Joulica Analytics Platform for Amazon Connect

Amazon Connect Analytics  addresses our main use case in the energy sector, where we provide energy and power for 95% of this particular state. Understanding why customers are contacting us and proactively improving our service to them, whether in terms of billing, outages, payments, or new connections, is very critical. Amazon Connect Analytics  includes Contact Lens, real-time metrics, and historical metrics, providing agent performance data and sentiment analysis. It helps any energy company monitor contact drivers, reduce call volume, and improve customer satisfaction, especially during high-stress periods such as outages, tariff changes, or seasonal demand spikes which we are experiencing as summer approaches.

Our main use case using Amazon Connect Analytics involves moving from reactive to proactive service. The shift Amazon Connect Analytics enables, from point A to our initial stage, involves transitioning from waiting for customers to explain issues to predicting and addressing potential problems even before they escalate. Amazon Connect Analytics shows us exactly what customers attempted before contacting us, where self-service journeys break down, and which intents consistently perform poorly. This information helps us redesign our digital journey based on evidence instead of guesses. Additionally, we're linking these contact insights to business KPIs; our advanced team integrates analytics with billing system data, metering systems, and CRM  tools such as Salesforce . This visibility illustrates relationships, such as identifying that calls with a negative sentiment analysis in the last three days correlate with three times the likelihood of churn, transforming the contact center into a strategic intelligence hub rather than just a call center.

Amazon Connect Analytics helped us reduce our billing complaint calls by almost 30 to 35% in just two weeks. We noticed a sudden spike in calls after issuing quarterly bills, leading to customer frustration, long queues, and overwhelmed agents. Amazon Connect Analytics assisted us significantly by providing automatic transcription and theme detection. Within hours, Contact Lens surfaced a pattern across the thousands of calls revealing the phrase "estimated bill" appeared unusually frequently, while terms such as "new smart meter," "readings not updated," and "higher than usual charges" emerged from callers. Customer feedback indicated highly negative sentiment scores for these calls, allowing us, using Amazon Connect Analytics, to identify the root cause. About 50% of the negative statement calls used the phrase "estimated bill," showing a correlation with the rollout of smart meter upgrades in two of our regions, where the system failed to receive automated meter readings for approximately 12,000 customers, triggering estimated bills. Fortunately, this insight took just a couple of hours, not days or weeks, thanks to Amazon Connect Analytics.

Some of the best features of Amazon Connect Analytics that I personally endorse include its real-time capabilities. The real-time conversational analytics allows Contact Lens to transcribe any call or chat live, analyzing sentiment in real time and showing if there is any trace of negative sentiment or distress. Above our agents, supervisors receive real-time alerts. For example, if a customer expresses frustration with words such as "disconnect notice," "I'm not happy," or "cancel," supervisors gain immediate alerts and can intervene during the call. This allows us not to wait until after a call to spot problems, preventing escalations and improving first call resolution. Additionally, post-contact summaries generated by Amazon Connect Analytics provide an AI-powered summary of each conversation, highlighting issues, next steps, sentiment, and outcomes to aid supervisors, agents, and other teams without needing to replay entire calls. This feature reduces manual effort while ensuring a faster response and follow-up, particularly during contacts involving multiple departments, such as billing, network, and hardship support teams.

Amazon Connect Analytics includes a valuable feature called silence detection and talk-time patterns. It flags any long silences, over-talking by agents, excessive interruptions from customers, or excessive hold times, quickly revealing process pain points. For instance, if an agent spends 40 seconds waiting for a slow billing screen to load, or if customers repeatedly interrupt scripted lines they find irrelevant or annoying, we get notifications of these issues. Another valuable feature is the automatic sensitive data redaction, which removes sensitive data such as credit card numbers and personal identifiers from transcripts and recordings, helping us comply with government regulations as an energy sector provider. Furthermore, we have a phrase-level timeline view, allowing us to pinpoint when specific phrases were used in the call transcript. For example, if a customer mentions "disconnect" at 3 minutes and 12 seconds, we can instantly check the precise timing, speeding up call reviews and making training significantly more efficient. We also receive per-agent behavioral insights, highlighting politeness markers and opportunities for coaching without supervisors needing to listen to multiple calls. Lastly, we have a heat map of issue frequency, tracking issues over time, revealing when calls peak, such as during outlier event times.

Amazon Connect Analytics could be improved in several ways. Sometimes, Contact Lens struggles with sarcasm. For instance, if a customer says, "Oh great, another billing increase," it might mistakenly register that as positive sentiment. Additionally, it may not effectively capture cultural or linguistic nuances. As ML models advance, enhancing the sentiment analytics tools to become more context-aware, recognizing human tone, stress, sarcasm, and frustration more accurately would be beneficial. Regarding the dashboards, while functional, they remain quite basic, presenting challenges in custom visualization. We occasionally must export data into platforms such as Tableau or Power BI for deeper insights, so having more powerful and customizable dashboards built into Amazon Connect Analytics could be incredibly advantageous.

