Transforming field services through AI-augmentation amidst skill shortage
Intro
Europe’s renewable energy installation sector is currently facing a massive shortage of skilled field workers. In these challenging times, the future of field services may very well be boosted by AI augmentation. With current technology advancements especially in Generative AI, applying AI augmentation for field workers does suddenly not feel far away anymore.
The biggest barrier today is a data challenge for existing field service companies before they can leverage AI. In this blog post, we’ll explore how AI augmentation can bolster the efficiency of field workers, and the steps necessary to achieve this.
Please note that by “field workers”, I refer to individuals who perform tasks on-site, such as technicians, engineers, inspectors, maintenance crews, and service providers. If you have any input, feedback or if you are building in the same area, I’d love to chat.
The climate crisis forces us to rebuild our infrastructure
To combat the climate crisis in Europe, we are in the process of completely rebuilding our infrastructure. This means, we are installing solar panels, heat pumps, energy storage systems, EV charging stations, sustainable housing units, carbon removal solutions and much more. Given how fast climate change is impacting our environment, we need to do that as fast as possible. A good example of this is the sharp increase of newly installed capacity per year for solar in Europe and the U.K.:
Lack of skilled worker shifts the market towards specialization
One of the main barriers though is the lack of skilled workers in Europe who can do all of the physical work such as the installation and maintenance of these assets. Germany alone currently has 216,000 jobs unfilled in the wind and solar industry. In many European countries, the workforce crisis will only get worse in the next few years, putting the energy transition targets until 2030 at risk.
Due to the lack of skilled field workers in the renewable energy industry, we see a market shift towards a more specialized value chain. The specialization of roles does not solve the labor shortage but it allows greater efficiency, hence a greater number of installations.
An example of the specialization of roles is that electricians focus only on the electrical part of a solar installation and not on customer acquisition, planning, finance, admin, or supply chain.
Field service companies are trying to increase the labor pool
While the specialization of roles increases overall efficiency of skilled labor hours, field service companies are trying multiple solutions to increase the labor pool. I’m mostly seeing a combination of the following three:
- upskill / reskill field workers themselves
- hire field workers from abroad
- collaborate with local subcontractors
Independent of these solutions, field service companies are facing the same challenge: they often need to work and collaborate with less experienced field service workers in one way or another. This can lead to suboptimal results, costly re-work, and inefficient collaboration.
We’ve learned that to get to maximum efficiency in field services, knowledge needs to be embedded in the processes and software as this leaves no room for mistakes even for inexperienced field workers. In the past, some field service companies have been using static guidelines, rigid online forms and standard operating procedures (SOPs) to ensure field work gets done efficiently. However, with recent technology advancements in AI, this can go much further.
Data as the basis for automation and AI
The basis to enable automation and AI application is the underlying data. PVCase, an energy modeling software for solar PV systems, rightfully pointed out the “data risk” that is hindering developers of large solar systems to design faster and more efficiently. Designers and engineers need to work with multiple systems and different data sources, some even in older formats, such as PDFs and hard copy records.
Field service companies are facing the same “data risk” challenge. Even today, field service companies often work in different siloed systems (think CRM, ERP, field service software, custom built solutions). A scattered data landscape makes it impossible to automate processes or enable AI applications. Moreover, it is important to collect the right data. For field work that means data about the field work and processes itself, ie. what and how the field worker is doing her job.
The data challenge actually reminds me a lot of my work on Industry 4.0. In the manufacturing industry, many companies wanted to automate processes in the past few years before doing the groundwork, ie. connecting machines, gathering process data from the shopfloor and bringing all data to the cloud.
How to get to AI augmentation
To get from manual inefficient field services to AI augmentation, field service companies need to take the following 4 steps:
1. Define a structured, repeatable process
Creating structure and defining field service processes is the first step. Companies need to define instructions on how to execute field service processes, ideally supported by forcing functions such as quality gates. Most of the field service companies are doing this by using digital forms and standard operating procedures (SOPs).
2. Capture process and image data
Capture performance data of field work is essential to know how field work is currently being done. Nothing is worse for a company than the lack of transparency of its processes and quality of field work. If work still happens on paper and PDFs, it is impossible to drive any conclusions about how the work is getting done. So the more data points companies can capture of how their field work is getting done — including image data — the better.
