•
Sep 4
Sand Technologies
Before any calculation can begin, define the scope and metrics of the AI project. Start with the “why.” Clearly articulate the specific business problem the AI solution will address. Is it to reduce outage durations, predict asset failures before they happen, or improve vegetation management efficiency? A well-defined problem is key to a successful business case.
Once the problem is defined, identify the Key Performance Indicators (KPIs) to measure success. These metrics must be tracked and collected before the AI implementation to establish a clear baseline. This step is crucial. These metrics include the following.
For companies or departments pitching the first AI project, don’t forget any intangible benefits. The first AI project helps a company understand how to utilize AI, including the process, data relationships and any skills gaps it needs to address.
According to the ABB Value of Reliability: Survey Report 2023, outages cost industries an average of $125,000 per hour, and 69% of companies experience at least one outage per month. With 66% still employing run-to-fail maintenance, preventing breakdowns can lead to significant savings.
In fact, McKinsey research (in the manufacturing industry) found that predictive maintenance can reduce unplanned machine downtime by 30%-50%, increasing machine life by 20%-40%. This data makes predictive maintenance a compelling starting point for AI.
To calculate the ROI of a predictive maintenance AI project, use recent company-specific repair costs to estimate the savings achieved through a preventive maintenance approach, considering both the low end (30%) and the high end (50%) to determine the range of potential savings. Include a separate calculation for extending the cost of the capital equipment by 20%.
Savings Calculation:
(cost of reactive repair – cost of proactive repair) x number of failures prevented
Once the AI-driven predictive maintenance system identifies that a preventive measure is needed, AI-driven scheduling and dispatch systems can optimize maintenance management by dispatching the right crew with the right equipment to the right place faster, minimizing travel time and costly overtime. Overtime rates vary by country and industry, but in the United States, federal regulations require 1.5 times the hourly rate for every hour worked beyond 40 in a week. Being aware of a potential breakdown and scheduling repairs during regular business hours can lead to significant savings. The formula below only calculates a 30% reduction in possible overtime. Keep in mind that catching problems early could result in higher savings..
Vegetation management is often one of the most significant costs for utility companies. However, new technology is helping reduce these costs. AI analysis of satellite or drone imagery can pinpoint high-risk areas for tree trimming with incredible accuracy, preventing vegetation-related outages and avoiding potential regulatory fines.
For this use case, determine the current cost of infrastructure inspections and the average regulatory penalties from the last three years, subtract the cost of AI-powered vegetation management.
Outages are expensive. They result in direct costs, regulatory penalties and crew deployment, as well as indirect costs, including reputational damage and a negative economic impact on the communities served. Utilizing AI is an effective way to enhance grid reliability. This improvement has a clear dollar value.
Many regulatory frameworks include performance-based ratemaking or other incentives tied to reliability metrics. By improving the System Average Interruption Duration Index (SAIDI) and System Average Interruption Frequency Index (SAIFI) scores, companies can unlock new revenue streams or avoid significant penalties.
Saving Calculation:
(reduced SAIDI points x per point value) + (avoided SAIFI events x value per event)
Now it’s time to bring everything together. Start by tallying the benefits to build the complete ROI model. Sum up all the quantified financial gains above. Next, account for the costs of the AI project. Be comprehensive. Include upfront costs for software, hardware, and implementation, as well as ongoing costs for licensing and training.
Calculate the ROI by using the standard formula to determine the return on investment and payback period.
Consider running a sensitivity analysis that models best-case, worst-case, and most-likely scenarios. This level of detail demonstrates the robustness of the business case and shows consideration of all potential outcomes.
Strong analysis is only half the battle; presenting the findings effectively is equally crucial. Tell a compelling story with data that resonates with your audience, whether it’s the board of directors, regulators, or internal stakeholders.
AI is not just another IT project; it is a strategic investment that can transform utilities into innovation leaders. The urgency has never been greater—aging infrastructure, rising customer demand and increasing regulatory pressure—are converging to make efficiency, reliability and resilience non-negotiable. Utilities that adopt AI early will secure the financial and operational advantages needed to lead in this new era.
The path forward is clear: start small, with a single high-impact use case, apply a disciplined ROI framework and scale from documented success. By proving value with measurable KPIs, utilities can turn AI from a buzzword into a boardroom-approved business driver.
For utilities struggling to define the right KPIs or build a compelling business case, Sand Technologies can help. We work with utilities to identify the most impactful use cases, establish the right performance metrics and guide pilot projects that demonstrate measurable ROI.
The future of AI in utilities is not about potential; it’s about proven results. The question is: will your utility seize the opportunity to innovate, or risk being outpaced by the industry?
Use this AI project ROI calculation checklist to ensure you cover all the essential steps when building your business case.
Other articles that may interest you
