AI and RPA Prep - Part 3 ... Data

Okay, Let us be frank here. I really had to rename the original part 3 as part 4 and do this one. When I discussed one day on Omni Channel and Digital profile of a customer. I had to reiterate rather loudly on what is involved in terms of data and security here. Lot of times, it is often under estimated but should not be. Data is the backbone of anything we do now-a-days. If your data base crashes. You can as well close the shop and go home.

Each of us from morning till night and even when sleeping are creating data and more data. It is called digital footprint. It got derived from the word carbon footprint of how much each human is creating on the Earth. We make calls, we browse internet, we shop online, or use cards, we are paper-less most of our life now. All that information is out there revolving around one person that is you. It is your digital footprint. For example I alone can create some gigabytes of data in terms of digital footprint every day just by my normal transactions and browsing activity etc., then imagine a family's digital footprint size? imagine the customer base digital footprint that is getting created? So if one has to have a 360 degree view of a customer and have real time data coming in from various sources that have that data to do analytics for various purposes, imagine the amount of data we have to deal with. We need to be able to consume the data, cleanse the data because we do not need all the data but only specific ones that will help us in our goal, then keep the data some place for our analytics to access and derive the results. And to add on top of it, for analytical models to give reliable results, the amount of training data that is required runs into millions of records or so.

One of the question I always ask customers is "Are you ready - data wise?". I even wrote an earlier article "Be prepared?" on the same when bots have to be implemented. Without prepping the data, it would be futile building the blocks that have to be supported. Lot of social data can be bought from data providers (basically data vendors) for a price. But there is lot of legality involved when it comes to social media data. Data involving sensors also has some legality involved. For example, if we are talking about data from cars like telematics then there was a question on who owns that data. Does the car company own the data captured by sensors in the car? Or does the car owner own the data? Or can insurance company access it and use that data for claim processing? Where is the privacy of customer here? Similarly for sensors in smart homes or wearables. If insurance company can access the information without your permission that you had burgers with extra load of cheese thus would not cover the hospital claim then what would happen?

So, data and digital footprint is a huge thing when it comes to RPA and AI. Because of following reasons

• Data is huge so we need to have data lakes and related bigdata software components like Hadoop etc.,
• Data analytics or also known as Data science knowledge and expertise to cleanse, manipulate the data that is obtained.
• Data or messaging layer to format, encrypt and decrypt the data flowing in and out. There are international laws limiting certain data crossing the borders.
• Analytical models to run on the data but first should have training sample sets without the sensitive information
• Data is always not straight forward data as text but it can be voice which needs to be converted to text. For example, call recordings or voice bot conversation history etc.,
• Backups and Archiving becomes important since there has to be some kind of logical archiving and retrieval of data logic in place.

Apart from this, data provider agreements and contracts should be in place legally protecting the accessibility.

Having data alone is not enough, This data has to be accessible to front end channels and other consumer systems. So services for each of the entities is required. Phew !! I feel exhausted just writing this one.

Next part we will discuss security involved.

AI and RPA - Part 1

AI and RPA - Part 2?