In addition, it is important to ensure data consistency. This means companies need to work on building integrations of different systems so that they always have accurate data at hand. Manual data entry from a CRM to a field service software is a nightmare as this leaves so much room for mistakes, especially if data is updated in one of the systems later on.
3. Human-guided field worker augmentation
Once structured data is captured in a digital format and data consistency is ensured, we see big efficiency improvements when large solar installers apply manual quality assurance to on-site visits. Field workers are not allowed to leave the site until all field work is checked and manually approved by an office worker. At first, this might sound cumbersome and time-consuming. However, the efficiency gains are way higher than respective costs and can avoid costly re-work and unnecessary site-visits. It’s important to implement feedback loops to capture and analyze common mistakes to improve field work processes over time (also known as “Kaizen” in manufacturing).
4. AI augmentation
The goal of AI augmentation for field workers is to automate quality assurance and admin tasks with AI. According to Bain, workforce augmentation can lead to a 21% to 30% reduction in costs.
AI augmentation can enhance field worker’s skills, streamline their workflows, and improve their overall performance. This means that software can transform inexperienced workers to field service experts. Imagine a world where a young technician shows up to a wind power turbine and gets briefed on the full technical specs and is guided step by step through the job — reducing training time, enabling higher worker utilization and reducing errors. In the short term, I’m very excited about the following two use cases.
Use cases of AI augmentation for field services
- Data collection and analysis: AI can automate data collection processes, enabling field workers to gather and analyze information more quickly and accurately. This could involve using AI-powered sensors, cameras, or devices to collect data, and then leveraging AI algorithms to process and extract meaningful insights from that data. With the advancements of LIDAR sensors in smartphones, companies such as magicplan already use that data to create a “digital twin” of a construction site or room. This might even make it possible to “outsource” on-site visits for residential solar or heat pump installations to the end-customer.
- Real-time decision support: AI can provide field workers with real-time guidance and recommendations based on data analysis, historical patterns, or predefined rules. This can help field workers to make informed decisions and take appropriate actions promptly while on the field. One example is real-time decision support when documenting all the data necessary for the planning of residential solar panels, also known as quality assurance.
Why now
Current advancements in AI, particularly with generative models, are fueling opportunities to augment field workers.
By leveraging the rich reasoning and language understanding of LLMs, companies need far fewer domain specific training examples to deliver valuable AI assistance to their users. In addition, diffusion models — a variety of generative models that’s useful for multimedia — can now generate realistic and synthetic data, enabling field service companies to train AI models on diverse datasets and augment their field worker’s decision-making processes.
Generative AI also enables field workers to create natural language responses and images. This helps them to enhance communication, documentation, and reporting processes. Advanced language models, such as OpenAI’s GPT-3.5/4, have made significant progress in generating coherent and contextually relevant natural language responses. This capability is leveraged to augment field workers’ communication by generating accurate and appropriate responses to inquiries, assisting in field service work, or automating routine documentation tasks.
With more innovation and technology advancement in Generative AI and large language models on the horizon, it’s just a question of time how fast AI augmentation will find its way into every software product.
Recent examples of AI augmentation for field workers
As an example, Salesforce recently announced that by the end of 2023 its new Field Service solution will leverage AI powered by Einstein GPT. Einstein GPT is a generative AI technology that combines Salesforce proprietary AI models with generative AI technology from partners and real-time data from the Salesforce Cloud. Einstein GPT will help field workers to automatically compose reports and generate step-by-step guides to address technical issues.
With Kwest, we are currently looking into ways how algorithms can do the quality checks automatically to give field workers feedback in real-time. This would avoid the need of having office workers manually checking all the work done by field workers. At the same time, we could make this available to field service companies that cannot afford to hire dedicated office workers to do manual quality checks.
Take-aways
The goal of AI augmentation for field workers is to empower them with intelligent tools and capabilities that enhance their skills, streamline their workflows, and improve their overall performance. Embedding all the knowledge into software will help even inexperienced field service workers to do the job at the same quality as experienced field workers.
By leveraging AI technologies, field workers can benefit from increased efficiency, reduced errors, better decision-making, and enhanced collaboration, ultimately leading to higher productivity. To make that work, it’s important that we work with urgency to solve the data challenge for field service companies. Creating data generating processes, ensuring data consistency, and capturing performance data of field work is essential.
Only then, we can start to leverage the technology advancements in AI to drive efficiency to field services, especially in a time of a severant labor shortage.
