Optimize power restoration crew productivity during storm, a snippet of GenAI

GenAI is a powerful technology due its ability to naturally interact with work and asset management field crew assisting their needful actions by producing outputs in text, SAP transactions, image, video and other formats. To make the base model “production ready” in an organization context, LLMs need fine tuning with specific contextual information and human intervention to train the model.

One critical real-time business scenario of GenAI is effective storm management by optimizing wrench time of field crew. Here is how,

Effective Storm Management

During the storm and emergent outage situations, power restoration is the prime priority for field workers. Any innovative technology that can maximize wrench time would be highly beneficial. The GenAI technology assist storm field crew by allowing them to start a chat dialogue from field mobile application by asking questions in a natural manner. Combining the language-based query with application entities like Equipment#, Job ID# and Current Location provides the ability to execute queries against enterprise-structured databases.

For Example - In the event of field crew needing help of tree trimming crew (Vegetation Management), all he needs to do is open chat bot and request for Tree trimming.

Behind the scene , the GenAI should be able to identify the user and associated WO based on current location. Further, update the Work order to include additional operation for "Tree trimming assistance" and auto dispatch to nearest crew on priority.

The GenAI functionality gives field crew the ability to combine structured work and asset management data with unstructured data, such as additional crew, unplanned materials, safey safety protocols and standard operating procedures, to provide a customized, context-sensitive input to an LLM and use it for its summarization and reasoning capability to generate a response.

It also minimizes data traffic between field mobile device and the back end ERP cloud by summarizing the information, resulting in lower capacity consumption over wireless networks.

The ability to deliver asset information to crews, in an easy-to-comprehend manner, in near real time, will improve crew safety and productivity. Another potential benefit is that field crews get expert-level answers without having to get a human expert online.

How to develop GenAI prompts -

To maximize the benefits of AI, it's crucial to grasp the art of crafting prompts that yield the most valuable outcomes leveraging detailed prompting, context and data base.

?

1. Detailed Prompts:

Optimal results are achieved when prompts are lengthy and detailed. Investing more time in clearly articulating your expectations for the AI enhances the quality of the outcomes.

Providing context to the prompt can significantly impact the outcome. Train a Generative AI model using the structured data. This involves exposing the model to a wide range of scenarios to enable it to recognize patterns and correlations.

Example : Generate a spreadsheet list of poles and pole assemblies under a given tax jurisdiction code with a first column as Serial number, followed by Equipment ID, Object type, and Installation date

?

2. Prompting with Context :

Providing context to prompt can significantly impact the outcome. Initiating the conversation with details such as "you are a storm crew, working on Work order# "coupled with relevant background information such as storm ID, current operation sets the stage for a more informed and tailored response to the inquiry. Identify key features in the data context that are indicative of location, tax jurisdiction, city, county etc.?

Example : Generate top 10 distribution poles in the current location in the descending order of number of impacted customers.

3. Dynamic data lake for AI :

AI queries data sets of various sources including SAP, Non-SAP, GIS, Outage Management Systems (OMS), Transaction information such as material reservations, work order assignments and proximity field crew . It's critical to continuously pool data and consolidate data indexes to feed to AI prompting. Gather historical and real-time data related to storm restoration patterns, commonly used materials and relevant crew factors. Ensure the data is diverse and comprehensive, covering various types and intensities of storms

Example : Generate top 10 distribution poles in the current location in the descending order of service priority (Hospitals, Fire stations, EMS, Large commercial and industrial) and number of impacted residential customers

In the advanced GenAI , one should be able to develop algorithms to assess and predict the likelihood of seeking tree trimming crew, critical materials and switching operations. This predictive algorithm can further lead to a decision support system that provides actionable insights to emergency responders, government agencies, and the public. This can aid in planning evacuation routes, resource allocation, and other preparedness measures.

Bonus is real-time updates to customers on restoration work via a simple AI Prompt of send communication to all connected customers. This can be done by Implementing a Power Alert System (PAS) via a SAP CRM , one can establish a communication system that disseminates timely alerts and information to the public on power restoration services on a real time basis.

Leveraging generative AI in work and asset management can improve workflows and decision-making processes, tackling tasks that were once cumbersome, intricate, or seemingly impossible.

Future Innovation : The evolution of cloud-based Generative AI for Utilities is ongoing, with capabilities expected to expand further. Integrating this technology into business applications such as S4HANA Work and Asset management holds the promise of addressing current challenges and enhancing crew productivity. The transformative potential of this technology in reshaping field service and asset management programs is readily apparent.