9 MIN READ
To guide their companies through economic shifts, senior executives must understand how technology supports business strategy. Today, that awareness includes artificial intelligence (AI). But without business context, AI will deliver limited value.
Forward-thinking executives are discovering that a complementary technology can contextualise AI information and help them discover new markets, better understand and serve customers, improve asset management, optimise their supply chains, and mitigate risk.
Most C-suite executives sense that AI is a crucial aspect of digital transformation. Three out of four business leaders say they are working to integrate company-wide AI initiatives, according to an Accenture survey. The same percentage believe that if they fail to do so, their companies will be defunct in five years.
The next half-decade will be crucial. McKinsey research suggests that the AI adoption rate among corporations will rise steadily during this period—then increase significantly.
The Business of Machine Learning
In a business context, AI often refers to machine learning. By studying enormous datasets, this type of AI learns to spot patterns, draw conclusions, and even offer predictions. Around one in five C-suite executives say they already use machine learning in some capacity.
What these early adopters—and those who follow—will discover is that the insight generated by machine learning, no matter how advanced, is not enough to accomplish strategic goals. For an AI program, the world is numbers, images, and patterns. For an executive, the world is a place where events occur in certain locations at specific times. Even the best machine learning analysis needs real-world context.
Location intelligence (LI) provides the foundation for that context. Recent technological developments have brought LI and AI together in a hybrid sometimes called GeoAI. Geographic information system (GIS) technology generates location intelligence, and through GIS, executives can bring geographic context to AI and discover insight on key decisions.
The combination of machine learning and location intelligence helps executives drive strategic initiatives including market growth, customer engagement, and risk mitigation.
Predicting and Discovering New Markets
According to a Gartner survey, 53 percent of CEOs list growth among their top three priorities. And while the COVID-19 pandemic has idled many growth plans, some executives are already planning for recovery. Once that phase arrives, companies that most effectively use insight and predictive capabilities will outpace their industries, according to McKinsey. An AI-fueled approach to location intelligence can support those plans and foster growth in three ways: by automating processes, detecting patterns within large sets of data, and making predictions.
Imagine a bank’s planning committee analysing the expansion possibilities in a particular city. Using GIS-based machine learning, they can explore the sales data of existing branches and the demographics of surrounding neighbourhoods—along with the site’s proximity to the bank’s own branches and competitors’. The machine learning program predicts locations that would yield the highest returns with the least cannibalisation of business at existing branches.
Renewable energy companies already use machine learning and location intelligence to understand potential markets. Decision-makers at a solar company, for example, can use AI-based image detection to assess a neighbourhood’s viability for solar panels, even crunching historical weather data to calculate the ideal placement of roof panels for maximum sun exposure.
Virtually any business can employ a combination of AI and LI for growth planning.
Uncovering Hidden Patterns in Market Data
Another combination of AI and LI uncovers patterns and offers predictions so executives can discover trends and hot spots lurking in enormous datasets.
A wireless service provider, for example, can combine machine learning and location intelligence to detect trends and guide business investments. Company planners might perform a GIS-based cluster analysis of dropped calls to identify areas where additional cell towers are needed. When they add demographic predictions to the maps, executives can see where population growth will create pockets of potential new subscribers. This allows the company to prioritise network build-outs to meet demand.
Prediction represents the cutting edge of machine learning and location intelligence. Analysts have even used GIS-based machine learning to solve one of the most complex phenomena in retail—the so-called “halo effect”—in which a store’s sales and nearby online sales are mutually reinforcing.
Through a process called “quantifying proximity,” location technology reveals where physical stores will maximise online sales. To do this, a GIS-based machine learning program models the likely performance of stores in different geographic areas—neighbourhoods in a city, for example, or cities across the country—by analysing factors such as demographic information and the travel time of likely customers.
This reveals which locations will produce the most successful stores. Planners can then perform further analysis of online spending patterns in those areas, producing a smart map that highlights which physical locations will experience the greatest halo effect—with revenue predictions for each location.
Another form of AI-based trendspotting applies to the insurance industry. GIS technology can create a training data set of past accidents with details on location, time of day, and road conditions. A machine learning program analyses the data and indicates the probability of accidents on specific streets at specific times. The insurer can then advise customers on alternative routes to reduce their risk.
Insurers are also using a combination of AI and LI to speed relief to clients whose homes have been damaged by natural disasters. After training on large sets of images, an AI program learns to distinguish a damaged property from an undamaged property, allowing insurers to automate and accelerate the delivery of payment, which traditionally took days or weeks to complete.
Understanding Customers and Competition
Every executive takes pains to understand a brand’s customers; location intelligence strengthens that effort.
