• News
  • Columns
  • Interviews
  • BW Communities
  • Events
  • BW TV
  • Subscribe to Print
  • Editorial Calendar 19-20
BW Businessworld

Any Form Of Growth Across Industries Is Dependent On Its Ability To Adapt To Digital Innovation - Shveta Raina, AVIZVA

With a niche client base in healthcare, Telecom, Logistic, etc, Avizva caters to varied areas of expertise i.e. Integration platform, cloud infrastructure, UI-driven applications, QA Automation, AI &ML, Mobile applications, BI & Reporting, etc.

Photo Credit :


With a niche client base in healthcare, Telecom, Logistic, etc, Avizva caters to varied areas of expertise i.e. Integration platform, cloud infrastructure, UI-driven applications, QA Automation, AI &ML, Mobile applications, BI & Reporting, etc. In conversation with Siddharth Shankar from BWBusinessworld, Ms. Shveta Raina, Associate VP - Marketing at AVIZVA talks about AI Chatbots, Technology usages in HealthCare, AI and ML helping in health insurance and future trends 


Artificial Intelligence based Chatbots, or conversational AI, how is it working in the Digital Healthcare Space?

Conversational AI is an absolute specific use case that are specifically coded not just to provide information but also to have an appropriate tone and take into account all conversational nuances. Its cumulative power comes from its capability to run through personalized interactions with a large number of individuals. Such a use case was layered actively while COVID-19 pandemic was flooding the globe, which was demonstrated MyGov Corona Helpdesk, a chatbot integrated with WhatsApp to answer a range of queries, by design a well-crafted architecture of such conversational assistance can reduce the need for human intervention in mild diagnosis and information sharing tasks by 80%, this enables experts to cater to severe conditions, backlashes and problems with an absolute focus.

Deploying these intelligent bots on a sector having such a large scale of user demographics as in Healthcare have entitled benefits:

Self help diagnosis- It’s the basic use case for a conversational engine deployment, users now, can have there first directive without any delay from the symptoms the AI procure from them during dialogs which in turn helps bot provide holistic information on the disease in natural language.

Automated Appointment Booking- When a patient’s case is categorised as emergency by the bot, it can book a bed just right there and then, scheduled appointments and reminders are helpful likewise

Personalisation - Can put orders to re-stock medication and groceries, monitor health and fitness, such as fitness bot Ziggy

Reduce Cost & Increased Efficiency- Cater to users 24/7, reducing need for human resources and operating costs significantly, the algorithms in artificial intelligence assisted conversational engine is organically trained which improves performance automatically thus not needing an human assistive enhancement for mildly required intent context boot ups.

Automated Billing & Registration

Automated Prescription Management

Patient engagement & follow ups

Claim processing & predictive accumulator’s adjudication

How is the integration of technologies like AI and ML helping in solving healthcare insurance challenges.

Technologies like AI & ML can be deeply integrated at every step of the way because Healthcare Insurance deals with heavy & sensitive data exchange between multiple stakeholders. Here are some of the interesting use cases that we have successfully implemented with certain scenarios and challenges encountered by our clients in their day-to-day business operations.

  • Images and documents can easily be tricked to inset manipulated information. The crucial and required business acceptable artifacts are not safe either from these manipulations. Especially when these artifacts are to be dealt with, in large numbers in a day, it is essential to label manipulations to that of a real information, that’s where machine learning and artificial intelligence clubs to help us keeping everything in check using a core from CNNs as base.
  • To get an insight on their plan & benefits information, Consumers have to sift through screens and tons of information being rendered to their eyes, in the form of csv’s, flat files and pdf(s). We have been able to give a turn to all those information access design that too in human classical way of information sharing - conversations. With potential in devices such as alexa, google home and chatbots, today’s IVR,  designing and coding a vui helpful agent once and deploying through all these platforms, keeps self-help information retrieval in check and delightful.
  • Tasks are heavy on individuals when it has to be reciprocated hours on hours in loop throughout the day. Given the business rules, it is aesthetically and contextually important that only correct and coherent information is being inserted in systems. But what if the source of this information is a pdf document, with layers of data, words and tables? How would a human brain comprehend and pick the pieces to be inserted in a dbms? Even tougher when this has to be done for thousands of documents in a day. Our piece of architecture solves this by relying on Q-learning, another method from artificial intelligence feature stack to automate this set of comprehensive tasks.
  • Sometimes there are problems that need an acute angle of reasoning to follow up with. Two systems with different business layers having similar user data metric - when data feed from these are fed to a meta system, there are definite chances of dublications and repetitive data with discrepancies on some attributes. It usually is a case when no unique identifier(s) are possible to draft, such a case was solved using expectation maximization algorithm as base on spark clusters.

