Written by Vinit Balani, Senior Specialist – Cloud Services and Software
Just like social distancing went on to become the new normal during and post the pandemic in the physical world, the cloud is slowly becoming (if not already) the new normal in the enterprise world. With remote working and a serious push to digitization, it has become inevitable for enterprises to continue delivering products and services to their end-users. COVID-19 has been a major driver in cloud adoption for many enterprises across the globe.
Worldwide revenues for the artificial intelligence (AI) market, are forecasted to grow 16.4% Y-o-Y in 2021 to $327.5 billion, according to the latest release of the IDC Worldwide Semiannual AI Tracker. By 2024, the market is expected to break the $500 billion mark with a five-year compound annual growth rate (CAGR) of 17.5% and total revenues reaching an impressive $554.3 billion.
If we look at India, India Inc.’s AI spending is expected to grow at a CAGR of 30.8% to touch USD 880.5 million in 2023 as per the IDC report. AI is now being used by enterprises to get a competitive advantage with BFSI and manufacturing verticals leading the race in terms of AI spending.
So, why is AI becoming mainstream across industries and has picked up drastically over the last decade? One of the major reasons behind this is ‘Cloud’. I would like to draw an analogy of AI and Cloud with a human body. If AI is the intelligence that resides in the brain, Cloud is the muscle that it needs to execute any action or any algorithm in this case. The advantage of AI on the cloud against doing it locally using on-premise infrastructure is that the more the data you train, the cost does not grow proportionately due to the economies of scale it provides. In fact in the world of cloud computing, more is better. This is also one of the biggest reasons for the increase in AI adoption by enterprises post cloud adoption.
Below are some of the areas which I believe will see major developments in the coming 5 years –
Niche AI services
With the democratization of data, we can already see AI services being developed for different industries and in many cases also specific use cases. Enterprises are looking for automation within their domains and AI (in some cases along with RPA) is playing a major role to address business challenges. The growth of industry-focused (retail, manufacturing, healthcare & more) and in certain cases even going one level deep into the AI category within industries (i.e. conversational AI, computer vision, etc.) has been phenomenal. With on-demand compute available at a click, entrepreneurs are picking up focused challenges that can be addressed with AI and build products/services around it.
Accurate AI models
Due to the massive boost to digitization post-COVID-19, a massive amount of digital data is being generated in multiple formats. Thanks to cloud storage, all of it is being stored in raw format by different enterprises. Bias is one of the most important factors in the accuracy of AI models. Bias in AI is also one of the factors that can hamper its update and application within enterprises. However, with most of the data moving to digital form and the high volume of this data getting generated (and now being available to train), the existing AI models are bound to get more accurate and reliable. With much happening in data quality space as well, we can expect the AI models or services to get smarter and more accurate by the day.
While there have been talks on using AI responsibly, there has not much been done in this space. The regulatory and policy landscape for AI is an emerging issue in jurisdictions globally. There have been certain initiatives and papers written in this space by national, regional, and international authorities, but we still see a lack of governance framework to regulate the use of AI. With the growing adoption, we can expect the frameworks and governance to be established and play a major role in the driving development of responsible AI.
There are interesting trends already seen on automating cloud environments using Artificial Intelligence. With enterprises migrating their workloads to the cloud, there is going to be an incredible amount of telemetry data and metrics being generated by these applications and the underlying infrastructure. All of this data can be used to train AI initially to pinpoint the issue, further to provide the resolution, and slowly automate fixing such issues without human intervention. However, it will be interesting to see if any of the hyperscalers (i.e. AWS, Azure, GCP) can build that intelligence from telemetry data of their client’s environment and create service for this automation.
With the adoption of IoT (Internet-of-Things) across industries like manufacturing, retail, health care, energy, financial services, logistics, and agriculture, more data is being generated by the devices with a need of analysis and processing near the device.
According to Gartner, companies generated just 10% of their data outside of the data center or cloud in 2019; this amount is expected to grow up to 75% by 2025.
As a result of this; IDC predicts that in the next 3 years, 45% of the data will be stored, analyzed, processed, and acted upon close to the devices. We already see our smartphones carrying chips to process AI. However, this growing IoT adoption will lead to AI models being deployed onto more edge devices across domains.
While enterprises are looking to create a self-service experience in different domains, on the other hand, the cloud service providers are building products to create self-service platforms for these enterprises to reach to market faster.
As per Gartner, 65% of the app development work will be done using low or no-code platforms and a big chunk of this is going to be platforms to build and train AI.
Hyperscalers have been putting their best efforts to create no or low code platforms for non-techies to be able to create and train their models on the cloud. From chatbots, computer vision to creating custom ML models, enterprises are making use of these platforms to create their offerings with on-demand resources on the cloud instead of re-inventing the wheel.
These are some of the areas where I believe we will see advancements in AI. It would be great to hear some thoughts on what you folks think about the trends in AI.
At LTI, we aspire to create the data-to-decision ecosystem to enable organizations to take quantum leaps in business transformation with AI. LTI’s Mosaic ecosystem has been created to provide enterprises with a competitive edge using its transformative products. While Mosaic AI simplifies designing, development, and deployment of AI/ML at enterprise scale, Mosaic AIOps infuses AI in IT operations to enable pro-active operational intelligence with real-time processing of events. With data being at the center of AI, Mosaic Decisions ensures ingestion, integrity, storage, and governance aspects with Mosaic Catalog ensuring ease of discovering enterprise data.
Mosaic Agnitio’s in-built deep learning enables enterprises to extract insights from data and automate business processes, while Mosaic Lens’s augmented analytics capabilities help uncover hidden insights within the data using AI.
To know more details on LTI’s Mosaic ecosystem, you can visit – https://www.lntinfotech.com/digital-platforms/mosaic/