Service Description

Version 2.1Last updated: 22 September 2020
  1. General Service Description
    • This document gives a high-level description of the products and services available from the Company. For the specific products and license entitlements you have subscribed to, please refer to the Service Order.
    • The legal terms of the licensing would be executed through a Software Licensing Agreement executed between the Company and the Client.
    • For a more detailed operational description of how Agara works, please refer to the user guides, the terms of which remain separate from this document are provided separately
  2. Agara: Autonomous Real-time Voice AI
    • Agara uses state of the art Real-time Voice AI technology to create autonomous conversation engines which can converse with customers and vendors of businesses across sectors autonomously and resolve their queries like a highly trained human agent
    • Agara also aids human agents with highly contextual suggestions in real-time so the agents can conduct their conversations without breaks or gaps
    • Agara operates in the following communication modes:
      • Assist Mode
        • Available over email and voice channels, Agara applies Real-time Voice AI technology to predict the next best action/response and presents it to the agent in a unified, highly optimized user interface
      • Autonomous Mode
        • Available over email and voice channels, Agara applies Real-time Voice AI technology to autonomously respond to emails or incoming calls without the intervention of a human agent
  3. Subscription Components
    • Base AI Subscription
      • This comprises of the artificial intelligence modules that address a particular communication mode (text/voice), level of autonomy (autonomous/assist), or functionality in the product (next-best-action, intelligent conversation, etc.)
    • Base Software Subscription
      • This comprises of common software components which are part of the license entitlement (for example, User Screens, UX, Datastores)
  4. Base AI Subscription
    • The following components/modules/capabilities are available as part of the Base AI Subscription:
      • Agara Assist for Email
      • Agara Assist for Voice
      • Agara Auto for Email
      • Agara Auto for Voice
    • Agara Assist for Email
      • This capability is designed to guide a human agent to respond to email queries faster and with fewer errors. It comprises of the following sub-modules:
        • Highly optimized, unified user interface
        • Case creation and management
        • Case coding and triaging
        • Knowledge prediction
        • Optimized Search
        • Product identification
        • Issue classification
        • Predicted response suggestion
        • Optimized text editor
        • Fulfillment module
        • Intelligent Alerts
    • Agara Assist for Voice
      • This capability is designed to guide a human agent to respond to phone conversations faster and with fewer errors. It comprises the following sub-modules:
        • Real-time speech transcription
        • Highly optimized, unified user interface
        • Case creation and management
        • Case coding and triaging
        • Knowledge prediction
        • Optimized Search
        • Product identification
        • Issue classification
        • Predicted response suggestion
        • Automatic prefetch of data
        • Autofill of forms
        • Automatically generated call summary
        • Fulfillment module
        • Intelligent Alerts
        • Surveys
    • Agara Auto for Email
      • This capability is designed to autonomously respond to email queries while completing multiple tasks in the background. It comprises the following sub-modules:
        • Instant personalized response to email queries
        • Case creation and management
        • Case coding and triaging
        • Product identification
        • Issue classification and identification
        • Configuration control
        • PII identification
    • Agara Auto for Voice
      • This capability is designed to respond autonomously to phone calls while completing multiple tasks in the background. It comprises the following sub-modules:
        • Autonomous voice response to phone calls
        • SDKs and toolkits for integration into other platforms
        • Case creation and management
        • Case coding and triaging
        • Product identification
        • Issue classification and identification
        • Automatic prefetch of data
        • Automatic capture of data
        • Automatically generated call summary
        • Automatic SMS and email triggers with contextual information
  5. Base Software Subscription
    • Software Components
      • Authentication mechanisms
        • Agara allows user access and authentication using multiple platforms. This includes access and authentication using Agara’s internal mechanisms as well as using 3rd party authentication services using appropriate security measures
      • Highly optimized user interface
        • Agara offers a user interface that has been optimized for its data and display requirements
    • AI Components
      • Automatic Speech Recognition
        • This component is used to convert speech into words
        • Agara may offer this capability using a component developed in-house or may work with appropriate 3rd parties
      • Custom Language Model
      • Natural Language Understanding/Processing
      • Conversational AI
