A-RICE Framework: An IT leader’s guide to evaluating AI initiatives for enterprise IT

When it comes to implementing AI and solving ITSM challenges, supply is plentiful enough that figuring out the right initiative is the issue.  

Time, money and effort are of the essence when it comes to picking an initiative that can help your IT teams deliver faster results but even accelerate innovation so that the business sees faster results.  

This is where a formula that product managers use to prioritize features can help enterprise IT leaders on the lookout to narrow down AI initiatives and make the case for them.

The Reach-Impact-Confidence-Effort formula (a slight AI tinkering as a nod to complexity) can be used to take difficult decisions and balance innovation against cost, efficiency and business growth.  

Stay with us till the end, where we share a free calculator to get started with the AI initiative prioritization.

Do I need AI for that?

Before we dive into a prioritization formula, we must address the elephant in the room i.e is AI needed for solving every ITSM challenge?  

Not every IT challenge requires AI. Before investing in an AI-powered solution, IT leaders should assess whether the complexity and benefits justify the investment.  

AI is the right choice for all initiatives where:

• the problem involves unstructured data,  
• the solution needs to learn and improve over time,  
• the solution does not need to be word-for-word accurate,
• the solution involves complex decision-making and  
• where static automations just aren’t flexible enough to handle variations.

Depending on the usecase, AI isn’t necessary—but when it is, choosing the right AI approach can make all the difference.

Another popular question that pops up is: should we build it internally or buy it from a vendor? Depending on the initiative and the offering, this can end up being an expensive monetary and resource operation. This decision is the most important one we can take and can make or break whether you read the rest of this article.

Let’s breakdown A-RICE

After categorizing your use cases and narrowing down your enterprise AI approach, we approach the formula. I’ve tinkered with the formula a little bit to adjust for the target demographic and the problem at hand. An IT leader looking to prioritize initiatives cannot go by the same formula as a PM looking to build AI features for value.  

The formula focuses solely on quantifiable metrics but doesn’t capture other qualitative factors such as vendor reputation, ease of integration, long-term support, and strategic fit. While these might be incorporated separately (as with a vendor multiplier), they aren’t naturally integrated into the basic equation.  

Note: Each factor—Reach, Impact, Confidence, Effort, and AI Complexity—is rated on a scale of 1-5 (or specified otherwise) that can be highly subjective. Different teams and leaders of differing organization sizes and industries, might score the same initiative very differently, which can lead to varied prioritization.

• Reach: How many employees, teams, or processes will benefit from this AI feature? 5 if it’s the entire company. ‍
• Impact: Will it solve a major pain point in ITSM, or is it just a nice-to-have? 5 if it’s a major pain point.  ‍
• Confidence: How certain are we about the success of this implementation based on available data and past performance? 5 if it’s a proven solution. ‍
• Effort: What level of resources—engineering, time, and integration—are required to make it work? (This varies significantly based on whether you're building a solution from scratch with services like Azure OpenAI or using an out-of-the-box platform like Atomicwork.) ‍
• AI complexity (A): How demanding is this feature in terms of data, compute power, and long-term scalability? ‍
• K (Complexity of solution): A risk modifier that adjusts for how directly you're handling AI complexity. Set k = 1 for custom-built solutions (direct risk) and k < 1 (for example, 0.5) for vendor solutions, where the vendor’s expertise reduces the effective risk.

By applying A-RICE, IT leaders can cut through the noise and focus on AI initiatives that truly move the needle for their organization.  

Failing to prioritize AI initiatives effectively can lead to wasted resources, technical debt, solution fatigue, and low adoption rates.

A-RICE scores for AI initiatives

To make prioritization clearer, we assign a score to each factor (1-5, with 5 being highest impact/complexity and 1 being lowest).

The final score can then be calculated as:

A-RICE score = (Reach × Impact × Confidence) / (Effort + k*AI Complexity)

For instance, if you calculate the A-RICE score for an employee onboarding use case, the initiative will score highly across the board. This is because streamlined onboarding significantly improves employee experience and IT efficiency and is considerably less effort to automate because it is largely a structured, rule-based template. The reach and impact are higher too as every new hire in your organization would benefit from this initiative.

We’ve created a calculator based on the above scoring model, giving sample data for three big ITSM initiatives.

Get the free calculator to check and prioritize your AI initiatives.

A-RICE in action: An IT leader’s perspective

AI is no longer a futuristic concept—it’s a critical tool that can transform IT operations when applied effectively. However, not all AI initiatives are worth pursuing, and IT leaders must be strategic in their adoption. The A-RICE framework provides a structured way to evaluate, prioritize, and justify AI investments, ensuring that resources are allocated where they will drive the most impact.

Have an AI initiative in mind but don’t feel like reaching for that calculator app to make your decisions?

Use our free calculator to figure out which AI initiatives need to be prioritized!

Curious how our agentic service management platform can transform your IT operations? Let’s talk.