If the future of business is AI, then 2018 will be the year when that promise has to be fulfilled. Research firm Forrester has famously predicted that the honeymoon for organizations trotting AI capabilities is over. It’s time for companies to put AI to work. Dubbed as the year of reckoning — 2018 will be the year when AI will have to make progress in real-world applications. With 2018 rapidly approaching, AI is definitely on the minds of many business leaders. Analysts emphasize senior management will have to draw a line between the hype and real use cases and drive the impact home by realizing value from this technology.
Pockets of Success
Even though 2017 hasn’t exactly been a dull year for AI on the world stage, what with several pockets of success, from groundbreaking research in Deep Learning, M&As and acqui-hiring led by the Big 3 (Google, Microsoft, Amazon), 2018 should ideally outdo 2017 in terms of developments. While it may not be the year when bots get to work alongside, it sure would be a year of rapid development in AI, with some worthy headlines along the way. Also, the year 2017 saw a spike in turnkey AI and machine learning solutions such as Gluon from Microsoft & Amazon, IBM’s Watson APIs and Microsoft Azure ML Studio among other tools debuted by IT giants that can potentially speed up AI implementations.
According to a recent Vanson Bourne study State Of Artifical Intelligence For Enterprises commissioned by Teradata, 80% of enterprises already have some form of AI (machine learning, deep learning) in production today. The findings also indicate enterprises are aware of significant barriers to adoption and are looking to strategize against those issues by creating a new C-suite position — Chief AI Officer (CAIO).
So Far, Narrow AI Rules The Day
Narrow AI, where machine-learning solutions target specific tasks, is the order of the day. Also known as Weak AI, Narrow AI carries out specific tasks brilliantly through a combination of advanced algorithms. Today, most of the breakthroughs are seen in the domain of Narrow AI — such as self-driving cars, AlphaGo, image recognition application, latest iPhoneX’s face recognition app, or virtual agents deployed to improve customer service, manage emails or schedule meetings. We are still far from achieving truly Intelligent AI – that can autonomously acquire learning from a single example, understands language, can reason contextually and boasts of EQ.
Even though Indian enterprises have been quick to adopt AI adoption in the last five years, adoption is set to rise, a report by Intel India indicates. The heartening news is that AI adoption in India is on the rise with 68.6 per cent Indian organizations planning to deploy it before 2020. Meanwhile, 71 per cent of organizations are looking at increased process automation as a key benefit that could further drive spends on this technology by 2020. AI has penetrated all sectors, particularly IT that saw an enterprise-wide adoption of RPAs earlier this year.
We list down a few ways for businesses to make AI work in enterprise settings:
1.Increased focus on business-led use cases: According to McKinsey’s Simon London, today every industry can make use of AI technology what with use cases right across the value chain and across the operations of most companies. But even though there are a lot of applications and portfolio-of-initiatives across every industry, it is important for business leaders to understand this technology, find the potential area of application and understand how it can be leveraged as a competitive weapon. This will be the first step before diving into AI-enabled processes and AI-enabled business. The turnaround a business case is fast and Mckinsey’s Peter Breuer believes starting off with simpler use cases can help tech leaders prepare for more advanced uses cases in the future.
2.Tackle the good data challenge: A recent report threw light on the data deluge hype. It highlighted how despite the wealth of data flowing in—about 2.5 quintillion bytes a day, majority of the data is not labeled or structured, which renders it unusable for supervised learning tasks. As emphasized by Daniel Shapiro, a machine learning enthusiast, when you have labeled data machine learning works. But what about cases when you don’t have any labeled data? So, researchers devised two strategies – transfer learning and unsupervised learning to tackle the lack of labeled data challenge. So, how can enterprises meet the labeled data challenge? There are two options a) crowdsource labeled data; b) outsource the task of labeling data; c) there are a slew of startups that have devised machine learning models to cleanse data. Today, it has become crucial to get access to well-labeled, statistically representative data sets and maintain it over time.
3.User Experience in Customer-Facing AI Applications Should Take Centrestage: Now that AI has made its way to the frontlines with a slew of customer-facing applications, it is the right time for businesses to think about how to weave AI into a great user experience. In fact, UX should be front and centre more than ever given how bots are going to become a big part of enterprise settings. And a great UX goes beyond design thinking into providing a more unified customer experience.
4. Getting an ROI from AI: While AI has definitely lived up to its promise in 2017 and the investment is only expected to increase over the coming years, a recent survey pointed out how 64 per cent of IT decision makers expect to see a return on investment from AI implementation within two years. It is a question that vexes VCs and the C-suite alike. Since the application and level of maturity of AI technologies can vary a lot, it is difficult for decision-makers and senior management to generalize how to measure the ROI on their investment. A recent report indicates that the C-suite should also sport a VC-type mindset and brace themselves for both failure and success.
5. Time to appoint Chief AI Officer to strategize AI adoption: According to a recent survey b Teradata that sampled 260 large organizations globally, enterprises today are seeing AI as a strategic priority that will help them outpace the competition in their respective industries. Atif Kureishy, Vice President, Emerging Practices at Think Big Analytics, a Teradata company. And to leverage the full potential of this ground-breaking technology and gain maximum ROI, businesses will need to revamp their core strategies. “So AI has an embedded role from the data center to the boardroom,” says Atif Kureishy, Vice President, Emerging Practices at Think Big Analytics, a Teradata company. This can be fulfilled by creating a new C-suite position — the Chief AI Officer.
Try deep learning using MATLAB