Utilities today are squeezed across multiple priorities—including reliability, cost, and safety—and are facing increasing challenges related to labor shortages, regulatory scrutiny, and a post-COVID-19 hybrid work environment. Companies have generally taken on efforts to address their priorities individually, and this one-metric focus often leads to the inaccurate perception that there is a trade-off among reliability, cost, and safety. As a result, significant optimization improvements remain for those who are ready to take a more holistic, end-to-end approach.
Our experience in process transformation efforts across utilities indicates that the largest drivers of execution waste relate to the initiation of work orders, planning and scheduling handoffs, and information silos. Many of these issues can be traced back to traditional work management processes, which rely heavily on many time-consuming and inconsistent manual processes.
Smart scheduling involves analytics-powered algorithms and user-centric interfaces that can be deployed in a matter of months and within existing systems to build better, faster schedules. AI-enabled smart scheduling that efficiently matches resources with work can transform companies’ ability to drive long-needed improvements across multiple competing priorities. It can free up scheduler time, boost worker utilization, and increase productivity by 20 to 30 percent. These additional resources can then be used to reduce overtime, insource contractor spend, or reduce job backlogs.
AI-enabled smart scheduling can transform companies’ ability to drive long-needed improvements across multiple competing priorities.
In our experience, successful deployment of smart-scheduling tools requires utility companies to learn five key lessons:
- Data are crucial but should not be a barrier to starting.
- Technology must work in conjunction with processes.
- Businesses must clearly specify their optimization criteria.
- Piloting, followed by intentionally scaling, a “light tech” scheduling solution is vital to increasing adoption.
- Solutions must be user-friendly and holistic.
Deploying new technologies can significantly improve scheduling
In a previous, industry-agnostic article, we laid out how optimizing work management—starting with smart scheduling or scheduling optimization—can improve grid reliability, the efficiency of capital deployment, cost, safety, and employee engagement.1
Schedule optimization, however, is one of the most challenging optimization problems due to variations in types of work and operations. This variation makes solutions hard to generalize and therefore hard to scale. Additionally, the mathematical complexity of optimization equations and the number of decision variables mean models take a long time to run. To be truly useful, optimization models need to operate in almost real time so that they can react to changes such as employee sick days and unexpected demand surges.
While classic optimization models have been around for decades, the advent of new technologies in AI and cloud infrastructure allows for the rapid development and deployment of tools that bring deep analytics and optimization engines to the scheduling process. These tools have also reduced the cost of deploying an end-to-end schedule optimization solution and can sit on top of existing work management systems. Additionally, using AI improves the quality and functionality of scheduling in a number of ways:
- offering the most optimal solution given a range of interdependent constraints and dynamic, ever-changing demand
- providing a consistent, systematic approach with no human bias
- delivering significantly faster computation than manual processes, which improves the ability to adapt to unexpected changes in operations
- lowering HR requirements, which frees up capacity to focus on other areas
Smart scheduling offers benefits for utility companies
For electric and gas utilities, scheduling is a central function that matches demand for services with the crews, materials, and equipment needed to perform those services. Utilities have a variety of different work types—including emergency jobs, short-cycle jobs, and long-cycle jobs—with varying scheduling dynamics. Smart scheduling provides benefits for each work type:
- Emergency jobs have high importance but low predictability and may require a crew to be immediately reallocated from another work site. These schedule “break-ins” require real-time juggling of crews and often cause churn and rework for schedulers. Smart scheduling can help block off capacity for these emergent break-ins via dynamic schedule loading. For example, only 60 to 70 percent of capacity may be allocated in a given week if algorithms predict, based on historical data, that 30 to 40 percent of time will need to be spent on emergency jobs. Smart scheduling can also identify the optimal crew to address the emergency job based on factors such as geographic proximity and the priority and state of the crew’s current job.
