auvious

Whitepaper Version 3.0

20/02/2021 www.auvious.com

Executive summary auvious has created a flexible and customizable video solution to meet market ​ needs for customer support and engagement.

Customers can connect with businesses or institutions with a single click on ​ our embeddable widget for a video or voice call, supplemented with features like screen sharing, co-browsing, follow-up call scheduling, and more. ​

Our continuous development and your invaluable feedback has allowed us to refine our solution with new features. We can now cater to nearly every ​ possible use case across all industries. Whether your client is a bank or ​ insurance company, a retailer, a healthcare provider, or a law firm – our solution has got you covered.

Within the span of a year, we have put forward a video solution to the Genesys ​ marketplace. It is already tailored for Genesys Cloud, integrating directly from ​ within the cloud application, and will soon be available for Genesys Engage. Additionally we can provide the solution as a standalone application or integrate with your existing contact center infrastructure.

The auvious team is dedicated to sustaining a state-of-the-art, customizable ​ solution.We are always open to accommodate particular needs that remain ​ unmet, and strive to adapt and deliver on that swiftly. ​

Simplicity and transparency is what we aim for – no hidden costs, no ​ activation fees, no limitations on licenses. With just $38 per agent*, your ​ clients can upscale customer and sales engagement, team collaboration, time efficiency in communication, workforce management, and so much more.

Solution description

Real time communication over the is not an easy task. The most common phrase on voip software is the “can you hear me now?”, and it’s not because the apps are bad, but because the internet was not designed to do real time audio/video in the first place, because the cost would probably prohibit it’s fast and wide adoption. The following table will make it more clear, why real time communication over the internet is difficult.

Internet PSTN/GSM QoS best effort (packet based) Guaranteed (real or virtual circuit based) Hardware Vendors many few Software vendors many few Device complexity big (os,browsers,versions) small End user cost cheap cheap nationwide, expensive world wide (roaming costs etc…) …. …

In a nutshell, internet users want free real time audio and video communication over a vast range of devices, and over a medium that does not make any guarantees about network throughput, latency, jitter etc…

auvious addresses these problems because: ​

● It’s based on WebRTC which is an open standard widely adopted by all modern browser/device vendors like Apple, , etc… This means we build our services on top of a solid stack, which addresses

most of the network and compatibility issues in the most efficient manner available, and we our platform benefits as this standard evolves. ● It’s carefully designed, with flexibility and scalability in mind, using state of the art design techniques and patterns. ● It makes heavy use of industry standard open source solutions. ● It heavily relies on automated testing to improve QA efficiency. ● It’s cloud native architecture and cloud infrastructure vastly helps in providing high availability and scalability. ● Active and passive intelligent monitoring ensures high availability, and fast detection and handling of system failures. ● Last but equally important, customer feedback is a first class citizen, when it comes to deciding what and how to improve in our solution. auvious uses a flexible architecture which is heavily based on modern microservices paradigms-patterns. Core components are often bundled together in a deployable service, while certain components are isolated in deployable microservices. This design decision benefits from microservice architecture where it is deemed appropriate, ensuring that we have a minimum burden from the maintenance overhead that tens or hundreds of microservices would impose. The bundled components however are implemented in such a way, that they can easily evolve/split into microservices when the time is right

The technology stack of auvious includes the following: ​ ​

auvious is built using industry standard open-source technologies. ​ User Interface components are based on Angular 7 which offers modern web platform capabilities, high performance and component routing. Backend components are mostly based on the Spring Framework and Spring Boot project. It addresses in a “magic” manner a lot of production ready service features, while being lightweight enough to be used for microservices. RabbitMQ is used as the message broker because it is fast,pluggable, simple in management & with detailed documentation, and a large active community supporting it. At the heart of the real time communication system is Janus SFU. This component enables efficient WebRTC media communication, relaying RTP/RTCP and messages between peers. It also includes pluggable modules which can support everything from a small nettop/, to a full-fledged WebRTC gateway on the cloud. InfluxDB is utilized for storing processing and visualizing important statistics related to video audio quality. All sessions push statistics both from the browser clients and the , and these are utilized to monitor and address audio,video quality issues. The software is built, tested automatically and deployed to Google Kubernetes Engine through Gitlab pipelines. Rolling updates ensure no service disruption during service updates. Blue/green/rainbow deployments is something we have in mind to add in the future. Operations are based on Stackdriver Logging and GKE tools. Logs from all components are collected, correlated and available for search. Operation dashboards are based on GKE.

auvious provides the following advantages: ​ ● Designed with customers and market value in mind. ● Engineered to provide the best applicable audio/video quality ensuring the best customer experience. ● Works with most modern desktop and mobile browsers (lightweight, works even with low-capability clients). ● Minimal bandwidth requirements. ● Flexible and adaptable to efficiently comply with potential security restrictions & regulations. ● Automation in every possible way, ensures a 24-hour functioning system quickly evolving into the tool of choice that users need.

