Side Effect of AI Overdose – a Positive and Healthy Outcome!

Artificial Intelligence is in everyone’s agenda now. Investment in AI research by technology companies and institutions is growing exponentially. It is rapidly becoming a mainstream technology that is helping transform and empower us in many ways. Success is motivating researchers to explore new ways to tame artificial general intelligence. A recent study (AI Index 2018 Annual Report) by leading academicians from Stanford, Harvard and MIT suggests that research papers published on AI continue to outpace that of computer science and has shown 8 fold growth in the past decade. The same report states interest in rewards based reinforcement learning is overtaking data intensive deep learning neural networks in the recent years. The IEEE publication recently reported how a new application of AI based on reinforcement learning is helping amputees walk with robotic knee. This is one such example of many medical advancements, powered by AI. The story is same across the industries – banking, healthcare, manufacturing, retail etc.

Going by the predictions, interest in AI will continue to accelerate. CES 2019 and Davos WEF 2019 reiterated the same. While the voice (assistant) had its voice felt in CES, Industry 4.0 morphed into Globalization 4.0 in WEF. CES showcased fierce competition between major technology providers to put voice assistant in every device, and reminded us of AI’s potential to transform experience. Likewise, much of WEFs slated objectives, like digital literacy, inclusive growth etc. can be achieved by AI powered transformation.

In enterprises, data and analytics are in executive conversation, and there is imagination all around on how to extract business value from these technologies. There is information deluge on AI. Use cases, success stories and doomsday advocacies are being received through multiple channels, sometimes curated by AI systems themselves. There are even chances that these systems might prefer their own race, the AI race! There are several articles written by eminent personalities on AI bias and ethics, and is not a discussion topic for this article.

What’s next? When it comes AI driven transformation, sky is the limit. But the story is not the same for all companies. In startups, the so called cloud native companies, it is about sustaining the pace and mindset. In established companies with years of history and technical debt, it is not an easy ride. In these companies, whether it is IT infrastructure, applications, security and skills - ground reality is often not a good match to the artificial imagination. But there is a silver lining though. The AI overdose created enormous visibility across organizations, and is helping to advance transformation messaging across the enterprise.

For a successful AI use case, several building blocks should come together.

• Data collection – IoT plays a major role in it
• Data aggregation – connectivity, storage and data lake solutions
• Data analytics – ML, NLP, cloud and edge solutions
• Data sharing – cybersecurity controls

What is the way forward? Approach should be simple and iterative. Assuming that companies have a working arrangement to share and brainstorm ideas – like digital center of excellence or organizational framework for innovation and continues improvement, pick a use case that gives maximum business value to the organization. It could be implementing predictability of operations, improve a specific product delivery lead time or enhance user experience of the product. Now assess the environment you have to produce the outcome against the building blocks above.

Yes it all starts with an assessment of what you have. But it should be done quickly and effectively without getting into the analysis – paralysis mode. It should cover both hardware and software assets, and their dependencies, with the aid of innumerable assessment tools in the market. Often marketed as cloud fitment analysis tools, these tools gives a clear status of where you are even if you are not inclined towards cloud yet. The outcome of this assessment should create IT modernization road-map.

Understand the data. What data you have, where it is stored and how it is accessed. From what you understood, formulate what needs to be done to make the use case work. The solution could trigger multiple inputs to IT modernization roadmap – connectivity, storage, encryption, localization, segmentation.

Data – what you need more? From what you have, trace the missing elements. What it takes to collect that data? Another input to IT modernization road map – IoT sensors etc.

Analyze the data – once the data is understood, aggregated and made accessible, analyze the data. Run through machine learning systems, and establish a proof of concept to validate the assumptions. If it is success, well scale it or celebrate failure and move forward to next one.

It is not only about AI itself, but on how it triggers organizations to look into the areas that lacked attention for years. It is also not true to state that before AI assumed center stage, nothing was done. Those implementations were not holistic. It was mostly due to business needing the agility that IT was not able provide – most of the early cloud implementations fall into this category, there was a budget to be spent before the quarter or year end, or launch of new product or service that required new systems. This is even more visible in cybersecurity implementations. Security controls are often applied as an afterthought and as Band-Aid, and mostly companies struggled to establish a good governance mechanism to reduce risk of exposure.

AI is changing the rule of the game. It is forcing organizations to think holistically. Just by gaining visibility, companies can channel the investment towards future goals and establish KPIs. Companies will be able to streamline business processes, rationalize their application footprint, retire unwanted applications and the associated infrastructure, making the IT footprint lean and agile. Companies might be able to even fund the AI use case from the savings.

By looking at holistically, companies should be able to identify what skills they possess, who needs to upskill, and who needs job rotation. It can be a huge morale booster to employees. Some will get to work on new areas and some will automate boring tasks and move up the ladder to take on new challenges.

Cybersecurity is another big winner here. Companies should now be able to identify risk, assess and formulate strategies to control the risk. By being part of the holistic assessment and use case development process, security engineers can analyze the building blocks and implement integrated tools, controls and governance, especially when attack surface is expanding due to IoT and cloud. By gaining visibility, companies should be able to establish security controls throughout the life cycle of the data, and adhere to privacy regulations. Having an integrated, holistic and automated cybersecurity framework is a must for digital enterprises.

In conclusion, it is always one step at a time, and AI is prodding companies to take that one step, taking stock. The outcome of that first step would be a modernized and healthy IT platform where future innovations can materialize seamlessly. Having this platform will change the narrative from “I can’t do” to “what can I do more?”