Even though the euphoria around AI is high, talk about AI adoption within enterprises is usually met with a confused gaze among business leaders. And rightly so — AI use cases, though very well storied, have not really well concretised for enterprises. They still appear very consumer focused.
Also, adoption for AI, just like any new technology, seems so confusing at first to enterprise leaders. Starting out, organisations often lack a clear business vision and data strategy and have to grapple with a slew of critical questions, Who would lead the baton, what are all the new tools one would have to acquire, how much budget should be set aside, competencies, how to acquire those skills, how to build a data-product mindset and the big challenge all enterprise leaders have to face — will it really have any impact on the business? A lot of these question eventually lead to a state of inaction in adoption for AI.
Very few organisations have yet realised that AI adoption is one of the easiest ones, given that most of them have an analytics unit in place (at least the larger organisations). AI for enterprises is not very different from what analytics has been doing. Well-entrenched analytics units can serve as a good starting point for organisations looking to build AI capabilities.
Many organisations that stack up high in the analytics maturity index can leap-frog their AI initiatives by embedding an AI-driven mindset.
Analytics Unit Can Double Up As AI Unit
We tell you how to lay the groundwork for extending the existing analytics unit into an AI unit. AI is an offshoot of analytics in more ways than one and enterprise leaders can maximise the value by broadening the scope of expectations from their current projects.
a) Where are the similarities: Well, firstly, both Analytics & AI feed over data, a lot of enterprise data. The whole premise of analytics adoption was that the enterprises are generating a goldmine of data that can be used to generate insights and eventually benefit organizations. The same data (and I really mean a lot of that same data) can be used for AI as well.
b) Both Analytics & AI have algorithms at the core of what they do: The idea has been to take the data and run algorithms on top of that for the desired outcome. The range of algorithms have evolved over time and might differ for Analytics & AI but the basic fundamentals of stats, data cleaning, memory management remain the same.
So, it’s very clear that analytics units have been doing what AI might be wanting to do for some time. And with this much apparent that analytics units should outgrow to AI use cases.
But then there has to be a difference, right?
How Do They Differ?
Analytics & AI though similar in how they function, differ in some aspects. Primarily, Analytics & AI differ in terms of their use cases. Analytics is an enterprise technology, that is used to fine-tune management strategy or help in better decision making. AI usage is different – can we have machines think/act in some shape and form like humans. So, computer vision is a big application of AI and is now being used by enterprises for a myriad of use cases like facial recognition and object detection.
Case in point, a manufacturing company scans its floor workers whether they are wearing safety gears, thus reducing operations risk. Or a retail company that uses Computer Vision to identify shoplifters thus reducing fraud. It is a strong technology and one with so many solid use cases for enterprises.
Question is, do we need a separate unit for AI or can existing analytics unit take that baton.
There are a lot of other differences too:
- AI is a lot about neural networks as algorithms, analytics has been a combination of them — classification, clustering, regression or just hypothesis testing
- As a result, AI requires much more computational bandwidth
- The mindset of creating insights or predicting outcomes for the enterprise is to a large extent missing in AI
How Analytics Unit Can Function As An AI Unit
Given all the chatter around how AI is different from analytics, the existing analytics unit should be the first combat zone that organizations need to win for a successful AI journey. The similarities are way too high for this to be ignored.
- Train existing data scientists on neural networks
- Create a data-product mindset within an analytics function
- Have a dedicated small team with the larger group that oversees AI implementation
- Chalk out AI use cases that can be quickly deployed given existing capabilities
When comparing resources to resources, it's much more efficient to transform existing analytics units into AI ones. There is so much overlap in capabilities that chalking out a separate AI unit would take up efforts of humongous scale for an enterprise. And a lot of organisations have been doing exactly this.
Most of the time, the existing analytics unit take up AI projects on their own. This also gives them a larger mandate, motivates existing data scientists. But more so it provides a sense of working on a tech that’s the next big thing.
Register for our upcoming events:
- WEBINAR: HOW TO BEGIN A CAREER IN DATA SCIENCE | 24th Oct
- Machine Learning Developers Summit 2020: 22-23rd Jan, Bangalore | 30-31st Jan, Hyderabad
Enjoyed this story? Join our Telegram group. And be part of an engaging community.
Our annual ranking of Artificial Intelligence Programs in India for 2019 is out. Check here.
Provide your comments below
What's Your Reaction?
Bhasker is a Data Science evangelist and practitioner with proven record of thought leadership and incubating analytics practices for various organizations. With over 16 years of experience in the area of Business Analytics, he is well recognized as an expert within the industry. Earlier, Bhasker worked as Vice President at Goldman Sachs. He is B.Tech from Indian Institute of Technology, Varanasi and MBA from Indian Institute of Management, Lucknow.