I have been using Amazon Connect Analytics for around two years now.

Amazon Connect Analytics is stable.

Amazon Connect Analytics is highly scalable, efficiently accommodating call volumes ranging from thousands to millions.

We contacted the Amazon Connect Analytics back-end support team for clarification regarding billing for Contact Lens usage and guidance on setting up transcription and sentiment analysis. They were responsive, providing answers within hours due to our business or enterprise plan, and displayed good expertise with both Amazon Connect  and Contact Lens analytics, which significantly aided our troubleshooting and configuration efforts.

We previously used a legacy on-prem contact center solution, which presented numerous limitations such as manual call monitoring, slow responses to spikes, and high operational costs. We ultimately switched to Amazon Connect Analytics because it is cloud-native and scalable, eliminating the need to manage any services. This transition simplified our ability to scale from thousands to millions of interactions while benefiting from integrated AI analytics and flexible pay-as-you-go pricing that aligns costs with actual usage, avoiding large upfront expenditures.

In terms of pricing and setup costs, Amazon Connect Analytics operates on a pay-as-you-go model, which is great because it gets billed per usage rather than through licensing. Voice calls are analyzed per minute and chat messages per message, with storage for recordings and transcripts billed per GB in S3 . The pricing remains transparent and predictable, making scaling up or down with call volume straightforward. There are no upfront licensing fees and no significant capital investment is necessary, maintaining alignment with actual usage throughout low periods. However, during high volume periods, such as outage surges or storms, costs can noticeably increase, especially if additional AWS  services such as Lambda or EventBridge are deployed, which might introduce complexity.

Amazon Connect Analytics results in significant time savings from faster issue detection, call summarization, and improved agent coaching. The average handling time has decreased by nearly 30% following the introduction of AI-powered call summaries. Supervisors can avoid the time-consuming task of reviewing every call manually, which saves considerable hours weekly. Our issue detection time has shrunk from days to hours, and even down to just a few minutes, resulting in accelerated resolutions and reduced follow-up calls. We have also seen a reduction in repeat calls by about 30 to 35%, improving our workforce efficiency and enabling better forecasting and agent allocation. This enhanced operational efficiency leads to an 8 to 10% reduction in staffing costs while still meeting service level agreements, and it has helped decrease churn in at-risk segments by around 6 to 8%, boosting revenue.

We contacted the Amazon Connect Analytics back-end support team for clarification regarding billing for Contact Lens usage and guidance on setting up transcription and sentiment analysis. They were responsive, providing answers within hours due to our business or enterprise plan, and displayed good expertise with both Amazon Connect  and Contact Lens analytics, which significantly aided our troubleshooting and configuration efforts.

I have witnessed a return on investment with Amazon Connect Analytics. The time saved from faster issue detection, call summarization, and improved agent coaching represents a significant benefit. The average handling time has decreased by nearly 30% following the introduction of AI-powered call summaries. Supervisors can avoid the time-consuming task of reviewing every call manually, which saves considerable hours weekly. Our issue detection time has shrunk from days to hours, and even down to just a few minutes, resulting in accelerated resolutions and reduced follow-up calls. We have also seen a reduction in repeat calls by about 30 to 35%, improving our workforce efficiency and enabling better forecasting and agent allocation. This enhanced operational efficiency leads to an 8 to 10% reduction in staffing costs while still meeting service level agreements, and it has helped decrease churn in at-risk segments by around 6 to 8%, boosting revenue.

In terms of pricing and setup costs, Amazon Connect Analytics operates on a pay-as-you-go model, which is great because it gets billed per usage rather than through licensing. Voice calls are analyzed per minute and chat messages per message, with storage for recordings and transcripts billed per GB in S3 . The pricing remains transparent and predictable, making scaling up or down with call volume straightforward. There are no upfront licensing fees and no significant capital investment is necessary, maintaining alignment with actual usage throughout low periods. However, during high volume periods, such as outage surges or storms, costs can noticeably increase, especially if additional AWS  services such as Lambda or EventBridge are deployed, which might introduce complexity.

Before choosing Amazon Connect Analytics, we evaluated Genesis Cloud CX and Cisco's Webex  Contact Center. Ultimately, we opted for Amazon Connect Analytics because it is fully cloud-native, managed by AWS, and seamlessly integrates Contact Lens analytics while providing easy connectivity to other AWS services such as S3, Lambda, and QuickSight .

