A longtime monetary companies agency with over 140 years in enterprise, Principal is a world funding administration chief and serves greater than 62 million clients around the globe. Principal is conducting enterprise-scale near-real-time analytics to ship a seamless and hyper-personalized omnichannel buyer expertise on their mission to make monetary safety accessible for all. They’re processing knowledge throughout channels, together with recorded contact middle interactions, emails, chat and different digital channels.
On this put up, we display how knowledge aggregated inside the AWS CCI Post Call Analytics solution allowed Principal to realize visibility into their contact middle interactions, higher perceive the shopper journey, and enhance the general expertise between contact channels whereas additionally sustaining knowledge integrity and safety.
Answer necessities
Principal gives funding companies by Genesys Cloud CX, a cloud-based contact middle that gives powerful, native integrations with AWS. Annually, Principal handles tens of millions of calls and digital interactions. As a primary step, they wished to transcribe voice calls and analyze these interactions to find out main name drivers, together with points, matters, sentiment, common deal with time (AHT) breakdowns, and develop extra pure language processing (NLP)-based analytics.
So as analyze the calls correctly, Principal had a couple of necessities:
- Contact particulars: Understanding the shopper journey requires understanding whether or not a speaker is an automatic interactive voice response (IVR) system or a human agent and when a name switch happens between the 2.
- Content material redaction: Every buyer audio interplay is recorded as a stereo WAV file, however might probably embrace delicate info equivalent to HIPAA-protected and personally identifiable info (PII).
- Scalability: This structure wanted to instantly scale to 1000’s of calls per day and tens of millions of calls per yr. As well as, Principal wanted an extensible analytics structure that analyze different channels equivalent to e mail threads and conventional voice of the shopper (VoC) survey outcomes.
- Integrity is non-negotiable at Principal—it guides the whole lot they do. In reality, doing what’s proper is likely one of the core values at Principal. Due to this fact, when the Principal staff began tackling this venture, they knew that guaranteeing the very best normal of knowledge safety equivalent to regulatory compliance, knowledge privateness, and knowledge high quality could be a non-negotiable, key requirement. The staff wanted to make the most of know-how with an identical stance on knowledge safety, and the power to construct customized compliance and safety controls to uphold strict necessities. Consideration to this key requirement permits Principal to take care of a protected and safe buyer expertise.
Answer overview
After in depth analysis, the Principal staff finalized AWS Contact Center Intelligence (CCI) solutions, which empower corporations to enhance buyer expertise and achieve dialog insights by including AI capabilities to third-party on-premises and cloud contact facilities. The CCI Post-Call Analytics (PCA) answer is a part of CCI options suite and match lots of the recognized necessities. PCA has a Solutions Library Guidance reference architecture with an open-source example repository on GitHub. Working with their AWS account staff, Principal detailed the PCA answer and its deployment, and arrange customized coaching packages and immersion days to quickly upskill the Principal groups. The instance structure (see the next diagram) and code base within the open-source repository allowed the Principal engineering groups to jumpstart their answer round unifying the shopper journey, and merging telephony information and transcript information collectively.
PCA gives a complete structure round ingesting audio information in a completely automated workflow with AWS Step Functions, which is initiated when an audio file is delivered to a configured Amazon Simple Storage Service (Amazon S3) bucket. After a couple of minutes, a transcript is produced with Amazon Transcribe Call Analytics and saved to a different S3 bucket for processing by different enterprise intelligence (BI) instruments. PCA additionally affords a web-based consumer interface that enables clients to browse name transcripts. PCA’s safety features make sure that any PII knowledge was redacted from the transcript, in addition to from the audio file itself. Moreover, all knowledge inside the S3 bucket may be encrypted with keys belonging to Principal.
Principal labored with AWS technical groups to change the Step Capabilities workflow inside PCA to additional obtain their targets. Name particulars equivalent to interplay timestamps, name queues, agent transfers, and participant talking instances are tracked by Genesys in a file known as a Contact Hint Document (CTR). Combining correct transcripts with Genesys CTR information, Principal might correctly establish the audio system, categorize the calls into teams, analyze agent efficiency, establish upsell alternatives, and conduct extra machine studying (ML)-powered analytics.
