Comprehensive software reviews to make better IT decisions
Lesson Learned From an Interview: AIOps Is Not Just About Tools
TietoEVRY is Scandinavia’s biggest IT and business consulting firm and outsourcing services provider with about 24,000 employees, serving more than 10,000 customers in over 90 countries. For users who are interested in IT automation, leveraging AI and ML capabilities, the following interview with Himadri Das, the head of automation at TieroEVRY, will be of interest. Areas that are covered in this interview include the important things businesses should consider for implementing AI-based solutions, the maturity level that their practices should have, the possible risks of adopting AI, and areas of AIOps that need more improvement.
As per Himadri, there are three categories of customers:
- First-generation outsourcers, who operate internal IT with no outsourcing experience. They are typically non-technical savvy, risk averse, and consist of small to medium-scale family-run or localized companies. They lack maturity and know-how of automation and its impact.
- Second-generation outsourcers, who have already used some AI capabilities and understand some basics, but don’t have a full stack of AIOps.
- Extremely advanced group. This group of users know the outsourcing world extremely well. They have been using AI in the past 20-25 years. They understand AIOps and are eager to implement more functionalities but are practically challenged by the tussle between business and IT priority and during these escalations, the vision towards a full-stack, integrated automation focus is lost.
IT automation has numerous benefits, such as efficiency enhancement, cost reduction, error decrease, risk mitigation, and end-user satisfaction improvement, which frees resources to focus more on business activities than operational activities. It also improves business process effectiveness. However, just like any other initiative, AIOps implementation has its own risks. Automation needs discipline and a high level of commitment. Customers should first identify their manual processes that need to be automated and be committed to give all the approvals and push delivery teams. Overall, the risks involved are comparable to the risks in any major IT process change. However, risk of failure in AIOps is minimal, as this risk passes to the partner or vendor.
The following picture depicts TietoEVRY’s AIOps-led automation architecture, covering all key elements like the situational dashboard for measuring operational and business SLAs and KPIs, NLP-based ticket analytics engine, chatbots, virtual agents, the ticketing tool, automation orchestration platform, event hub, monitoring systems, data lake, and integrated DevOps to orchestrate RUN, BUILD and OPERATE catering to events, incidents, and standard/non-standard service requests.
Image 1: TietoEVRY Integrated AIOps Solution. Source: TietoEVRY
The COVID-19 outbreak has disproved many myths and has been acting as a game changer. For instance, prior to the COVID-19 pandemic, many people thought that service delivery is better only when staff is working in the office; however, the majority of employees are nowadays working from home and are able to provide services without major disruptions. That said, businesses have the ability to meet most of their responsibilities by leveraging AI-based solutions and technology. Nonetheless, AI/ML looks like a buzzword for clients. “AI is not a tool,” said Himadri. “It's a holistic approach to building automation. You have to do the hard work and have the use case to make a success of AIOps.” Customers want to see not only an automation solution on the backend, but also the real impact of automation on the support function as well as business operation. They want to see the real use cases. The success of automation initiatives will be even higher when customers break silos. For instance, automating a specific process will improve only that specific workflow or IT process. However, the real benefit is when customers have a full-stack automation aligned with business KPIs, like automation of the service desk and self-service portal. Nevertheless, Himadri strongly suggests taking baby steps to implement AI capabilities based on the level of requirements, then taking the further steps to improve other processes.
Many professionals believe that AI will seamlessly automate all IT processes and resolve all problems. AIOps certainly facilitates these initiatives. However, it needs to be implemented thoughtfully, through carefully identified use cases supported by a detailed production rollout plan, otherwise, it will lead to a waste of time, money, and resources. Initially, your IT processes need some foundations to be able to leverage the benefits of AI/ML.
AI and ML need excessive time, investment, and resources to show their real benefits to the business. The best place to start to show the real benefits of AI to your business is aligning its capabilities with your business requirements. Improve your practice maturity, start with low-hanging fruits, and continue automating other processes as you go ahead. For your process improvement, analyze your internal capabilities, and if it’s not possible to do it in house, outsource the process. This way, you will improve the chance of success, and the risk of failure will be transferred to the MSP.