      • Rules Engine
      • Text to Speech
        • This component is used to convert words into speech
        • Agara may offer this capability using a component developed in-house or may work with appropriate 3rd parties
    • Platform Components
      • User Management
      • Usage Monitoring
      • Billing Management
      • Workflow Management
      • Analytics/Dashboards
  6. Infrastructure [if applicable]
    • Agara core service is delivered via cloud infrastructure hosted on Amazon Web Services
    • Agara would host its service in a Virtual Private Cloud offering that will isolate the service from those being offered to any other customer of Agara
    • All client-specific data is separated from the core services in a client-specific access layer. The client-specific data access layer can be hosted in a cloud environment as per customer preference
    • In addition to the Production system that will be used to serve final customers, the Company will also provide optional and chargeable Sandbox systems for testing and rollout management purposes.
    • All systems will be fully managed by Company for appropriate load handling, backups, and redundancy
  7. Integration with Client Systems
    • In order to operate effectively, Agara may need to be integrated with specific IT systems owned and operated by the client
    • All integrations are enabled via standardized APIs. Non-standard setups may be provided by Company in specific circumstances
    • The integrations may be done in one of three ways:
      • Self-service by client IT team
      • Directly completed by Company
      • Executed by a suitable third-party identified either by Company or the client
      • The typical systems that are integrated with Agara include, but not limited to:
        • CRM systems
        • Telephony systems
        • Case Management tools
        • Email gateway
        • SMS Gateway
        • Order Management Systems
        • ERP systems
        • GDS
  8. Charge Metrics
    • Voice
      • Company will license Voice Subscription services based on the number of calls processed during the year
      • A Charge metric can be defined which gives customers the benefit of committing to larger volumes
      • This is applicable for both Assisted mode and Autonomous mode
    • Email
      • The Company will license Email subscription services based on the annual estimated number of email queries to be responded
      • A Charge metric can be defined which gives customers the benefit of committing to larger volumes
      • This is applicable for both Assisted mode and Autonomous mode
    • SaaS API Addons
      • The licensing metric will be on a case to case basis and will be decided based on the type of API subscribed to (example: Call summarization API)
  9. Professional Services and Set-up Charges
    • The integration of Agara into the client’s enterprise system will need professional services. Company or optionally Company partners can provide this service. The service provided by Company will be as per the fees mentioned in the Service Order.
  10. Peak Concurrent Capacity (for Voice subscriptions only)
    • The Peak Concurrent Capacity – max number of calls that can be in an active state at a given time - for the client will be mentioned on the Service Order and committed to by Company
    • Additional Peak Burst Capacity can be provisioned for limited periods or Peak Concurrent Capacity can be permanently increase based on client request. There may be charges applicable for both changes
  11. AI Base Subscription Overage
    • As mentioned in the Charge metrics section – Entitlements are decided based on Annual Call minutes or number of emails processed. Overages are applicable in case annual entitlements are exceeded. Overages will be as mentioned in the Service Order
  12. AI Adaptation Service Requests
    • The Company’s AI solution on deployment in a client’s context/function will need data Adaption services for the ML models to function. As part of the base product, 40-man hours of Adaptation services are included in the base licenses
    • 40 man-hours are packaged as 1 unit of adaptation services. In rare cases, if additional expert AI adaption services to incorporate activities beyond normal scope are required, it may be made available as additional chargeable units
  13. Training Data
    • The Company may request for data related to its services from the client to strengthen and contextualize the services (Training Data)
    • The data types that may be requested include the following:
      • Historical emails from customers
      • Historical call recordings of conversations between the contact center and its customers
      • Relevant metadata stored in IT systems including CRM, order management systems, knowledge systems, and others
    • The Company will utilize the data for various product development and AI training purposes only for the purpose of providing services to the Client
    • With written permission from the client, the Company may share the data with select third parties as specified below, to undertake specific tasks like speech transcriptions and annotation
    • Any training data received from the client will be stored in the Company’s cloud infrastructure where access will be limited and controlled to employees who have a need to view or handle the data
  14. Agara Generated Data
    • During the course of operations, Agara will process queries from end customers and other counterparties. There will be two kinds of data generated during this process