The groups constructed a brand new knowledge ingestion mechanism, permitting the CTR information to be collectively delivered with the audio file to an S3 bucket. Principal and AWS collaborated on a brand new AWS Lambda operate that was added to the Step Capabilities workflow. This Lambda operate identifies CTR information and gives a further processing step that outputs an enhanced transcript containing extra metadata equivalent to queue and agent ID info, IVR identification and tagging, and what number of brokers (and IVRs) the shopper was transferred to, all aggregated from the CTR information. This further info permits Principal to create a map of the shopper interplay all through the lifecycle of the dialog and deal with the vital speech segments, whereas excluding much less related ones.
Moreover, this postprocessing step enabled Principal to additional enrich transcripts with inside info equivalent to agent and queue names and increase the analytics capabilities of PCA, together with customized NLP-based ML fashions for subject and buyer intent identification, deployed utilizing Amazon SageMaker endpoints, and extra transcript augmentation utilizing foundational generative AI fashions hosted on Amazon Bedrock.
PCA is open supply on GitHub, which permits clients equivalent to Principal to increase and keep their very own forks with personalized, non-public enterprise code. It additionally permits the group to submit code again to the primary repository for others to make use of. Principal and AWS technical groups partnered to merge the Genesys CTR and postprocessing placeholder options into the primary launch of PCA. This partnership between Principal and AWS enabled speed-to-market for Principal, whereas guaranteeing that current and incoming enterprise necessities could possibly be quickly added. The contributions to the open-source venture has accelerated different clients’ Genesys CTR workloads.
Reply enterprise questions
As soon as PCA was in place, Principal analysts, knowledge scientists, engineers, and enterprise house owners labored with AWS SMEs to construct quite a few Amazon QuickSight dashboards to show the information insights and start answering enterprise questions. QuickSight is a cloud-scale BI service that you need to use to ship easy-to-understand insights from a number of datasets, from AWS knowledge, third-party knowledge, software program as a service (SaaS) knowledge, and extra. The usage of this BI device, with its native integrations to the present knowledge repositories made accessible by Amazon Athena, made the creation of visualizations to show the large-scale knowledge comparatively simple, and enabled self-service BI. Visualizations have been rapidly drafted to reply some key questions, together with “What are our clients calling us about,” “What matters relate to the longest AHT/most transfers,” and “What matters and points relate to the bottom buyer sentiment scores?” By ingesting extra knowledge associated to Principal customized subject fashions, the staff was capable of increase their use of QuickSight to incorporate subject and correlation comparisons, mannequin validation capabilities, and comparisons of sentiment primarily based on speaker, section, name, and dialog. As well as, using QuickSight insights rapidly allowed the Principal groups to implement anomaly detection and quantity prediction, whereas Amazon QuickSight Q, an ML characteristic inside QuickSight that makes use of NLP, enabled fast pure language quantitative knowledge analytics.
When the preliminary initiative for PCA was full, Principal knew they wanted to instantly dive deeper into the omnichannel buyer expertise. Collectively, Principal and AWS have constructed knowledge ingestion pipelines for buyer e mail interactions and extra metadata from their buyer knowledge platform, and constructed knowledge aggregation and analytics mechanisms to mix omnichannel knowledge right into a single buyer perception lens. Utilization of Athena views and QuickSight dashboards has continued to allow basic analytics, and the implementation of proof of idea graph databases by way of Amazon Neptune will assist Principal extract insights into interplay matters and intent relationships inside the omnichannel view when applied at scale.
The Outcomes
PCA helped speed up time to market. Principal was capable of deploy the present open-source PCA app by themselves in 1 day. Then, Principal labored along with AWS and expanded the PCA providing with quite a few options just like the Genesys CTR integration over a interval of three months. The event and deployment course of was a joint, iterative course of that allowed Principal to check and course of manufacturing name volumes on newly constructed options. Because the preliminary engagement, AWS and Principal proceed to work collectively, sharing enterprise necessities, roadmaps, code, and bug fixes to increase PCA.
Since its preliminary growth and deployment, Principal has processed over 1 million buyer calls by the PCA framework. This resulted in over 63 million particular person speech segments spoken by a buyer, agent, or IVR. With this wealth of knowledge, Principal has been capable of conduct large-scale historic and near-real-time analytics to realize insights into the shopper expertise.