Features

Click to video : Cross browser, cross platform and mobile customer video support.

Device audio fallback : When joining a call if no camera device is detected the call will still be allowed by falling back to audio only by selecting one of the existing audio devices.

Mute audio: The ability for a participant to mute/unmute their microphone ​ during a call and immediately propagate this change to all participants.

Disable camera: During a call a participant can explicitly disable/enable their ​ video stream and immediately propagate this change to all participants.

Mobile switch camera: During a call from a the participant can ​ easily switch between the front & back camera.

Pause call (OnHold): Ability for an agent to instantly suspend audio & video of ​ all customers participating in a meeting.

Quality & security enhancements: Under the hood improvements for greater ​ security and enhanced video call quality and overall stability.

Screenshare: Both way screen sharing functionality. Supporting latest browsers ​ without the need of a plugin installation.

Embeddable Widget: With our revolutionary widget the customer initiates a ​ video conversation directly with an agent, bypassing the need to go through other channels first. Using our floating widget your customer never leaves your website with chat conversation also supported. Additionally, customers are able to initiate a voice call, a webchat or schedule a video call.

Scheduled callback: Using this feature the agent can schedule a video call for ​ a later time. The customer receives the meeting url along with the scheduled date and time via email & the agent receives an incoming callback interaction. White-label : Ability to change the look & feel of the app ( colours, logos , fonts )

Augmented Reality pointer: Enables agents to interactively add markings ( ​ pointers ) to customer screens for enhanced, step by step customer assistance.

Network indicator: Real-time call quality monitor & feedback e.g. slow network ​ detected

Multiple participants: Ability for multiple agents to participate in the same call. ​

Omnichannel support: Agents can initiate calls from a single system regardless ​ of the preferred channel of engagement.

Recording service: Ability to record audio & video media of all participants in a call. The agent will have the ability to save the output in their preferable storage provider (S3 Amazon, Google etc.).

The recording service has been designed as a separate component, engineered to efficiently and robustly manage all media related recordings and storage of associated files. Having a designated recording service ensures that high quality recordings occur seamlessly in the background with no performance cost to either the agent or customer.

Storage Providers: We intend to support multiple cloud object storage providers in order to provide our customers with flexibility of choice when it comes to securing and storing their recorded files. Initial storage provider support will include Google and Amazon S3, with more provider options planned in upcoming releases.

Storage Format: Recording files are stored as objects in the given provider's object storage device. Some providers refer to these objects as BLOBs others simply as objects. Regardless of the name object storage stores data as objects and is the recommended means for storing media files such as photos, videos etc. Object storage provides many benefits in terms of data integrity, scalability and performance. This facilitates our constant efforts in ensuring efficient distribution and storage integrity of your high quality media files at the lowest possible cost. The ability to export a recording to a viewable format (such as mp4, mkv etc) on demand is also planned in an upcoming release. Analytics of all recordings will be available through our system for any retrieval or investigative purposes. All information stored is anonymous i.e. we do not process or store any information that relates to an individual who can be identified from that information.

Recording size estimations: Average 1 minute duration video stream: 1.8 MB Average 1 minute duration audio stream: 0.3 MB Average 1 minute duration audio/video: 2.11 MB

Example: 1 minute duration consisting of 2 participants (agent & customer) • Basic video call with will have x2 audio/video streams: 4.22 MB • Video call with screen sharing will have x2 audio/video streams + x1 video stream: 6.02 MB

Please note that these are rough estimations. Variations in background, movement, lighting and fluctuations in sound can affect the final size slightly.

Dynamic audio fallback: Minimize poor calls and improve customer experience by automatically falling back to audio in response to poor network conditions or low bandwidth situations e.g a customer is on the move and is using 3g. The video stream may be re-established if network conditions improve.

AI Background removal : Real time background masking with the ability to replace with a static image or blurring.

Recordings playback service: Ability to compose and playback previously ​ recorded audio/video sources with searching and filtering capabilities.

Co browse: Ability to see what your customer sees, and even request remote control to troubleshoot your apps remotely.

Mobile SDK: Native mobile SDKs will allow customers to embed auvious ​ features in their apps.

Mobile App: Mobile apps will allow customers to use auvious features on their ​ ​ (agents will be able to receive video calls from customers on their mobile device).

Snapshot Tool: The ability to take a screenshot of the customers video stream ​ for further processing. The agent will have the ability to save the output in their preferable storage provider (S3 Amazon, Google etc.)