What are your thoughts on AI deployment?

 Development and Deployment of an artificial intelligence model is a complex workflow, an output of a model development process is only a set of files, some scripts, source code, and takes a very small fraction of the whole superset of things that are absolutely necessary to deploy and maintain the model. The total set of deployable and data injection pipelines are:

  • Configurations
  • Collected data
  • Verified data
  • ML code
  • Extracted features set from data
  • Process management pipeline
  • Serving Infrastructure
  • Monitoring and Feedback
  • Analysis tool set

We keep two pipelines of any ai solution architecture, one that drives value and one which drive new model and innovation, the first pipeline takes care of maneuvered data which produces the output, analytics and deliver value to the business. In the second pipeline, there is a process layered to create new model, innovation pipeline is also tested and deployed into the value pipeline. Regardless AWS have now achieved a good traction on reducing the ai/ml pipeline efforts.

What are the sectors you are catering to and what are your expansion plans?

Since it’s inception in 2011, AVIZVA has built numerous custom solutions, each of its products employ Technology that simplifies & expedites care, Design that uplifts the synergy between consumers & service, Strategy that keeps organisations ahead of the curve. 

In an ever-changing healthcare landscape, each business enterprise has its unique vision and unique challenges. We have partnered with many healthcare leaders to understand their challenges & develop products customised to their needs. 

In the building years of AVIZVA, it has had the opportunity of working with many govt and non-govt establishments in aviation, mobile and logistics sector. Through our experience, we have realised that business challenges are of very similar nature across various sectors. With our experience & proven success of working with various sectors, we are certain that our innovations can be tailored to any demanding sector. 

What are the big factors accelerating technological innovation in the healthcare sector

Availability of data in correlation with absolute and precise structure is the key factor that is driving an intrinsic acceleration in the healthcare sector. The emerging tech stack deepens the layers of personalization; it culminates to the growth and innovative problem tackling. Such a use case can be realized by integration platforms. Further fueling big data with modern computing and algorithms can stream metric to cognitive switches on nodes, this also adds tremendous effect by which computing devices develop an ability to learn, giving an extra edge over decisive tasks, such as detecting a claim to be fraud or not, predicting adjudication results prior to claim processing.

What tech trends do you foresee gaining traction in the future in your sector?

Any form of growth across industries is dependent on its ability to adapt to digital innovation being applied to businesses. In the backdrop of COVID, Healthcare specifically was the one to quicken its pace & adapt to digitalisation in no time. From Data Analytics to Telemedicine to Self Help portals to AI enabled conversations - the world at large experienced, consumed & benefitted from the innovations in Health-Tech space.

There are many innovations that are underway to revolutionise the healthcare industry. The list is by no means exhaustive and it will be an injustice to just name a few. Sticking to our immediate area of expertise i.e Healthcare Insurance - the two most notable trends that continue to make a futuristic headway are:

Healthcare Data Analytics especially predictive analytics. The effective employment of predictive analytics can offer significant improvement in access to care, reduction in costs, streamlining of healthcare operations, efficacy of decision making and policy making. Predictive analytics has great potential to transform healthcare. 

The other is Artificial Intelligence that has massively exemplified its impact by improving the pandemic response and recovery. With many wins to its credit already, we believe it’s just the beginning of endless & unimaginable possibilities that AI especially conversational AI, can bring to consumer focussed industries. The outbreak of COVID-19, also gave us a compelling use case highlighting the unimaginable impact of conversational AI.

Tags assigned to this article:
AVIZVA artificial intelligence machine learning