- Short-cycle jobs can typically be completed within the day. They range in complexity: some jobs may require one crew for an hour or two, while others—such as hydro-vacuum excavation—may require several crews for a full day alongside coordination with third-party contractors. The scheduled duration for a short-cycle job may often be several hours more or less than the actual requirement, leading to either a schedule backlog or underutilization. Smart scheduling can better estimate the durations of these jobs using a combination of historical performance and factor-driven adjustments. For example, data on local soil composition can be used to estimate the time needed to dig.
- Long-cycle jobs may require multiple days to complete, and the main challenge is to ensure continuity by scheduling the same crews for the whole duration. These jobs often come with multiple crews and pieces of equipment, plus third-party contractors, which means that smart scheduling can ease the significant mental burden on schedulers.
Schedulers need to coordinate the availabilities of crew, materials, and equipment ahead of time to ensure that all components are ready on the day when the work is to be done. Depending on the type of job, schedulers may need to create crews of different sizes—generally one to four full-time equivalents (FTEs). Additionally, crews may be qualified only for certain types of jobs, and some jobs (particularly those related to electrics) may also require materials that are not in stock and that have a long lead time once ordered. Most gas jobs, on the other hand, can be done with the materials readily found on trucks. Finally, jobs may require special equipment such as backhoes or diggers.
One of the largest pain points for crews is job delays or “false truck rolls,” which occurs when a job cannot be started or completed on time due to the unavailability of the right crew, materials, or equipment. Smart scheduling can help ensure all job components are ready before jobs are incorporated into the schedule.
The tangible benefits of smart scheduling for a US utility
In our previous article, we laid out the significant, tangible benefits accrued by a US electric and gas utility after it piloted a machine learning–based schedule optimizer2:
- Lowered HR requirements for scheduling. Scheduler productivity increased by 10 to 20 percent, which is the equivalent of freeing up one to two scheduler hours per day.
- Increased automation for flexibility. AI models automate initial schedule builds and ongoing optimization and can react to changes in the system (for example, COVID-19, seasonalities, or workforce changes) within one to two days. Manual schedulers may take much longer to adjust to such shifts.
- Reduced waste. Over the six-week pilot, dynamic schedule loading and a decreased number of prematurely scheduled jobs meant that break-ins were down by 75 percent, job delays by 67 percent, and false truck rolls by 80 percent.
- Increased crew utilization and field productivity. Prior to the pilot, crew members at one of the utility’s sites spent 44 percent of their time actually working on jobs (as opposed to being unassigned, training, or traveling). In the automated, optimized schedules, crews could expect to spend 65 percent of their time on jobs. Overall, the pilot achieved an approximate 20 to 30 percent increase in field productivity (Exhibit 1).
Five lessons for utility players in deploying a smart-scheduling solution
Based on our experience, there are five core lessons to keep in mind during the development and deployment of smart-scheduling solutions in a utility context.
1. Data are crucial but should not be a barrier to starting
Many utilities often delay analytics-based scheduling efforts due to a lack of trust in data quality. Most leaders have a misperception that data need to be rich and easy to digest to begin to get value from AI-based tools, but the opposite is true: a small amount of data can yield disproportionate insights. In fact, new data-processing methods can take existing data and make them usable for AI models. To achieve optimal results, utility players will need to map their data landscape and find resolutions to any issues that compromise the quality or usability of the data (see sidebar, “Mapping the data landscape”). This process frequently highlights the relative importance of specific data that can then be prioritized for better data governance and stewardship, which can further increase the accuracy of AI outputs.
These processes can be conducted in as little as three weeks, but space must be built into any smart-scheduling rollout timetable. This time is used to prepare and process data related to timesheets, HR, and job backlogs, as well as to evaluate data quality and to run preprocessing modules to prepare data sets to be used by the AI engine.
2. Technology must work in conjunction with processes
Smart scheduling will be effective only if it works for the end user. Therefore, new technologies must be developed in conjunction with efficient scheduling processes, which should be codesigned with frontline employees. Digital tools, such as smart-scheduling engines, codify the underlying processes, meaning that organizations that do not optimize their processes in tandem with the development of technical tools are at risk of codifying inefficiencies.