The measurement method for tracking improvements in customer sentiment scores is critical, especially when presenting credible metrics. The primary method we have used involves Amazon Connect Contact Lens sentiment scoring to assess sentiment improvements. Contact Lens automatically assigns customer and agent sentiment scores, tracking the call's emotional trajectory. It indicates whether sentiment moves from positive to negative or vice versa, highlighting key moments during the conversation. The sentiment is typically expressed on a scale ranging from negative to neutral to positive, backed by ML models trained on conversational patterns. We compare sentiment data from a baseline period before major fixes to data collected post-change. For instance, in a six-week project, the first two weeks are baseline, while the middle two weeks reflect the impact of a billing portal fix, and the last two weeks show results after training on empathy scripting. This method allows us to track the percentage of calls classified as negative, the percentage with improving sentiment trajectories, and the average sentiment score per call. Furthermore, we segment sentiment analysis by call type, identifying specific journey areas in need of improvement. Amazon Connect Analytics provides dashboards that deliver sentiment heat maps and trends.

My advice for others considering Amazon Connect Analytics is to research thoroughly before implementation. It is essential to start small and scale quickly; begin with a subset of calls or agents to comprehend capabilities and workflows. Testing transcription accuracy and sentiment analysis through pilot programs before full deployment can yield immediate ROI by addressing pain points in specific areas, such as billing inquiries. Prioritizing high-impact use cases and planning integrations early can mitigate future bottlenecks with BI dashboards or automation workflows. Additionally, making use of the available AWS documentation and webinars is highly beneficial.

I chose a rating of eight out of ten for several reasons. Firstly, the built-in transcription and sentiment analysis are excellent. The accuracy of the speech-to-text feature is remarkable, even understanding Australian accents and background noise, delivering tremendous value compared to manual QA processes. The rapid time to insights with minimal setup is impressive, as I can access dashboards for agent performance, customer sentiment, and trending issues immediately upon opening. Its scalability is outstanding, accommodating anywhere from 10,000 calls to 10 million without compromising performance due to its native AWS integration. Furthermore, it supports CX and cost improvement; in our case, it helps reduce repeat contacts and quickly identify billing system concerns, while also assisting in agent coaching. The reason I deduct two points is due to the limited customization of the dashboards and the complexity involved with advanced queries, which often necessitate external tools. Moreover, while real-time analytics perform well, deeper insights would enhance the experience, alongside some limitations in UI integrations.

My main use case for Amazon Connect Analytics  is for agent time, including the times they are engaging with customers, and their reviews and analytics. We also send that data to our team to find out information.

A specific example of how I use Amazon Connect Analytics  in my day-to-day work is whenever a recent case is opened; I go and check the analytics, looking at how the performance or reviews and how the sentiment analysis are working these days.

The best features Amazon Connect Analytics offers include the filtering out and the visualization of agents and the data transition.

What I appreciate about the visualization and data transition features is that it provides a user interface that is easy to use. Instead of going and searching and running the queries manually, visualization and easy filtration help us.

Amazon Connect Analytics has positively impacted my organization, with noticeable changes in efficiency and customer experience.

We introduced AI-generated responses in chat, so we have better AI-generated responses and specific responses from feedback. We can see that from the analytics and the sentiment analysis across the board.

With the sentiment analysis and feedback, I've seen measurable improvements in customer satisfaction and efficiency, increasing by about ten percent more than before.

Amazon Connect Analytics is good; I don't have anything regarding how it can be improved.

I chose eight out of ten because there could be more generic functionalities and giving specific responses. I'm expecting more AI-related responses, natural language processing, or the visualization.

I've been using Amazon Connect Analytics for about two years.

Amazon Connect Analytics is stable.

The scalability of Amazon Connect Analytics is good, as it auto-scales based upon needs.

The customer support for Amazon Connect Analytics is quite good.

Before Amazon Connect Analytics, we were using some Genesis features, but I'm not sure exactly which ones.

I've seen a return on investment, as fewer integrations will not need our specific employees.

I purchased Amazon Connect Analytics through the AWS Marketplace , and my experience has been good.

Regarding pricing, setup cost, and licensing, I think it's good, but I'm not sure about the specifics.

I wasn't aware of other options evaluated before choosing Amazon Connect Analytics.

My advice to others looking into using Amazon Connect Analytics is to look for positive feedback. I rated this product eight out of ten. I found this interview to be good; there's nothing I think you should change for the future.

My main use case for Amazon Connect Analytics  involves one of our retail companies who is ordering several shoes, where customers order several shoes for them. We have to build a dashboard on how many return customers, how many customers are returning the shoes, how many customers are buying the shoes, what age group they buy the shoes, and what's the demographic of every customer. It can be very region specific; for instance, let's say they're buying shoes only from Boston. Therefore, we try to put some analytics in place to say which region is servicing or ordering the most, and then we stay focused on those markets with the help of that analytics.