      • Case Data
        • This includes data pertaining to the customer/counterparty, emails, call recordings, call transcripts, and information fetched from client IT systems
        • Agara will immediately send Case Data back to the client’s IT systems at appropriate intervals using its integrations with client Systems
        • Upon writing Case Data, Company will purge all such data from its own systems permanently post completion of the mutually agreed upon retention period
        • The current retention period is determined to be 90 days
      • AI Performance Data
        • While handling a specific interaction, Agara would generate additional data related to predictions made by its AI modules, actions undertaken by its IT systems, and system/application logs
        • Agara will maintain this data in perpetuity
  15. Data Security
    • Company uses appropriate forms of secure technologies at all times in connection with processing and storage of client data, including in connection with any transfer, communication, remote access, or storage (including backup storage). The underlying cloud infrastructure follows the security industry's best practices.
  16. Usage and Ownership of Data
    • Any data classified as Client Data will remain, at all times, the sole property of the client, and the Company will not have or obtain any rights to this data
    • Client Data includes data
      • Training data Provided to Agara by the client
      • That the client collects, processes, generates, or uses for or on behalf of Agara
      • That Agara collects, processes, or uses on behalf of the Client
    • Client Data also includes metadata from Client’s use of the Agara services and any derivatives of any of such data (including aggregations of data, profiles of users, or Client’s websites) (collectively the "Client Data") and will at all times remain the sole property of the Client, and Agara will not have or obtain any rights to the Client Data. For purposes of this document, "Client Data" includes any information in any media or format relating to an identified or identifiable individual (including name, address, phone number, fax number, email address, Social Security number, date of birth, financial account, and credit card information, personal health information, IP addresses, or other identifiers used by Client, Agara, or a third party) ("Personal Information"). For purposes of this document, Client Data excludes the following derivatives: machine learning models and algorithms created by Agara or Agara personnel that do not contain Client Data or allow Client Data to be reverse-engineered or inferred (“Excluded Derivatives”). It is further provided that Agara is free to utilize Excluded Derivatives in any form or manner, without limitation, for commercial or non-commercial purposes, for internal or external uses, and towards providing services to the client or to other parties
    • Anonymous and aggregated usage data – Data Agara will capture and retain for any purposes, for performance tuning, management of infra, maintaining and monitoring usage and analytics
    • The Company’s exposure to client data is temporary and only for the duration of the email processing or live call. Once the email is responded to or the call is closed, all data including inferred, captured, and sensitive data is written back into master systems of record and purged from Agara data stores. Storage of PII or Client Data is not allowed in Agara
    • The Company will process Client Data in accordance with applicable laws in compliance with Industry best practices.
    • The Company will capture and maintain data about the performance of Agara’s solution in terms of key metrics about Call indicators, establish causality/performance improvement in key metrics like average handling time, agent productivity, call resolution times, wrap times – which are laid out in base software – analytics section.
  17. Third-Party Services
    • The Company may utilize services from select third parties from time to time. Any third party utilized by Company will be bound by substantially similar confidentiality terms as agreed upon by Company with the Client
    • The third parties currently in use by Agara are:
      • Amazon Web Services (AWS)
      • Google Cloud Platform
      • Amazon Mechanical Turk
      • Flitto, Inc.
      • DeepGram
      • Several individual contributors to create specific voice assets
    • The Company may add new third parties to this list at any time and notify the Client in writing of such addition.
  18. Client Reference
    • The client agrees to participate in a Case Study with the Company. The Client may be asked to participate in a video or phone interview about their experiences working with Agara, which may be utilized in a written case study and/or video testimonial
    • Any video or written Case Studies may be displayed on the Company’s website or distributed as part of the Company’s sales and marketing practices and may include use of the clients’ corporate logo
    • The Company may identify relationship with customer on its website and on sales and marketing collateral
    • The case study content is subject to written approval by the client before publishing
    • The Client also agrees to act as a reference for prospective customers, investors, or analysts. All of the items above are subject to written approval by the client

