Artificial intelligence will impact every industry and every business discipline — and that includes field service management. But how quickly will practical solutions be available that enable the typical medium-to-large field service organization to take advantage of AI? Let’s take a look at AI solutions already helping improve processes, deliver knowledge efficiently and automate repetitive activities to free human workers to focus on more personalized service and complex problems.
The future of artificial intelligence holds much promise for enterprise systems and we are already beginning to see its significant impact in automation. In three areas of field service, there are already commercially available, practical AI solutions delivering real business value: enhancing customer interactions, enabling management by exception and quickly masterminding complex scheduling.
AI for customer interaction
First impressions matter, but the initial interaction a customer has with a service organization often involves several missteps. Chief among these is long on-hold wait times. Customers are also reaching out through multiple channels including email, chat and social media; when these data streams go into disconnected siloes, the result is disjointed communication.
Today, AI solutions can solve both these problems — but it requires more than just chatbots. Commercially available AI software, which ties into chatbots is capable of learning which answers are appropriate for each question and automating a significant majority of chat interactions.
A chatbot can be taught to answer commonly encountered questions, such as inquiries about when a technician is scheduled to arrive. This enables customers to have a digital, convenient experience in their own time, while also providing call-center efficiencies and the ability to use your agents to deliver a personalized service.
AI alone can handle, typically, between 50 and 60 percent of requests, but at some point, an AI chatbot may get stuck. This is when personalized service is required, with a human agent taking over the discussion thread without missing a beat. It should be seamless not only to the customer, but for the internal customer service: ticketing and support systems.
AI-based chatbots can enable a good agent to handle up to five or more chats at a time. It can consolidate Facebook messages, emails, telephone and tweets and direct them to an agent or AI for intervention. The beauty of this AI functionality is that it learns from answers provided by human agents and gets better at answering questions.
Integration between an AI chatbot, email, voice, social and enterprise applications is important for another reason. It enables one version of the customer record. Lacking this, a customer can send an email and get no response. They send a direct message through Twitter. Then call and sit on hold. Then initiate a chat. All these interactions may not appear in a central customer record, yet there have been three attempts to contact the company.
Manage by exception to drive ROI
In the case of AI applications for the service organization, a primary ROI driver is that it enables humans to manage by exception. A high volume of activity can be automated, with human intervention limited to when a situation falls outside business rules or logic built into service management software. AI doesn’t eliminate the need for human interaction; it makes human interaction more focused on what humans do best — handle escalations and complex decision-making for unique cases.
Management by exception is more successful when an AI application has access to extensive information about each customer. So full integration with enterprise resource planning, field service management and other enterprise tools is essential. AI tools can be more effective if they have more rather than less information on the status of the customer’s account, including their maintenance or service history, warranty or service-level agreement entitlements.
Human agents can excel in serving customers directly but, in the case of scheduling technicians in the field, humans are sometimes just not able to manage the constant numerical challenge of optimally adjusting a schedule.
Manual or traditional software-based scheduling may be a workable solution for service organizations with a very small number of technicians engaged in a small number of jobs. But it does not take many technicians or jobs for the number of possible solutions to outstrip human computation capabilities, either individually or as a group.
A dynamic scheduling engine driven by AI algorithms is designed to solve complex scheduling problems in real time — problems much too complex for any human dispatcher or customer service agent to handle. Even at the low end of the spectrum, a human dispatcher cannot quickly identify all the possible solutions and pick the best one.
Two technicians and five service calls yield 720 possible solutions. Four technicians and 10 service calls present a dispatcher with 1,037,836,800 possible solutions. By the time you get to five technicians that must complete six calls each — a total of 30 calls — you have 12,301,367,000,000,000,000,000,000,000,000,000,000 possible solutions!
Finding the optimal solution becomes even more complex as rapidly changing factors are added into the mix — including emergency jobs, service level agreements and other contractual requirements, technician skill sets, or tools and materials currently in stock on each service vehicle.
An AI-driven scheduling engine automates the schedule, making adjustments in real time based on priorities set by service organization management and real-time information. This frees up human dispatchers to manage by exception and deliver meaningful customer interaction that builds loyalty and deepens the relationship.
Service management for many businesses relies on inventory. When a service request cannot be closed on the first visit, it is often because the right part is not on the truck or immediately available.
Service management software should encompass inventory management functionality and that functionality should include automated reorder points for each part. The ability to consider parts availability is a critical data set for AI to work on, as parts are a critical determinant in first-time fix and job completion. It’s also key to successful SLA and outcomes-based commercial relationships.
Once inventory data is available and integrated, better inventory logistic can be configured so parts and materials are housed in warehouses, satellite offices or inventory drop locations closer to anticipated demand, while inventory is matched to jobs in a forward or current-day schedule.
Service organizations should recognize the tremendous potential AI holds — they can harness it to transform their operations, outflank their competitors and disrupt their markets. We are only starting to tap into the different ways AI can be used to better solve the problem of delivering optimal service in a rapidly changing environment as adoption is still lagging despite the real benefits AI brings.
The good news is there are several straightforward and easily accessible ways service executives can harness AI technology today.