Getting the most out of new digital tools may require some or all of the following process improvement initiatives:
- clear job readiness checklists that take into account the specificities of each electric or gas job
- break-in management processes that reduce nonemergency break-ins and quickly reorder the schedule to address emergencies
- efficient handoff meetings to align stakeholders such as schedulers, field supervisors, and warehouse managers
- prejob walkthroughs to ensure site readiness
Additionally, the successful deployment of digital tools requires continuous maintenance of the technical models. This work will require a number of different skills profiles, including capable data scientists, data engineers, and cloud engineers.
3. Businesses must clearly specify their optimization criteria
A smart-scheduling engine can optimize frontline schedules based on several evaluation criteria. For example, the engine could maximize the number of jobs scheduled, minimize operating costs related to shifts or travel time, or maximize service levels by reducing customer wait times. To achieve the best results, it is imperative that business leaders feed clear objectives into their smart-scheduling engine.
To achieve the best results, it is imperative that business leaders feed clear objectives into their smart-scheduling engine.
4. Piloting, followed by intentionally scaling, a “light tech” scheduling solution is vital to increasing adoption
Smart-scheduling solutions can be developed as “light tech” overlays on top of existing systems and do not require platform overhauls (“heavy tech”). Algorithms can often be tested in an isolated testing environment to pilot the efficacy of a scheduling optimization solution. This piloting, which should be done in conjunction with schedulers, is essential to train the model for the specific utility company and context. For example, variations in regulations or union-specific requirements can have a significant impact on the details of an optimized schedule.
A key metric during the pilot period is the frequency of manual schedule overrides by schedulers. These overrides can indicate an issue with the underlying model and should therefore happen as seldom as possible. However, some manual intervention will always be required to address last-minute contextual changes such as sick days or employee holidays.
In our experience, it takes at least four to six weeks for smart-scheduling algorithms to reach a 70 to 80 percent match with the final schedules previously created by schedulers, as measured by the percentage of jobs and crew pairings that are the same in each (Exhibit 2). While an optimal schedule is unlikely to exactly match the existing manual schedule, a relatively close match is a good indication that the new algorithm is factoring in the right parameters and will not require frequent manual overrides.
Pilots can also be an important way to build support for the new scheduling methods within the organization, which can make the subsequent rollout easier. In the US utility example used above, schedulers—who were spending four to seven hours a week building and updating the manual schedule—saw that the new technologies could build automated schedules that closely matched their own within minutes.
After a successful pilot, it is important to execute a well-thought-out scale-up plan. This plan should take into account factors such as overall deployment speed, deployment across work types (that is, there may be different considerations for electric versus gas jobs or for short-cycle versus long-cycle jobs), the differing challenges of rural and urban service centers, and resourcing the scale-up (for example, potentially hiring change champions or trainers). Tools and processes can be scaled up in an agile fashion because making iterative improvements over time is generally preferable to trying to perfect the algorithms during the pilot period.
5. Solutions must be user-friendly and holistic
To operate successfully, schedule optimization needs to be integrated into a user-friendly, end-to-end digital solution. The final, holistic solution must update constantly, forecast accurately, and incorporate an easy-to-use, interactive front-end interface. Many schedules are currently based in Microsoft Excel, and the benefits of the automated, AI-optimized schedule can be multiplied if companies can incorporate features such as daily or weekly drag-and-drop schedules and a metrics dashboard.
Scheduling interfaces should incorporate user-friendly features, which could include the following:
- preloaded, optimized schedule
- prioritized work orders
- simple and real-time edits
- precise information display
- flexible crew management
Inflation, supply chain issues, and ongoing labor disruptions are making work optimization—which has long been one of the most challenging problems for consumer-facing industries such as utilities—even more complex. When deployed thoughtfully as part of a holistic solution, AI-driven schedule optimizers can significantly improve work management processes, smooth out operations, and boost overall productivity.