Notice: Function register_block_script_handle was called incorrectly. The asset file (/home/myneaexs/public_html/wp-content/plugins/seo-by-rank-math/includes/modules/schema/blocks/faq/assets/js/index.asset.php) for the "editorScript" defined in "rank-math/faq-block" block definition is missing. Please see Debugging in WordPress for more information. (This message was added in version 5.5.0.) in /home/myneaexs/public_html/wp-includes/functions.php on line 6031

Notice: Function register_block_script_handle was called incorrectly. The asset file (/home/myneaexs/public_html/wp-content/plugins/seo-by-rank-math/includes/modules/schema/blocks/howto/assets/js/index.asset.php) for the "editorScript" defined in "rank-math/howto-block" block definition is missing. Please see Debugging in WordPress for more information. (This message was added in version 5.5.0.) in /home/myneaexs/public_html/wp-includes/functions.php on line 6031

Notice: Function register_block_script_handle was called incorrectly. The asset file (/home/myneaexs/public_html/wp-content/plugins/seo-by-rank-math/includes/modules/schema/blocks/schema/assets/js/index.asset.php) for the "editorScript" defined in "rank-math/rich-snippet" block definition is missing. Please see Debugging in WordPress for more information. (This message was added in version 5.5.0.) in /home/myneaexs/public_html/wp-includes/functions.php on line 6031
Computer Vision to help in WFH challenges in Post Covid World
Tue. Dec 3rd, 2024

Computer vision (sometimes called machine vision) is one of the most exciting applications of artificial intelligence. Algorithms that can understand images– pictures and moving video – are a key technological foundation behind many innovations. From autonomous, self-driving vehicles to smart industrial machinery and even the filters on your phone that make the pictures you upload to social media look more pretty.

In recent years attempts have been made to emulate human senses through computing. There is quite a bit of success in vision, speech and listening areas but very limited in the savory and olfactory areas.

The following treatise will deal with the usage of Computer vision to solve post-Covid human productivity assurance solutions.

We all know that “Work from home” is not new, and with the changing situation, it has come to the forefront with new challenges.

WFH is not just enabling people to work remotely but also allowing business to operations to be managed remotely.

Initially, the work from home was seen as a stop-gap measure. Still, after the emergence of multiple variants of Covid-19, we now need to provide employee options to get 100 % work from home, work from home or remote and hybrid working models.

As we are thrust into a Work from home solution way of working post-Covid, different kinds of challenges propped for senior management of organizations to ensure both employee safety and productivity. In settings where an employee’s presence on the desk is mandatory to serve the business process, human surveillance is insufficient and inefficient.

It is imperative to have an automated monitoring tool in place to track productivity and ensure that endpoints are secured & data is protected.

Tracking tools must be ready to provide data at all models; hence a single tool with multiple work models is the need of the hour.
A picture containing text, businesscard

Description automatically generated

Some of the things predominantly done from the office will now have to be managed remotely

•       Floor Walking

•       PC health

•       Compliance audit

•       Login time and loss of productive hours 

Tech Mahindra realized the need to develop a computer vision-powered monitoring solution that generates data to perform desktop utilization analytics.

Enter KORNEA, an application resident on the end user’s laptop/desktop to monitor only their login activity and applications aided with Computer Vision algorithms.

Need of hour:

Figure 2 Key Features

Under the hood:

Post Installation, two aspects enable KORNEA to perform its core function. 

Registration:

  1. Face scanning will be done using deep neural network & Caffe model
  2. Face-scanned images gets processed using ML, and extract the encodings out of them
  3. A pickle file is created by merging all the encodings extracted from the face images

Authorization:

  1. A new Face image gets compared against the pickle file, and a confidence value in face similarities is calculated.
  2. This confidence % value is compared with a predefined threshold match % to finalize whether the face matches the pickle file and thus authorized.

The solution is configured to comply with security and sensitive data capture.

During the tool’s operation, it provides near real-time reporting, enabling better decision-making for Ops teams.

These statistics reflect employees’ productivity along with hosting machine’s health data and its compliance with respect to current IT landscape.

Thus, this solution can be compelling when integrated with ITSM, WFM, QMS and CCT systems and key stroke level information can be used for Prediction models can be built on historical trend analysis.

With automated and accurate data capture and classification without manual dependency, the operations team can avoid the manual effort required for monitoring, reporting and analysis and replace it with proactive triggers of critical events.

Key Benefits:

• Risk mitigation – The report generates a lot of IT security compliance status along with any risks to be mitigated

• Provides an independent and thorough review of system status, eliminating potential security risks, improving data quality, and increasing system performance

• Serves as a preventive measure and helps in tightening security

• Cost efficiency – Saves resource costs otherwise required for auditing compliance checks manually

• Customer focus – Provides assurance that the compliance obligations and controls set are met as desired

• Cloud Ready – KORNEA Data gathering can be configured to both Cloud and On-Premise as directed by clients. We work with all cloud providers like Azure /GCP, AWS

Employee Privacy is of utmost importance to us. Solutions like KORNEA can be perceived as an intrusive tool to monitor agent or employee activity or hold employees accountable for non-compliant behaviours. For e.g., Checking the Amazon site, since transparency is the key here

Objective of a solution like KORNEA is to improve efficiencies by building trust. It enables the implementors to gain visibility and insight into what data is captured and reported. KORNEA can be configured to classify data to a level of ‘Productive’ and ‘Non-Productive.’

Surely, any granular level details which are defined as not in the scope of the solution, like specific URLs, screen capture etc., can be prevented from being captured or reported.

KORNEA’s implementation always complies with the ethics, codes and standards of the customer organization and consequently ensures we have the trust of our employees. Bolstering the fact that the employees can remain confident that no personal information is captured and shared with anyone, including their managers

In several of KORNEA’s implementation, it has been found that it is also extremely effective for Back-office teams. Operations, especially back office does not have great means to monitor, capture agent efficiency and most of the capture is manual and advisor led. Hence there is always an absence of insights on non-productive work activities

This creates Lack of transparency in employee’s time utilization, lack of knowledge of productive hours and Missing demarcation on application usage.

Using KORNEA across many customers, we now have Insights on time spent on accessed applications, thus visibility on logged-in hours v/s productive hours. 

There is an inbuilt Time track & engagement feature which helps provide advanced insights on the time spent on each activity with the identification of behaviour patterns and tracked outliers. We have been able to achieve up to 7% Uplift in productive hours and an additional 10% improvement in Login hours.

Future is bright

KORNEA’s core engine is a Deep Neural Network which learns from sample images and can match with an incoming image with high accuracy. Our algorithm’s accuracy is more than 99%.

The use case for face match is tremendous, as there are use cases where in we have been asked to provide the solution for sifting through photographs and creating a report on duplicates. 

There are possibilities where Kornea can expose the API for citizen development which can save quite a bit of consulting and development costs for, customer organization.

As you can see the future of Computer Vision is BRIGHT….. 

Author’s Bio –

Diptojeet Mukhopadhyay is Digital Process Automation expert with 23 years of end-to-end automation delivery experience across the globe. He has experience in Customer management, Practice &CoE Management, and Technical Delivery Management etc. Diptojeet comes with deep expertise in DPA Transformation leveraging market-leading tools and technologies. He has also been engaged in multiple consulting studies across the globe for some of the leading brands across North America, UK, the Middle East and Australia. He is also a Togaf 9 certified Enterprise Architect with additional certifications in UiPath, Appian and Pega.

Leave a Reply

Your email address will not be published. Required fields are marked *