The best features Amazon Connect Analytics  offers include the granularity of the calls that we upload into the system, so if we upload more than 10,000 customer interactions in a day, it provides us the information back.

Amazon Connect Analytics has positively impacted my organization 100% as the results are phenomenal; now my operations team can actually see how many orders are getting rejected so that they can work on those cases and satisfy the dissatisfied customer.

I can share that our CSAT scores have increased from a 70% baseline to around 85% today, and the returns have reduced from approximately 1,000 returns we get in a month down to almost 50%.

I believe Amazon Connect Analytics can be improved, especially regarding the interface, as it would be helpful for us since we have to do a lot of development work to make sure the interface looks exactly the way our customers want. I'm talking about the user interface.

Further about the needed improvements, the UI does give us the entire pattern, but you have to customize that pattern. What we prefer is that if those patterns or designs can be pre-built or available in the form of stencils, allowing us to leverage each stencil one by one instead of recreating from scratch every time.

I choose an eight because, as I said, it gives us a lot of benefits, but the journey to reach that level is a little painful. It took us roughly two months to fine-tune a particular analytics dashboard, and once it's fine-tuned, then it works great. That is the process we have to go through, and that's the reason I suggest maybe we could implement a little bit more stencil-based or pre-ready designs into the analytics dashboards, which would be helpful.

I have been using Amazon Connect Analytics for three years.

Amazon Connect Analytics is stable.

We haven't tested more than 400 users, but I hope it will work fine even for 4,000 or 10,000 users as well.

The customer support for Amazon Connect Analytics is phenomenal; we get support from NAMER SIs services very quickly, typically receiving a response back in just 15 minutes.

Before switching to Amazon Connect , we used Avaya, which was not able to offer us any cloud-based platform, prompting our full switch from Avaya to Amazon.

I have seen a return on investment with around 22 use cases on Amazon Connect Analytics, as almost every use case is benefiting in terms of reducing our attrition by 10%, which is helpful for a call center.

Before choosing Amazon Connect Analytics, we evaluated a platform called Jolika, which is a relatively new platform.

We receive information from Amazon Connect Analytics through various formats, including dashboards, which will be shown to our operations people. The contact center operations will look at these dashboards and then decide the outcome or the journey and what's their next action based on that.

Regarding the features, I would say that dashboards are difficult to build but easy to use. We have a specialist whose job is to build those dashboards and ensure that they meet all operational requirements. Once it meets that, it's very easy to use, but going through that journey is a little bit of a pain. We have to go back and forth with the operations and the system a lot. It typically takes us one to two months to fine-tune a report, but once it's fine-tuned, then it just keeps working.

I would advise others looking into using Amazon Connect Analytics to do good due diligence and meet customer expectations and operational demands before starting to build the dashboard, ensuring they gather the requirements first. I have given Amazon Connect Analytics a rating of 8 out of 10.

Amazon Connect Analytics  allows customers and users to gain insights into their contact center data, understand their volume, and build dashboards to achieve visibility.

The best features Amazon Connect Analytics  offers are the custom dashboarding capabilities, which allow me to build, cut, and slice the data according to my specific needs rather than relying solely on the out-of-the-box dashboards.

The custom dashboarding helps me and my clients in day-to-day work by providing visibility based on custom attributes on the contact, enabling us to build reports tailored to our requirements.

Thanks to the visibility and flexibility of the dashboards in Amazon Connect Analytics, I can act upon the actual events occurring in the solution. We now have more operational insights into what is happening in our environment.

Amazon Connect Analytics could be improved by being even more flexible, particularly in allowing dashboard embedding in third-party applications. If we want to embed the dashboards in another tool, that capability would be valuable.

We have used Amazon Connect Analytics on some projects a few years back.

Amazon Connect Analytics is stable and performs well in my experience.

The scalability of Amazon Connect Analytics is performing well.

The customer support for Amazon Connect Analytics is sometimes not as responsive and helpful as we would prefer, but it is not the worst. We have had a few occasions where we received great help from their support team.

Before Amazon Connect Analytics, we used Calabrio , but we wanted a more cost-effective solution. Amazon Connect Analytics data analytics seemed to be a better option for our needs.

My experience with pricing, setup cost, and licensing for Amazon Connect Analytics has been satisfactory so far.

Before choosing Amazon Connect Analytics, I evaluated Calabrio , which we had used previously.

We identified that too many fraud calls were coming to our contact center. Thanks to the identification of those calls through Amazon Connect Analytics, we were able to make changes to our flow, adjust the experience, and prevent long-lasting fraudulent calls.

My advice for others looking into using Amazon Connect Analytics is to deploy it to your account and ensure the data resides within your organization under your control, allowing you to customize your reports. I would rate this product an 8 out of 10.