AI has witnessed its evolution from the pages of science fiction to being an actual business enabler producing tangible business benefits. The profound success of infusing AI into business operations ushered in the concept of AI-led operations, or AIOps. Recent reports highlight the success of AIOps-based incident management, operation automation and enterprise observability solutions, with a 33% reduction in enterprise cloud consumption expenses.
Unsurprisingly, markets are bullish about the growth of AIOps adoption and predict a surge from $11.7 billion in 2023 to $32.4 billion in 2028 at a CAGR of 22.7%. As this growth unfolds, Manish Shah, Managing Director at State Street and Eswar Subramanian, Sr. Vice President and Global Partner at HCLTech discuss the demand for AIOps, including its emergence, scope, challenges, as well as its future during the latest iPredict session from HCLTech.
The promise of AIOps
Shah recounts his first-hand experience of witnessing the evolution of AI from labs to the space of business operations. Several enterprises are already leveraging AI operations, even though many are only starting the journey. CXOs across industries are intent on finding the right fit for their specific use cases and objectives. Speaking on the potential of AIOps, he says: “I think it’s going to change the way all of us operate. Every day and in every part of our lives.”
On CXO imperatives, he adds: “I think every CCIO and CTO is reserving budgets and financials to understand how AIOps is going to impact them, as we move forward.” The promises of AIOps are leading business leaders to plan and strategize their expenditures, with a focus on fostering operational optimization and efficiency in their enterprise.
AIOps improving IT operations
The conversation progressed to the real-world benefits of AIOps, especially in financial services. Answering Subramanian’s question on the mechanism of improving IT operations with AIOps, Shah emphasized the criticality of uninterrupted services.
“Financial organizations have critical business times for high transaction volumes and may have federal commitments in the market, involving significant financial settlements. So, these are critical times, and as a result, this requires minimum time disturbance in the meantime to restore and detect errors. It's essential to minimize both and improve upon all that,” Shah said.
Across industries and in BFSI, AIOps can be the go-to solution that integrates disparate and siloed data sources and automatically generates actionable insights quickly on root cause analysis and remediation. This not only saves time and money but improves efficiency and customer experience as well.
Finding the right match with AIOps
With the benefits of AIOps well established, the conversation shifted toward the best implementation methodologies. Shah reckons the secret to enterprise success lies in preparing themselves; both in terms of strategizing and implementation. This needs careful analysis of the current state of an organization encompassing its data structures, application services, business services as well as overall culture.
As he pointed out: “It all depends upon where the enterprise journey is. They need to make sure they’re adopting to the right culture and a mindset change where they’re able to bring all that together and do the right analysis.” Organizations that are lagging with siloed data structure and mindset challenges will need to catch up to make their IT operations AIOps-ready.
Shah then clarified the role of CIOs, CTOs and other business leaders charged with this organizational overhaul in drafting the right technology criteria and planning investment by accounting for the desired outcomes and priorities. Essentially, it is all about marrying the desired outcomes with the current state of IT operations. The next step involves bringing in the right people with the right skill set, including industry experts and consultants to smoothen the implementation and arrive at the total cost of operations.
Challenging stymying AIOps adoption
However, the hosts agree that it’s not always smooth sailing for everyone. First, there are several challenges with data itself. Numbers indicate that 39% of businesses across the US and Europe grapple with data quality challenges while 35% of them face data access challenges. There are also significant issues around identifying the right data. As Shah stated: “AI systems are only as good as the data you get into them. And organizations are struggling to get the right data into the system.” Other challenges include integrating data from disparate sources, storing the data as well as keeping it secure.
And as with all AI-based solutions, AI models are only as intelligent and effective as the humans feeding in the data. Shah focused on another issue plaguing enterprises — the lack of suitable talent. Moreover, since enterprises will need to feed copious amounts of data in petabytes, costs can be substantial. An observation that Eswar Subramanian agreed with: “I think I was blindsided by the cost factor of it. And once you bought up the amount of data that these could consume, it's a very significant decision as the adoption for AIOps is considered.”
Getting the most out of AIOps
Despite the challenges, the promises of AIOps are hard to overlook, according to Shah. With predictive service management and analytics, AIOps help enterprises reduce their Mean Time to Repair (MTTR) while lowered downtime reduces operational expenses. AIOps also helps streamline IT operations, IT security and service delivery and improves the overall customer experience.
Streamlining of ITOps and security means more time for productive analytics and higher efficiency. Shah stated: “As you streamline your IT operations and security, you can do more predictive analytics. You don't require segmented multiple teams and now can optimize the efficiency across your enterprise and your whole plan on IT operation.”
He also elaborated on how organizations can build a Center of Excellence (COE) by upskilling resources who can help in continually training the AI model. With machine learning (ML), organizations can anticipate issues with service management including hardware failures, prevent downtimes and reduce operational risks.
“So, I think there are a lot of benefits if enterprises use AIOps effectively,” he said.
The future of AIOps in operations
Wrapping up his views on the future of AIOps, Shah was confident about the growing role of AI-enabled solutions. However, he expected more cross-domain integration to emerge.
“Where I still see more evolving to happen is in cross-domain integration, such as AIOps and DevOps. So, you'll have a seamless integration development process. This will help to deliver faster and reliable software development and deployment.”
The second part of this is AIOps’ role in security. Shah elaborated on the capability of AIOps to play a larger role by detecting anomalies to take immediate action and further the understanding of internal and external threats.
With comprehensive insights into operations, AIOps is also expected to evolve past IT operations into the domains of finance, HR and customer support. Here, Shah provided examples of use cases with a sequence of questions that AIOps helps answer:
“Which customer of mine had a problem? What in the chain has created a problem? How can I win this business and what is my customer looking for? How am I making my financial decision? How do I make my HR decisions? Based on the individual I'm hiring, what are the challenges? What are the biased approaches we are taking?”
However, at the same time, he doubled down on the importance of organizations eliminating biases, while prioritizing ethical considerations and ensuring fair decision-making with proper training. Shah concluded by asserting how AI solutions, with an already established foothold across industries, are being segmented with different players and industries with organization-focused requirements.
“And I think that would change the way industry's operating. We'll have clarity about which particular model is effective for which use cases. So, that's changing, and depending upon their interest, I think they should be learning those products more in and out,” he said.
As the discussion closed, one thing became clear — businesses armed with an understanding of the scope, the opportunities, the challenges and the future of AIOps will have the best chance of implementing these solutions and driving value from them.
