When we talk about Artificial Intelligence Strategy (AI Strategy) and Machine Learning (ML), more often than not, it is about the technical aspects involved in developing models and implementing the algorithms. Whatever small talk is there on the business front, it’s either about potential or about the first steps smaller businesses are taking. Any examples of a mature implementation involve large companies. But AI adoption is a challenge for small & medium businesses on many fronts. We are going to discuss a few such problems and the possible first step for such companies.
Table of Contents
The Understanding Problem
To a large extent, most smaller businesses are in the stage of “unconscious incompetence”. We don’t know what we don’t know. There is a minimal idea about what exactly AI can do for the business. The exact mechanics of how to roll out AI/ML projects, how to execute them, and how to extract value out of them. This skill gap prevents small businesses from starting the right projects. Right execution of such projects also becomes a challenge.
The Data Problem
The primary challenge for the adoption of AI in the data. Larger companies have massive data collected over the years. That data is also much cleaner and organized. Availability of the data makes it easier for them to get started on supervised machine learning problems. Smaller businesses have the issue of data sufficiency and cleanliness. AI and ML efforts need extensive data labeling efforts. These efforts can prove to be time-consuming & costly for smaller businesses. Time & cost put tremendous constraints on the kind of AI initiatives small & medium companies can undertake.
The Capabilities Problem
The third challenge is the challenge of affordable skills. One, even bigger organizations have this issue. They do not have a skilled enough workforce at this point. But at least they have more money. Availability of such resources is a constraint for smaller businesses.
Even in larger organizations, the challenges are more or less similar. Here are a few insights from McKinsey.
Should smaller businesses even care about AI Strategy?
These challenges come only if smaller businesses want to embrace AI. But should they even worry about it right now? Is it time to start spending efforts on acquiring knowledge & skills and thinking about rolling out an AI initiative? Is AI going to leave up to its potential, or is it going to be another hype?
From a business perspective, this is a strategic decision. A lot depends on the use case too, where AI can start making an impact. If we go by what the trends suggest, then it’s worth getting started sooner rather than later. ML & AI capabilities is an area where businesses can get a competitive edge all of a sudden if everything goes right. And if a competitor gets an advantage by being the first to put it in place, it can have a catastrophic effect given the power of AI. So yes, this is the right time to get started on utilizing AI when the adoption is still maturing.
What are some quick-start use cases?
One of the most straightforward use cases involves conversational AI agents (aka chatbots) and all kinds of virtual assistants. Conversational intelligence is an area where a certain level of maturity already exists when it comes to the tools. Enough APIs & tools are available, which can help the rapid development of chatbots & virtual assistants. (You can check out chatfuel.io). Businesses can get started on these with little data they have. They can then iteratively build a more capable agent.
Businesses can use chatbots in a variety of ways. e-Commerce businesses can automate the entire flow from browsing the products until buying and payment. Another compelling use case is to automate the hiring process (or at least a significant part of it). The process can include application, online test scheduling & actual tests, and scheduling of an interview.
The travel industry is also using chatbots in a variety of ways. Check out this article Top 3 chatbots that are changing the travel industry to view some examples.
The second area where business can get started is personalization. The most obvious implementation would be the product/content recommendations. Personalization is an area where you can start with little data historical data. You can then keep on increasing the accuracy of the model over time as you gather more and more data.
The third area can be useful if you have much user-generated content. Such content can be in the form of reviews, comments, or feedback. Sentiments analysis is mature enough, and you can get started with relative ease.
Let’s also see a comparison between organizations where digital transformation is at a mature stage versus the organizations where there is still a vast scope.
What will be the future?
The AI technology scene has already started to get commoditized. The tools for building chatbots, as mentioned earlier, are already available. Microsoft Azure ML Studio provides a host of services out of the box with a visual interface. Google AutoML, though currently supports only three functions, makes it possible to dump data without worrying about the algorithms and get the desired output for even an uninitiated. Again, Google provides ML Kit for Firebase, which allows the implementation of areas like text recognition, image recognition, face detection, barcode scanning, landmark recognition, and language identification in Android apps. Other MLaaS platforms too provide similar services.
We will stop here with the following comparison. But more discussions are coming up on the same theme. Stay tuned!
I wish to give as many LIKES as possible to this post, especially for the presentation part
And it is very informative, clear and to-the-point
Thanks so much, Uma. Coming from you, it means a lot.
This is my first read about AI, and I totally understtod your article. Thanks to the simplicity and clarity
Glad you liked it, Swati.
Pingback: Digital Transformation And Its Astonishing Impact On Business Growth
Pingback: What Is Digital Transformation? A Beginner's Guide