Service Levels

  1. Service Levels
    • The Company commits to make the service available 99% of the time (Uptime Commitment)
    • If Company fails to achieve this Uptime Commitment in any month, Company will provide credits (Service Level Credit) to clients equivalent to 2x the applicable per minute, per email, or per-call charge, as applicable, for the entire duration of the disruption. Service Level Credit will be applied to extend the client’s usage of Company services by an equivalent amount
    • Scheduled Maintenance with at least 7 days of notice will not be counted against the Uptime Commitment. Scheduled Maintenance means any system downtime and/or unavailability mutually planned and agreed upon by Company and the Client for technical or business reasons including, but not limited to, new update deployments
    • Service Levels apply to Production systems only and not to Sandbox systems
    • The Company will provide the client with on-demand access to up-to-date problem management data and other data regarding the status of service problems, service requests, user inquiries, and Service Level performance
  2. Incident Management Workflow
    • All identified issues must be raised by users to the designated people in the Matrix of Key Personnel – Technical
    • The designated personnel must estimate the severity level of the identified issue and file a report with Company via the available channels as mentioned in the Response Time & Channels table above
    • All incident reports must be accompanied by the following information:
      • Error description
      • Details of users affected by the issue
      • The time period over which the issue has been noticed
      • Steps to reproduce the error
      • [Optional] Screenshots, videos of the issue in action
    • Agara team may request additional information if the report does not contain enough data to pinpoint the issue. For critical issues, the person reporting the issue and the person raising the support request must be online and contactable during this period to provide additional information
    • Agara team, in consultation with client designated personnel, may escalate or de-escalate a reported incident based on information received
    • Depending on the findings of the Agara team and discussions with the Client team, fixes to an identified issue may be pushed into the next release cycle or deployed as a hotfix. The process for release will remain the same as outlined in the Release Management section in this document
  3. Updates and Release Management
    • Silent Updates
      • Machine Learning Model updates: Company will carry out regular retraining/optimization of ML algorithms/models on recent training data
    • Scheduled Updates
      • The Company will make monthly releases to its platform
      • Releases will be made on specific dates in consultation with the client
      • For Changes which have any one of the following components, the Client will be notified and the Process mentioned below will be followed for such releases
    • Changes:
      • Changes to the user interface
      • Changes to any component which interfaces with client systems or relevant third-party system
      • Changes to underlying data structures
    • Process:
      • The releases will be made initially to the sandbox environments and opened for testing by the client team
      • There will be a fixed period during which all testing must be completed
        • If no bugs or errors are found, the changes will be automatically deployed to production on the designated date without additional notice
        • If bugs or errors are found and rectified before the designated production deployment date, the changes will be deployed to production with notice to client
        • If bugs or errors are found and NOT rectified before the designated production deployment date, the changes will NOT be deployed and Company will either:
          • Push the changes to the next deployment cycle, or;
          • Where possible, disable the changes for client
      • For the purposes of testing and confirmation on changes being okay for deployment, a matrix of key personnel will be created. This matrix will serve as the final authority for all confirmations
      • It is expressly stated here that in the case of no response from the client, Company will do its best to notify the matrix of key personnel. However, in case no advice to the contrary is received prior to the designated deployment date, all changes will automatically be deployed to production
    • Client Specific Updates
      • At times, the client may request new features or modifications to existing features (‘change requests’) to achieve its business objectives.
      • All change requests must be logged into the Agara ticketing system with the appropriate label by the referred client personnel to be adopted for consideration and tracking
      • Agara will revert on new change request tickets within a maximum of 15 days from being logged
      • Change requests are taken up on a best-effort basis only, subject to the product roadmap and there is no upfront confirmation on whether or when the change requests will be taken up. The possibility, extent, and timing of a change request being accepted will remain the exclusive decision of Agara
      • All change requests will be discussed between Company and client before being considered for insertion into the platform
      • Change requests will only be considered for addition into a Regular Platform Release cycle
      • The following process will be followed for including a change request into the release cycle
        • After initial discussions, a Product Requirement Document (PRD) must be provided by the client covering all aspects of requested changes. If there is more than one change requested, a PRD must be provided for each of the changes
        • Depending on the substance of each PRD, Company will, in consultation with the client, decide which release cycle the change can be inserted in (if the change is found to be feasible)
        • Further, both parties will also decide when third-party changes need to be ready, where applicable, for the Company to work on the requested changes
        • Company will target completing all work for the change request the sandbox deployment date in the target release cycle. However, if, for any reason, work is not completed before the sandbox deployment date, the changes will be pushed to the next release cycle
    • Change Management Process from Client
      • Given its nature, data or structure level changes (by the client as well as relevant third parties) may have an impact on the smooth functioning of the Agara solution. To this extent, an important part of the change management process covers changes in these external systems
      • The Client must alert the Company about an impending change as early as possible. This is important to ensure timely action and minimizing disruption to its platform
      • All such changes must be logged into the Agara ticketing system with the appropriate label to be considered as a notice
      • Minimum notice period to be provided for changes:
        • Data related changes: 30 days
          • This includes new product/comment paths or changes in existing paths
        • System related changes: 90 days
          • This includes data structure related changes, CRM related changes, and changes to any third parties that Agara connects to
      • Agara team will consult with the client team (including requesting additional information) to establish the extent of disruption the proposed change may cause and the requisite changes to be made in the Agara platform to minimize the disruption
      • All decisions of making changes in the Agara platform will be made as per the process outlined in the Client Requested Releases section above
    • Scheduled Maintenance
      • From time to time, Company may need to undertake maintenance on its own infrastructure including changing machines deployed for the client, upgrading systems, archiving of data, and more. The Company will endeavor to complete such maintenance without impacting operations of the clients
      • In case the maintenance requires a downtime, Company will inform clients at least 7 days in advance so appropriate measures can be taken

Start optimizing your voice support today

One of our friendly and knowledgeable experts is always at hand to partner you through your transformation journey.

Schedule a Demo

Learn how Real-time Voice AI can help you

Schedule a Demo

Have an urgent requirement and wish to talk to a sales rep? Call +1-512-333-4634 now