Accelerating adoption of automation in Enterprises - Full Keynote - Botminds Fusion02 December 2022
An exciting keynote session by Aditee Rele, ISV Sales, Microsoft, where we will hear about the challenges enterprises still face in adoption of automation and in harnessing unstructured data hidden in their documents and processes. Aditi also talks about the way forward for enterprises to adopt AI-powered platforms to unlock value from the data.
Note: This transcript has been edited for clarity.
The need to infuse technology in every business process, in every business function, in every business application has never really been so urgent. So, this is a fantastic depiction of how we see data from one of the top four analyst firms. In a typical enterprise, over 80% of data is unknown to us or is unstructured.
A significant amount of data that is lying in unstructured formats such as Word documents, Excel documents, PDF images and so on. There, harnessing the data is extremely difficult. Of the 80% of the data that is unstructured today, only 3% of the data has been harnessed. So, that is a huge opportunity for us as technology providers, and all partners of Botminds and Microsoft. But it's also an opportunity for our customers to figure out if that data is going to help them develop better business insights. So, how do we make sense of all this data?
If you take an example in healthcare, today a physician’s notes are still handwritten. So, if you're building out intelligence to figure out how should a patient's journey look like, what should be predictive analytics on that patient to see what the patient is going to suffer from, how do you make sense of that? In the insurance industry, when your car is hit on the road, somebody will come with a paper or these days, and put data onto the app. But the recording of the incident still happens on video and that is not harnessed.
If you look at manufacturing as an industry, the quality assurance still happens on the shop floor on paper. How does that get translated into systems that can then do predictive maintenance algorithms on those systems? So, every industry, every business process, has an abundance of unstructured content. A typical company processes, say 5000 or 3000 invoices a month, each invoice takes about five minutes to process by humans. The invoices are (in) different (formats) because they're coming in from different vendors, different vendor systems - that's 250 hours. The objective is to figure out how to repurpose these 250 hours into something more intelligent than what the worker can do and use process automation to fix this problem.
There are other problems to worry about. For example, in India and in many developing geographies, there are continuously changing compliance requirements. For example, a new GST law has come in. All your invoices would need to have the company's GSTN number. Now, that's one of the problems - regulatory compliance.
The other problem I like to think about is (analytics). We are building a lot of systems that give us the ability to do predictive analytics on the data. Bring in data, understand it, describe it, diagnose it - decide what actions you should take going forward.
But that data does not include any unstructured data. So, that's just 20% of your enterprise. Are you truly able to make business decisions to drive your growth with just 20% of your data, which is structured? What if you had the ability to bring in the remaining 80% of which only 3% has entered, and do the same to analyze that data and then build business insights? That would be like a true reflection of what data as a new currency really looks like.
The third problem I feel is the ability to search this unstructured content, whether it is images, videos, or even documents - many of these are simply unsearchable. I mean, you would not be able to really find the data that you're looking for if you just look at plain videos and images today. So, with today's technology, we can reduce time to market, increase efficiencies, reduce costs, and importantly, future-proof our growth.
Over the years, we've seen significant changes to technology. Gokul talked in the beginning about all the technologies we've been through - OCR, ICR, RPA, etc. But IDP is still an evolving technology.
Today's challenge is a little different.
Not every organization may have the right skills to build out these AI solutions that we're talking about. And not every organization should even invest in those skills per se. That's where the power of our partner ecosystem comes together. So, Botminds and Microsoft work together to help our customers solve all these issues using Azure and AI capabilities.
Our objective is simple - we want to make this unstructured data available to harness business insights. Our CEO, Sathya (Nadella) said this beautifully a couple of weeks ago. He said, “In an inflationary economy, technology is the only deflationary force.” And that's exactly what we're trying to do.