Mobile phones, combined with the world’s 30 billion IoT sensors, have created an avalanche of anonymised human movement data that helps retailers understand their customers better. Add psychographic and demographic data—aggregated at the group level, not by individuals—and the result is a bevvy of information with useful truths about customers. But that data is often too vast to explore without machine learning.
To better understand how location intelligence adds context to machine learning, consider some of the questions executives ask when planning brick-and-mortar locations:
How easy is it for someone in one area of the city to reach a particular store?
How far will customers from other locations travel to visit a new store?
To what degree is a store accessible by the people likely to be its customers?
If the customers of one store are demographically distinct, where should the company place another store to attract similar customers?
How do these calculations vary by time of day or day of the week?
In the pre-AI era, planners could comb through data to answer those questions, but that often took months. Machine learning algorithms instantly run these calculations and display results on GIS dashboards.
With a detailed understanding of customers, retailers can adjust the merchandise they carry and how they display it. At one sporting goods store, a robot scans RFID-tagged products on shelves throughout the day and creates an updated inventory map of the store. The robot’s machine learning technology notes patterns that help workers stock products based on demand. Meanwhile, stores that offer pick-up services—which have become ubiquitous during the COVID-19 outbreak and may remain popular afterward—are using location data to predict when a customer will arrive to retrieve an order—of coffee, groceries, electronics, or furniture.
Understanding customers also means understanding the competition’s customers. For instance, one AI-based program analyses aerial imagery and classifies the types of cars in a parking lot, revealing information about the demographic groups who patronise a particular store.
A similar brand of intelligence can come from analysis of much larger areas. An oil-and-gas company could discover exactly where its competitors are working, anywhere in the world, by using a location-aware machine learning program to spot drilling rigs, well pads, and site activity from satellite imagery.
Asset Management—in Less Time
With machine learning and location intelligence, a company can understand the real-time conditions and locations of its assets. One utility company offers a striking example of the dividends this can yield. Under innovative leadership, the company enlisted drones to monitor the condition of poles and transmission wires, taking LIDAR-based measurements along the length of its network.
This left specialists with the daunting task of reviewing those measurements manually. Instead, the company added AI to the mix. Analysts taught a machine learning program to detect the power lines and any intrusion—for instance, vegetation—on the lines’ safety corridor. The utility company’s machine learning program analysed 10 billion data points and accomplished, in short order, what the company estimated would have required 50,000 hours to do manually.
Similar innovations are happening at companies large and small, with modern GIS technology that features built-in AI capabilities. Delivery companies, for instance, can use machine learning to analyse drivers’ routes and plan better ones based on real-time and historic traffic conditions. Rail companies can use remote sensing technology like LIDAR to detect track problems, while machine learning programs can sense conditions that will likely damage train wheels.
The value of location-powered AI is not merely in monitoring the current condition of assets, but in predicting what they will be.
Risk Measurement, Risk Management
Economist Avi Goldfarb says a reluctance to take risks creates untapped potential for a business. Machine learning’s predictive capabilities can lower the cost of risk, thereby unleashing that potential. The effect on a business can be transformative, Goldfarb told Harvard Business Review in 2018. Companies that integrate machine learning and location intelligence will have an even better sense of how to take measured risks.
Consider climate change. As Grant Mullins of Regions Financial Corporation noted in a recent WhereNext article, companies such as Regions are increasingly concerned about the connection between climate change and environmental, social, and governance (ESG) risk.
Floods, high winds, and rising temperatures and sea levels will affect road conditions and building integrity, eventually driving populations out of areas that become disconnected or uninhabitable. This will have a profound impact on business operations—affecting everything from a company’s office and retail locations to its supply chain and marketing practices.
AI-based programs can process this big data and create predictions of weather impacts in coming decades. GIS-based smart maps contextualise those predictions, allowing business leaders to preview coming years and see how climate change will affect specific business locations—including offices, stores, distribution centres, and infrastructure. In this way, machine learning and location intelligence provide executives a risk map that guides investment decisions.
Company leaders will increasingly discover how machine learning and location intelligence can guide the allocation of resources to minimise risk. An insurance company concerned about liability can use a GeoAI program to analyse weather history and topographic features and predict which homes might see an increased risk of flooding. An agribusiness can comb through millions of data points from soil samples to historical yield data and sales to determine how much of a certain crop to plant where.
AI and LI for Strategic Context
With smart maps that add context to AI-fueled insight, companies become better positioned to achieve strategic growth, strong customer engagement, efficient asset management, and more accurate risk assessment.
Amid the uncertainty stoked by population and climate changes, shifting customer expectations, and variable economic conditions, business leaders can count on technology to evolve, too. Those executives who stay informed, especially of advances in AI and location intelligence, stand to gain a competitive edge.
Learn how else growth opportunities can be identified before they emerge at the Esri Singapore User Conference 2020.