Last Updated Sep 01, 2020 — AI-Powered Analytics expert
AI-Powered Analytics

Digital systems have never been more critical or central to your business than in today’s rapidly changing world. Businesses used to rely on best practices for operations they performed over and over, but that’s no longer the case. Current scenarios are constantly changing, and it’s hard to say what “best practices” are any longer.

Successful product teams should be structured so they move through uncertainty and avoid the mistakes of the past by focusing on all stages of the product lifestyle. Rather than concentrating on the outcome of projects, which can hide important aspects of value creation, they can focus on value creation at all stages of the product development lifecycle by following the EPOCCA model:

  • Empathize – understand what delights customers or stakeholders 
  • Persist – last as long as they need to, and don’t rely on the whim of the project management office or annual funding
  • Outcome – focus on profit and loss, rather than only on deliverables 
  • Collaborative – communicate with each other
  • Cross-functional – bring multiple skills to the team
  • Autonomy (relative) – have a certain freedom to operate because they’re working in new territory

Product teams have a need to not only constantly adjust the product, but also to proactively anticipate problems and opportunities in light of data feedback from the entirety of the product life cycle. Artificial Intelligence (AI) provides the data analysis capabilities IT operations need to see emerging trends, identify relationships within complex systems, and know the outcome of certain decisions to determine the best choice for that moment ⁠— in advance of being forced to make that choice. 

There is no longer one “best practice” that can be applied ubiquitously. Instead, there’s an optimal choice for the time and place we’re at now, and AI can be used to clarify the options for what that choice might be.

Moving Our Mindsets Beyond Frameworks based on the Industrial Era

Traditional Information Technology Infrastructure Library (ITIL) frameworks for Change Management are based on industrial era values like consistent processes. However, these frameworks are not designed for an Agile world.

Software release frequency is increasing worldwide. Currently, 47% of respondents are releasing new product iterations at a rate of two-to-four times a year, a number affected by their “packaged” software approach and their legacy systems that are still in use, like on-site corporate mainframes. Companies that release more often – at the rate of once a day – have tripled in volume over four years and are steadily rising, growing from 1% to 3%, a small, yet statistically significant, amount. 

This rapid release schedule stands in contrast to the old industrial way of managing products as a single release with minor changes ⁠– or no changes ⁠– once initial R&D had been completed. Even digital products were released with this mindset until very recently.

Organizations must address these rapid, hard-to-predict changes by moving away from the legacy mindset of “business as usual.” That state no longer exists. With constant change in our environment, we have to be on top of emerging trends and risks before they impact the value change.

Because of the need for visibility and accountability throughout the product lifecycle, we also can’t continue a siloed approach to business responsibilities. Every part of the value chain matters, and every part can potentially contribute to an IT problem or incident. We can’t wait to respond to feedback from other departments or stakeholders; we need to generate our own feedback by viewing relevant data from across all major systems of record across the organization.

Modern Change Management Processes Require More Agility and Proactive Risk Management

No one expected businesses to suddenly be disrupted and work to shift largely to remote teams in the middle of Q1 2020. This provides an excellent example of unpredictable and emergent market conditions. In reaction to this recent volatility, many management processes and teams have instituted a temporary change freeze during the shift to remote work and/or a distributed workforce. 

It’s important for organizations to note that those slowest to adapt to volatility probably won’t successfully make it through that volatility in good shape. A reactive approach of freezing changes then releasing them all at once carries a greater risk of change-related incidents – like releasing water from a dam carries a  risk of causing downstream damage. 

To address change risk without instituting global change freezes, IT operations teams need a proactive change risk management strategy.

Ideally, modern change management encompasses three key disciplines:

  1. Predict and mitigate change risk. This reduces the risk of individual changes and removes the systemic positive change failure. 
  2. Increase change velocity and productivity. This will improve the value creation within IT change processes by allowing your governance stakeholders or Change Advisory Board (CAB) to focus on truly risky changes while increasing change velocity for low-risk changes to respond more agilely to evolving business needs.
  3. Accelerate change failure recovery. If you know which changes are truly high risk, you can proactively monitor the outcome of a change and fix issues before your customers find them.

Control Complexity Without Bottlenecking Value Creation

The three IT change management priorities above illustrate the shift in perspective in strategy that IT operations teams need. Legacy institutions, like a CAB that reviews any and all changes for manual approval, are simply too slow and cumbersome to allow businesses to keep pace with today’s modern, rapidly altering business ecosystem.

