Contact centers, in order to meet rising customer expectations and manage operational efficiency, have been adopting AI-powered tools that help them understand and optimize customer conversations. Amazon Connect Contact Lens, first launched at AWS re:Invent 2019, provides machine learning (ML) capabilities for analyzing call recordings and customer interactions within Amazon Connect. Contact Lens also allows supervisors to gain real-time visibility into live customer interactions, including sentiment progression and alerts based on predefined words and phrases.
This article explores some of the key features of Contact Lens, a refined AI-powered analytics service built into Amazon Connect, and how companies are using it to improve service quality.
What are the Main Features under Amazon Connect Contact Lens?
Amazon Connect Contact Lens records, transcribes, analyzes, and evaluates customer-agent conversations in real time and after the call. It supports both agents and supervisors by enhancing workforce efficiency and optimizing service quality. Using speech-to-text (STT) transcription and sentiment analysis, Contact Lens detects customer emotions, tracks keywords, and identifies compliance risks. It also enables data-driven decision-making in customer service operations.
If you navigate the Amazon Connect Console, you will mainly find these tools:
- Dashboards and Reports: Amazon Connect provides built-in dashboards for Contact Lens analytics, offering summaries of performance metrics and other insights.
- Real-time Metrics / Historical Metrics: These metrics track contact center performance in real time or over selected time periods.
- Login/Logout Report: This report logs agent activity, including login and logout times, to monitor agent availability.
- Contact Search: This feature lets users review data such as transcripts, sentiment scores, and conversation recordings—essential for quality assurance and training.
- Rules: Contact Lens can trigger real-time alerts based on predefined sentiment or keyword detection.
- Custom Vocabularies: Users can add organization-specific terms to improve transcription accuracy.
- Evaluation Forms: Built-in QA forms support consistent agent evaluations within the console.
These features support both agents and supervisors by offering real-time and post-contact analytics, searchable records, and supervisor alerts. Contact Lens tracks sentiment and compliance issues during live calls, enabling quick action. After each call, transcripts with emotion markers, keywords, and talk-time stats are available for analysis. Past conversations can be easily found by filtering for keywords, customer emotions, or agent behavior. Additionally, supervisors receive alerts if a live call appears to encounter issues, allowing them to respond promptly. Based on the customer’s needs addressed by MegazoneCloud, several features were tested. You can check the Contact Lens availability, by region(here) and by language(here) in the documentation provided by AWS.
Feature Enablement and Test
Conversational Analytics in Amazon Connect Contact Lens
1. Enable Contact Lens on the Amazon Connect instance and check security profiles.
First, you need to enable Contact Lens service to the instance.
| AWS Console → Amazon Connect → Select instance → Applications → Analytics tools → Contact Lens → Click on Enable → Save |
Once the service is enabled, then check the security profiles since agents need specific permissions to access Contact Lens data.These security profiles should align with the organization’s policies and requirements.
2. Enable call recording and speech analytics to a flow
Once Contact Lens is enabled for the instance, a Set recording and analytics behavior block can be added to a flow, as shown in the image.
In this block, you can also configure options such as agent screen recording and chat analytics, with detailed settings including language, redaction, sentiment, and Contact Lens GenAI capabilities.
3. Check Contact Lens analytics results
Once these settings are complete, agents can view the real-time transcript on the CCP panel. If you switch to the administrator role, you can view contact details by navigating to Contact Search.
| Amazon Connect console → Analytics and optimization → Contact Lens → Contact Search |
You can select a time range of up to eight weeks and choose the time zone if needed. The table can also be downloaded as a .CSV file.
By clicking on a Contact ID, a detailed view will appear. It includes information such as Contact Details, Sentiment Analysis, Call Recording, Speech Transcript, and other relevant metrics.
The Contact Details section provides key information such as the contact’s status, contact ID, channel type and duration, as well as the queue name and agent name. It also breaks down agent talk time, customer talk time, and periods of silence. In the Sentiment Analysis section, machine learning-based natural language processing (NLP) assigns sentiment scores to both the customer and the agent on a scale from -5 (very negative) to +5 (very positive). The Call Recording is stored in an Amazon S3 bucket, allowing external access when needed. The Speech Transcript provides a text version of the conversation, making it easier to review and analyze. Finally, the page includes additional relevant contact metrics, such as connection details, queue and agent information, or custom tags.
Generative AI-powered post-contact summaries
Generative AI-powered post-call summarization enhances agent productivity by displaying a summary of the conversation once it ends, helping agents save time and reduce manual note-taking. It also improves consistency by providing summaries of past interactions. This data can be integrated with other applications via API.
It also improves managers' efficiency in reviewing interactions. Managers can refer to highlighted key elements—such as reasons, actions, results, and follow-ups—within the Contact Details section on the Contact Search page. With this feature, managers can reduce the effort required for contact reviews and quickly access summaries.
Evaluation Forms
Evaluation Forms are used to assess agent performance during customer interactions by providing a customizable form to review and score various aspects of a call, such as adherence to scripts, customer satisfaction, and compliance. Supervisors can create or modify criteria specific to their business needs. Integrated with Contact Lens, Evaluation Forms provide automated or manual call assessments and generate reports which help enhance overall agent performance.
Use Case
Contact Lens is one of the most favored features of Amazon Connect among Users. While not highlighted in this article, various features of Contact Lens are actively used by customers. For example, one of our customers uses dashboards and reports to monitor contacts on a regular basis and agent adherence. Supervisors receive email alerts about potential issues such as waiting calls, and they can intervene through the real-time metrics dashboard when identifying problems during calls. The customer has also enabled the Evaluation Form feature for quality assurance and regularly reviews call analysis provided by Contact Lens.
Summary
Amazon Connect Contact Lens is an AI-powered analytics tool that helps contact centers improve service quality through real-time monitoring, post-call insights, and quality evaluations.
| This article was written as part of the AWS Ambassador Candidate activities. 이 글의 작성자는 AWS Ambassador Candidate로 활동중입니다. |
