
Mulay Ahmed is an Assistant Director of Engineering at Principal and well-versed in architecting and implementing advanced enterprise-grade options on AWS Cloud.
Ruby Donald is an Assistant Director of Engineering at Principal and leads the Enterprise Digital Assistants Engineering Crew. She has intensive expertise in constructing and delivering software program at enterprise scale.
