Meet the women shaping the future of the power industry.
AI grid inspection helps utilities turn millions of drone and aerial images into prioritized, actionable maintenance decisions. According to Kaitlyn Albertoli, CEO and co-founder of Buzz Solutions, load growth, faster equipment failure, wildfire risk and workforce shortages have moved AI from a source of skepticism to a necessity. Success comes from starting with focused use cases, integrating insights into existing GIS, asset management and work management systems, and keeping experienced linemen in charge of decisions.
What You Will Learn
- Why load growth, wildfire risk, storms and labor shortages are accelerating AI adoption in utilities
- What an “AI-first” approach to grid inspection means in practice
- How utilities move AI beyond the pilot stage and into daily workflows
- How drone docks and image analytics make more frequent inspection cycles possible
- Why technology providers need to “meet utilities where they are”
Meet the Expert: Kaitlyn Albertoli
Kaitlyn Albertoli is the CEO and co-founder of Buzz Solutions. She graduated from Stanford alongside her co-founder, has a background in sustainable food and finance, and has worked in the energy industry for about 10 years. Before launching Buzz, she and her co-founder spoke with more than 35 investor-owned utilities about their maintenance and inspection practices.
Introduction
Utilities are collecting more inspection data than ever before. Drones, helicopters and fixed-wing aircraft generate enormous image libraries, yet analyzing them has traditionally been a manual task. That gap is becoming harder to ignore as aging infrastructure, data center load growth and workforce shortages stack up.
In the inaugural episode of Powered By Her, hosts Audra Drazga and Samantha Gilman speak with Kaitlyn Albertoli about how AI, drones and inspection data are changing the way utilities monitor and maintain critical grid infrastructure. The podcast is produced in partnership with APC Media.
The conversation covers what utilities are asking for today, how AI image analysis plugs into existing systems, and why inspection cycles that once seemed impossible are now within reach. It also offers practical advice for women and newcomers entering the power industry. Below are the key insights, explained for utility, TSO, DSO and engineering audiences.
Why Are Utilities Adopting AI Grid Inspection Now?
Utilities plan in five to ten year cycles, so shifts in technology adoption tend to be gradual. Albertoli says the conversation has changed noticeably over the last 18 to 24 months. Three or four years ago, AI met hesitancy, skepticism and some fear. Today, she describes it as becoming a necessity in some ways.
Load Growth and Faster Equipment Failure
Albertoli points to stress and strain from load growth, with equipment failing at faster rates than before. Aging infrastructure and data centers coming online were also raised by the host as part of the changing landscape.
Wildfire, Storms and Labor Shortages
Storms have long driven inspection programs, and wildfires have been a major catalyst. Utilities also want to reduce the risk of sparking a wildfire during pre- and post-storm restoration. On top of this sit significant shortages of linemen and field technicians. There are simply not enough people to analyze all the data, triage findings and direct crews to the field.
What Does an “AI-First” Approach to Grid Inspection Mean?
An AI-first approach means building the algorithms from scratch for the electric utility context, rather than adapting generic image tools. Buzz Solutions trained its models on more than 15 years of historical utility data across different asset types. That training allows detections to go beyond “this is a pole” or “this is an insulator.”
Granular Detection and Defect Severity
The system identifies what type of insulator is present, what kind of anomaly exists, and how severe it is. Albertoli gives an example: bird droppings on an insulator carry a different severity than flashing or a broken disc, which can be much more critical depending on location. This level of detail builds confidence among linemen who are looking for specific critical defects.
Keeping the Human in the Loop
Albertoli argues that AI moves experienced field staff from manual data labeling to reason-based decision making. Linemen and technicians often have 20, 30 or even 40 years of experience. With AI results in front of them, they decide whether a finding warrants a truck roll, should be revisited in six months, or is actually fine.
How Do Utilities Move AI Beyond the Pilot Stage?
Albertoli says utilities have moved past demonstrations. Many are now processing hundreds of thousands of images per year, with a goal of millions. In one example with an East Coast utility, more than 100,000 images were processed in under 10 hours. Utilities have gone from collecting thousands of images to collecting millions.
Integration With GIS, Asset Management and Work Management
The shift from standalone image processing to an integrated workflow is central to scaling. Insights are pushed into systems utilities already use:
- GIS (such as Esri): geospatially identifies assets, sends crews to the field and supports autonomous inspection programs.
- Asset management systems: the system of record where components and their prioritization are cataloged.
- Work management systems: where work tickets are drafted and crews are dispatched.
Direct integration keeps these systems up to date and reduces redundant manual data entry.
A Common Misconception About Implementation
Many utilities assume a program like this requires a multi-year rollout, heavy investment or a separate tool that every lineman must learn. Albertoli says programs are taking off faster when they start with a couple of focused use cases, demonstrate success, and expand over time.
Can Drone Programs Enable More Frequent Inspections?
Yes. Albertoli says one customer is looking to move from a five-year inspection cycle to a one-year cycle, something that would have seemed almost impossible a few years ago. Previously, that change would have required a linear increase in drone pilots, who are hard to find and train.
Drone-in-a-Box, Docks and Targeted Reinspection
Drone-in-a-box solutions and drone docks, combined with image management and analytics, unlock capacity for more frequent inspections. That includes a baseline annual inspection of an entire asset base (she mentions 30,000 to 50,000 assets), plus closer monitoring of higher-risk assets such as those in critical wildfire zones or those affected by a recent storm.
How Did EPRI’s Incubatenergy Labs Help Validate the Technology?
Buzz Solutions took part in two cohorts of EPRI’s Incubatenergy Labs. Its first larger-scale commercial demonstration was with a Canadian utility and a drone service provider, validating the algorithms on both transmission and distribution infrastructure. The program gave the company a validated public report to show other utilities, which helped reduce perceived technology risk. This came after about three years of building and training the algorithms.
Advice for Women and Newcomers in the Power Industry
Albertoli says more women are entering the sector, noting that at a recent utility conference nearly every panel included multiple women. Her advice is to dive in if you are passionate about the problem, and to ask early questions, especially “why” a process has been done a certain way for 30 or 40 years. She also argues that the industry is more eager to evolve than its reputation suggests, and that technology providers should meet utilities where they are to win early results and buy-in.
Conclusion
The interview points to a broader shift from reactive to proactive utility operations. AI and drone-based inspection are no longer experiments; they are becoming tools to cope with load growth, wildfire risk and a shrinking field workforce. The practical lessons are clear: integrate with existing systems, start with focused use cases, and keep experienced people making the final calls. Albertoli says the industry is only scratching the surface of what AI can do.

