Tuesday 14 November 2017

Analyze, Assess, Adapt: How to Implement AI in a Business Environment

For years, we’ve heard that AI will change the way we work and live. Yet for all its promise, few companies outside the tech space are using the technology in truly meaningful ways.


But that’s about to change.


We’re now seeing AI used primarily in two ways: to automate work that would otherwise be performed by humans or to augment the work of humans.


Everything that can be automated will be automated, and unless you’re like the auditor I met recently who appreciates repetitive tasks for their meditational qualities, automation is a good thing for business.


Repetitive and mundane (i.e., boring) tasks like data entry and even mowing the lawn have already been automated to a great extent. And thanks to continuous advancement in the fields of machine learning and robotics, machines are increasingly able to take on tasks that require more robust cognitive abilities — like driving.


Moreover, machine learning technology is now enabling AI to augment human decision-making. AI-powered systems are being implemented to help people manage energy use, analyze X-rays for cancer detection, and even make investment decisions.


Undoubtedly, the technology is still in a relatively early stage. While some jobs have been completely eliminated by automation, there aren’t many business processes that can be performed entirely by machines. Instead, AI is replacing portions of work that would otherwise require a human touch, allowing (human) employees to be more productive in other areas. But there are still hurdles involved in AI adoption.


You’re Going to Need More Data


Aside from the obvious concern that someone else will beat you to the punch in implementing AI, the technology poses a number of challenges for executives.


It starts with your data. Before thinking about investing in the technology, consider how you gather data (including the privacy issues that data collection could present), the quality of that data (whether it’s user-generated or generated automatically by bots or sensors), and how you structure it.


Your data must be accessible in order for AI to analyze it, and it also needs to be reliable. If you have accessible, high-quality data, that’s a good start. But realize that you’re going to need a lot of it if you want an AI-powered machine to be a truly useful asset. The more you have, the more powerful your algorithms become.


Likewise, you’ll have to be aware of the legal and ethical issues that accompany the use of AI, as well as inherent biases that may influence its output. For example, the diversity problem plaguing the tech industry is well documented, and the fact that most developers are young, white males has potentially (let’s be honest, probably) created biases in the algorithms powering many of the AI systems being implemented today.


So you don’t need to be afraid of super smart AI running the world; you have to be afraid of stupid (or biased) AI running the world.


Analyze, Assess, Adapt


Obstacles aside, AI is here to stay. If you don’t want to be left behind (which you don’t, right?), I recommend using the “analyze, assess, adapt” approach to understanding technologies and adopting AI applications:


1. Analyze the technology.


Spend time studying the different AI applications that are available today or will be available in the near future. In general, there are four applications you should consider: AI for understanding unstructured data like text, images, audio, or video; AI that allows your system to learn from its interactions with the environment; AI that can understand a user and help you create personalized user experiences; and AI that can interact with a user, like chatbots or virtual assistants.


Focus on the one or two that are most relevant to your company, and try to understand roughly when these could be adopted given current technological constraints.


2. Assess the business implications.


Consider how the above applications could affect your industry, customers, and company.


From an industry perspective, ask yourself what your competitors are doing. Are they moving in new directions? Attacking your customers in a new way? Do you see signs of partners moving into your arena? Who owns the data that’s being collected in the value chain? Will your part of the value chain grow or diminish?


Then, think about your customers. Are they adopting AI? Are you seeing a demand for AI-powered products and services? Are there opportunities to create new products or services that cater to customer needs?


Finally, consider your own business objectives. Could you automate certain functions to cut costs and make your business more efficient? Would AI allow you to make better use of your data? (Answer: Yes. Always.) Asking the right questions will help you make the right decision.


3. Adapt to a new reality.


Based on your technology analysis and business assessment, determine what adjustments need to be made in the face of new uncertainties and risks. Think about how your leadership might have to change. Think about your place in your industry. If you’re a large player, you might be able to use AI to shape your space. If not, ask yourself whether AI technology would allow you to be more equipped to adapt to what the shapers are doing.


Understand what’s out there, when it’s happening, how it will change your business in different ways, and how you should adapt to this new reality.


While people often think of automation as a threat, for business leaders, it opens up countless opportunities. AI will continue growing and getting smarter whether you’re ready for it or not. Don’t be afraid of AI running the world — be afraid of AI running the world without you.



Source: B2C

No comments:

Post a Comment