eWEEK content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.

In its simplest form, AIOps (Artificial Intelligent Operations) is a marriage of AI technology that’s been around for decades. AIOps arrived at its current incarnation from two different directions.  For example, the jointure of machine learning with monitoring and operations technology.

Along the way, existing monitoring and ops players who may now support both traditional systems and cloud-based systems initially just “AIOps washed” their existing product lines.  In most cases, they bolted on an AI engine and provided some type of integration with the existing tools, databases, and repositories to assist in the operations of cloud and non-cloud-based systems.

It’s the net-new startups that pushed purpose-built AIOps engines into the marketplace.  These AIOps players typically built systems designed primarily for cloud-based monitoring but they also added some new features.

Some of the new features include self-healing capabilities and built-in orchestration and automation which allow the CloudOps teams to remove most of the human interactions that were formerly required to do CloudOps. The days of ops teams carrying around company cell phones to look for critical errors are over, and it’s about time.

What’s the real cost?

AIOps implementations can cost about $10,000 per system per year, whether your enterprise upgrades to AIOps through a single vendor strategy, or if you replace your current ops automation using new AIOps tooling, or, more likely, if the enterprise adds AIOps for the first time. Aside from the cost of the tool, which can now be implemented as-a-service, you must also consider the cost of outside consultants and the disruption to the organization as new tools and processes are introduced.

An enterprise with 2,000 systems, including cloud and non-cloud compute, databases, and other required systems, often ends up with a $20,000,000 AIOps bill per year, all factors considered, for 2000 systems. While this may produce sticker shock in many CIOs, this cost is in-line with what other digital transformation components cost, such as data integration, process orchestration, and leveraging industry clouds, to name just a few of the current trends that enterprises follow.

The trick to justifying the cost of AIOps is to understand what AIOps is, as well as how to define the required value. Keep in mind, you can rarely take baby steps into the AIOps space.  You must use all of its capabilities across most of your systems.

The cost to support, say, 10 percent of your cloud and non-cloud systems using AIOps tooling and approaches does not result in a sufficient benefit to justify the expense. Indeed, enterprises that take a partial AIOps implementation approach usually make things more confusing for the ops teams and more expensive for the enterprise.  The old saying, “in for a penny, in for a pound” definitely applies to this situation.

What’s the real value?

We can define the value of AIOps in the same ways we do other strategic technologies.  First, define the “as-is” state of things without the technology.

Then, define the “to-be” state with the technology, including any changes in value that affect intangibles like business agility, revenue growth, rising customer satisfaction, and even employee morale.

Experts and consultants (including yours truly) can provide insight into many of the intangible value questions, but the values will vary greatly depending upon the type of business involved, the size of the business, the industry, as well as the regulatory environment the enterprise exists within.

In other words, the value that AIOps will bring to your enterprise is largely unique to your enterprise, and thus the models that you leverage to understand the value to justify the price tag will be unique as well.

With that said, it’s important to understand the most likely points of value or the tactical problems AIOps could solve for the enterprise. What value multipliers can you attach to these value points?

Let’s go through the major points of value that I see for most AIOps projects and attach T-shirt sizes to define the degree of value that most enterprises will see.  Again, your actual milage will vary.

Point 1: Better collaboration.  One of the benefits of AIOps that many don’t see or understand, is the fact that this technology enables collaboration and workflow in and between IT groups, as well as in and between other business units.

For example, if a database server is about to fail, it’s important to quickly determine the root cause of the imminent failure and automate a fix. But the system must also alert the business units and employees who will be affected by the issue, and often incorporate their assistance to solve the problem (such as by stopping data entry for 10 minutes while the databases is switched out or rebooted).

While this seems like a minor thing, the proactive actions of the ops team can provide much better productivity gains and confidence in the systems. In the old days, an ops team would have waited for the failure, which meant waiting for irate calls from the database users to report that something was wrong. This costs time, money, and worse…marginalized opinions of IT’s reliability.

For most enterprises, this as-is value point falls in the Medium range.

Point 2: Better productivity.  It goes without saying that fewer outages will result in increased productivity. However, better productivity in this context goes beyond the traditional uptime definition.

AIOps is no longer about “report-analyze-respond-repair” as it was with traditional ops tooling. Today’s AIOps is about detecting issues using a “big brain” AI engine that identifies problems most humans cannot. It’s like having 1,000 operations experts collaborating on an issue in less than a second. Using this big brain mechanism, the AIOps tool can detect issues before they become issues, as well as automate the fixes before a human ever comes into the loop.

So, what does this mean in terms of value? Happier customers and happier staff.  i.e., when making a purchase, customers no longer need to deal with system slowdowns or outages or interact with customer services. Employees don’t feel as if they are the designated fall guy if or when outages do occur. Finally, the revenue line sees the benefit as well when AIOps automation results in more sales via happy customers who buy more product and the retention of staff that reduces the costs of employee turnover.

This is perhaps the biggest area of value for most enterprises, and the core reason they add or enhance AIOps to the enterprise.

Point 3: System performance benefits. Most ops teams don’t pay attention to performance as closely as they should. The primary metric often seems to be a yes/no answer to the question, “Do the systems work?”  Instead, ops teams should question how well the systems work and how well they perform.

AIOps brings a few pluses to performance:

• AIOps can spot performance issues that most users and customers will never notice or identify and correct those issues on the fly.  An AIOps systems may find that memory allocation needs to be increased, or other virtual servers should be added to enhance performance, etc..
• AIOps can predict what resources are required to maintain a base level of performance, and auto scale those resources before a performance issue occurs.  So, if performance is no longer an issue, what’s the value of that situation?

Ops teams and IT in general usually underestimate the impact of performance on revenue.  A 10 percent increase in system performance can result in 100 million dollars in additional sales over a year for some retail companies. Huge value can exist here. Be sure to understand the point value to your enterprise’s particular problem set.

Additional AIOps Benefits

• The value of business agility that AIOps can bring by shifting ops around what scalability and changeability are needed.
• The value of not having to replace hardware and cloud services as frequently.
• The value of predicting ops costs and modeling future costs around budgeting limitations.
• The value of maintaining compliance with regulations and policy, thus avoiding fines.

Each point has its own value that is based upon the operations of a specific enterprise where AIOps is being evaluated or implemented.

Value is a malleable concept. While AIOps is valuable most of the time, the degree of value must be determined for each enterprise.  The purpose here is to provide you with some common places to look for intangible values of AIOps.  Understanding the soft costs and benefits is how you will find the true value of your new or enhanced AIOps tools.