AWS CCI options are a game-changer for Principal. Principal’s current suite of CCI instruments, which incorporates Qualtrics for easy dashboarding and alternative identification, was expanded with the addition of PCA. The addition of PCA to the suite of CCI instruments enabled Principal to quickly conduct deep analytics on their contact middle interactions. With this knowledge, Principal now can conduct superior analytics to know buyer interactions and name drivers, together with matters, intents, points, motion objects, and outcomes. Even in a small-scale, managed manufacturing atmosphere, the PCA knowledge lake has spawned quite a few new use instances.
Roadmap
The information generated from PCA could possibly be used to make vital enterprise choices concerning name routing primarily based on insights round which matters are driving longer common deal with time, longer holds, extra transfers, and detrimental buyer sentiment. Data on when buyer interactions with the IVR and automatic voice assistants are misunderstood or misrouted will assist Principal enhance the self-service expertise. Understanding why a buyer known as as an alternative of utilizing the web site is vital to enhancing the shopper journey and boosting buyer happiness. Product managers liable for enhancing net experiences have shared how excited they’re to have the ability to use knowledge from PCA to drive their prioritization of latest enhancements and measure the affect of adjustments. Principal can be analyzing different potential use instances equivalent to buyer profile mapping, fraud detection, workforce administration, using extra AI/ML and enormous language fashions (LLMs), and figuring out new and rising traits inside their contact facilities.
Sooner or later, Principal plans to proceed increasing postprocessing capabilities with extra knowledge aggregation, analytics, and pure language era (NLG) fashions for textual content summarization. Principal is at the moment integrating generative AI and foundational fashions (equivalent to Amazon Titan) to their proprietary options. Principal plans to make use of AWS generative AI to reinforce worker productiveness, develop belongings beneath administration, ship high-quality buyer experiences, and ship instruments that permit clients to make funding and retirement choices effectively. Given the flexibleness and extensibility of the open-source PCA framework, the groups at Principal have an intensive checklist of extra enhancements, analytics, and insights that might prolong the present framework.
“With AWS Put up Name analytics answer, Principal can at the moment conduct large-scale historic analytics to know the place buyer experiences may be improved, generate actionable insights, and prioritize the place to behave. Now, we’re including generative AI utilizing Amazon Bedrock to assist our enterprise customers make data-driven choices with greater velocity and accuracy, whereas lowering prices. We sit up for exploring the put up name summarization characteristic in Amazon Transcribe Name Analytics with a view to allow our brokers to focus their time and sources participating with clients, somewhat than guide after contact work.”
– says Miguel Sanchez Urresty, Director of Information & Analytics at Principal Monetary Group.
Conclusion
The AWS CCI PCA answer is designed to enhance buyer expertise, derive buyer insights, and scale back operational prices by including AI and ML to the contact middle supplier of your selection. To study extra about different CCI options, equivalent to Live Call Analytics, seek advice from AWS Contact Center Intelligence (CCI) Solutions.
About Principal Monetary Group
Principal Financial Group and associates, Des Moines IA is a monetary firm with 19,000 staff. In enterprise for greater than 140 years, we’re serving to greater than 62 million clients in varied nations around the globe as of December 31, 2022.
AWS and Amazon will not be associates of any firm of the Principal Monetary Group Insurance coverage merchandise issued by Principal Nationwide Life Insurance coverage Co (besides in NY) and Principal Life Insurance coverage Firm. Plan administrative companies supplied by Principal Life. Principal Funds, Inc. is distributed by Principal Funds Distributor, Inc. Securities supplied by Principal Securities, Inc., member SIPC and/or unbiased dealer/sellers. Referenced corporations are members of the Principal Monetary Group, Des Moines, IA 50392. ©2023 Principal Monetary Companies, Inc.
This communication is meant to be instructional in nature and isn’t supposed to be taken as a suggestion. Insurance coverage merchandise and plan administrative companies offered by Principal Life Insurance coverage Firm, a member of the Principal Monetary Group, Des Moines, IA 50392
In regards to the authors
Christopher Lott is a Senior Options Architect within the AWS AI Language Companies staff. He has 20 years of enterprise software program growth expertise. Chris lives in Sacramento, California, and enjoys gardening, cooking, aerospace/normal aviation, and touring the world.
Dr. Nicki Susman is a Senior Information Scientist and the Technical Lead of the Principal Language AI Companies staff. She has in depth expertise in knowledge and analytics, utility growth, infrastructure engineering, and DevSecOps.