For a long time, change management was done entirely through CAB oversight. This practice originated from ITIL methodologies, which had the intent of manually controlling stages of work where risk could be introduced.

What many don’t realize is that the original intent wasn’t for the CAB to handle all production changes, but instead to handle only high impact, higher risk changes. However, over time (and partially in response to the Enron/WorldCom scandals and Sarbanes-Oxley) the CAB has become the repository of all change management decisions in many organizations.

However, research shows that performance drops when all changes go through a CAB. Peer reviewed changes – or even no change process – is correlated with higher performance. 

Another issue with CABs is that they react to information on a case-by-case basis, leading not just to latency, but also a mindset that could fail to recognize emergent issues. 

This is where the value of AI analytics comes in. AI can detect emergent issues more quickly than human reviewers. It can present information – such as a scorecard revealing the risks of certain changes – in a way that allows us to proactively anticipate situations and address them before production issues can affect the product. AI scoring can also allow us to determine which changes can be automated, which ones need a closer look, and which ones should be frozen or scrutinized further before deployment

The Maturity Model for Analytics Driven Change Management

Recent volatility has highlighted the fact that organizations need to think about strategy from a top-level perspective. Otherwise, they may take things for granted and fall back into the reactive mindset, with the assumption that repetitive best practices are enough until they’re presented with a problem.

To reach this level, organizations must make deliberate moves to adjust from a reactive to a proactive stance on change management. A model of process maturity across organizations can look like the following:

  • Reactive – Defined by metrics, with focus on outputs of processes rather than outcomes, and reliance on user reported incidents instead of preventing them.
  • Proactive – A focus on proactively finding and mitigating risk at the time of change requests and deployment resulting in lower change failure rates and higher productivity. This is achieved by consistently adopting analytics across teams and processes, driving continuous service improvement initiatives, and using a system of applied metrics.
  • Optimized – Automatically detecting outcomes and linking them back to root cause, with the AI model allowing teams to predict outcomes and drive improvements.

An example of optimized change management is using change risk scoring for teams and individuals to establish metrics that can be used to benchmark performance and enforce accountability on a micro level. Proactively addressing sources of change risk helps reduce the effects before changes have a chance to impact production or slow down continuous integration/continuous delivery.

A Machine Learning-Driven Approach to Success

Change success is hard to achieve, primarily because of siloed systems in many organizations. While there’s generally a lot of data available to help organizations predict change success, siloed data can lead to incomplete views, bias, skewing, and blind spots. These flaws collectively not only make data harder to analyze but they also obscure insights in a way that affects development, deployment, and customer experience.

An approach driven by AI and Machine Learning (ML) technologies can be fairly effective at solving the problem of change risk without introducing latency. ML can mitigate the risk of individual changes, while helping to remove systemic causes of change failure. However, there’s only so much machines can take over. This statement is especially true if the processes they are controlling are flawed or inefficient in some way. Because of this risk, human oversight of change processes isn’t going away anytime soon, even in the era of DevOps. 

By using AI analytics with CAB oversight, you can elevate the function of the CAB, so that it:

  • Instead of gatekeeping, holds teams accountable for success
  • Helps the team mitigate change risk
  • Helps everyone develop and adopt best practices

For the change process to be successful, it’s vital to have clear communication, as well as transparency so the team knows what they’re being measured on and why. Many companies have found that by defining, measuring, and centralizing the right behaviors, they can progress down the road to more proactivity and, eventually, continually optimizing their change processes.

Learn more about how AI analytics can help your organization achieve process maturity with our recent webinar: “Change Management: From Reactive to Proactive with AI

Read the Report

Are you ready to scale your enterprise?


What's New In The World of

May 5, 2023

AI Risk Management Means Less Stressful Work Weeks

Roughly 41% of tech professionals working have a high risk of burnout, which is why the need for AI predictive analytics is greater than ever. Read on to learn more!

Learn More
March 2, 2022

Automate low-risk change teams to create a “fast lane” for release velocity

Increase deployment speed and efficiency by using data models and AI-powered analytics that enable DevOps leaders identify which change categories and change teams have a high rate of change success. Read on to learn how our Change Risk Prediction solution makes the difference.

Learn More
February 23, 2022

Explore with DORA: Leveraging DORA analytics to their fullest

In today’s completely digital world, organizations and individuals alike are acutely aware of the speed of change and how it affects them. Read on to discover the importance of DORA metrics and how AI/ML can make DORA metrics more actionable.

Learn More