Automated decision making is broadening its scope and capabilities. The Internet of Things (IoT), with its sensor capability and the ability to collect, share, and transfer data using the Internet, is creating many new possibilities. Decision automation is often integrated with Internet-connected devices. Creating a data-intensive environment in a hospital, retail store, home, or other discreet area where sensors monitor changes in state provides many possibilities for decision automation. The goal of a sensor-rich intelligent environment is generally stability and homeostasis. An ambient intelligent environment refers to a pervasive computing environment that enables interaction with, and appropriate responses to, the people in that environment.
Manufacturers are adding sensors to product components to transmit data about how they are performing via automated monitoring software. Decision automation, sensors, and embedded devices create opportunities for intelligent manufacturing, personalization for customers, field service automation, industrial system consolidation, and robotic assembly.
IoT devices, sensors, digital identifiers, and monitoring devices can serve several functions, including providing input for decision automation and initiating changes. For example, these devices can improve:
Identification. Radio-frequency identification (RFID) with decision automation can automatically identify and track tags attached to objects. RFID helps in updating inventory, self-checkout, and loss prevention.
Monitoring. Devices transmit data in real time, including video cameras, motion sensors, and temperature and moisture sensors. Uses include healthcare, security, and surveillance.
Location services. These include GPS locator, finder, and tracker devices. Satellite-based radio navigation can help locate, find, and track people and other mobile things (e.g., cars, trucks, pets, kids, and people with disabilities).
Control of distant objects. Such devices include teleoperations, telepresence, and telemonitoring with remote robotic systems.
These systems can provide security, deliver in-home patient care, and provide remote consultations and training, among other uses.
Change often creates challenges. Inherent in many challenges are one or more opportunities. Resolving challenges associated with implementing decision automation and sensors can help identify opportunities for digital transformation and operations renewal. Managers must assess what is needed, what is cost-effective, and what is most useful with new decision automation technologies.
The IoT, when combined with decision automation, can be transformative. The following are five important challenges that, when resolved, may result in innovative opportunities from adopting automated decisioning technologies:
Managers must prioritize possible decision automation applications and manage the integration of rapidly obsolete technologies over time. They must identify the “low-hanging fruit” to show the potential business value of decision automation, but also have a broader deployment plan.
Managers must secure automated decision environments from internal and external disruption but remain flexible enough to manage new sources of data. Security breaches are increasing, and the growing complexity of networked environments creates vulnerabilities. The major opportunity from resolving the security challenge is the possibility to monitor and control decision automation from anywhere at any time.
Managers must maintain data control, retrieval, and storage for a decision automation application. Meeting this challenge is important for auditing decision automation implementations. IoT devices generate high volumes of data for decision automation often captured in real time. Managers need to develop strategies for these new sources of data.
Many managers seem to feel that decision automation with chatbots, IoT, and analytics will increase customer engagement. The rationale for deploying decision automation must be supported by a robust business rationale, and appropriate resources must be allocated to the project. There is a danger that decision automation will be poorly conceived and deployed which could ultimately alienate rather than engage customers. For example, going through many layers of a hierarchical information menu using a phone can be annoying.
Managers want to use more data with the goal of making better decisions. The challenge is identifying the rules and monitoring the decision outcome variables. Making better decisions is an ambiguous goal. It is far more likely that a well-designed decision automation system with IoT will result in more systematic, faster, and evidence-based decisions. The rules used in decision automation and the assumptions made linking data to decision automation systems will ultimately determine the quality of decisions. Managers should be prepared to monitor multiple decision outcome variables, like speed and quality, to determine if an automated decision is better.
With an expanding role for IoT, machine learning, and algorithms in decision automation, it is important to assess the role of people in networked, distributed decision automation settings. Some advocate for a hybrid person-computer approach to decision making when there is limited or ambiguous sensor data. In these situations, shared decision making and decision support are frequently superior to decision automation. Decision automation should have rules to involve people when creativity, assumptions, nuanced facts, and judgment are required. Decision automation systems should have referral rules about when to consult a person.
In some parts of the world, the IoT and decision automation are helping to create smart cities with smart buildings. These automated, data-intensive environments are self-regulating and self-sustaining. These environments offer the potential to reduce resource use and increase quality of life for people. The IoT and decision automation are expanding, and that expansion and the increasing sophistication of IoT devices create new opportunities. Algorithms and analytics can help make sense of machine data and can use that data to automate and support decision making. The right mix of people, AI, analytics, and smart things is important to operations and organization success.
[For more from the authors on this topic, see “Decision Automation: Challenges and Opportunities.”]