THE RISE OF

AI AND LOCATION
INTELLIGENCE
How organizations are using the combined
power of AI and location intelligence for
sustainability, efficiency, and growth
WHAT’S INSIDE
What Makes AI and Location Intelligence Difference-Makers...................................3
What AI and Location Intelligence Can Do for Decision-Makers.............................................................................................................. 4

Global Awareness of Change Averts the Worst of Climate Impacts ........................5

CASE STUDY: AT&T: AI and Location Intelligence Help Telecommunications Giant Get Ahead of Climate Risk ................................. 6
CASE STUDY: Microsoft, Impact Observatory, and Esri: Seeing Near Real-Time Changes on Earth..................................................... 7

Achieving Automation for Sustainable Business Operations....................................8

CASE STUDY: Finnish Forest Centre: Robots for Better Forest Management ........................................................................................ 9
CASE STUDY: Clearwater Seafoods: Data-Driven Sustainable Marine Development........................................................................... 10

Pinpointing Predictive Maintenance to Enhance Efficiency and Safety .................11

CASE STUDY: Autobahn: Bavaria Gains Insights into Roadways with Predictive Maintenance ........................................................... 13
CASE STUDY: US Army Corps of Engineers: Predicting Where to Dredge to Save $100 Million per Year.......................................... 14

Keeping Supply Chains Moving..............................................................................15

CASE STUDY: FedEx: Location Intelligence Helps Keep Shipments on Time....................................................................................... 16

Utilities Inspect Assets and Reduce Network Loss with Insights from AI...............17
CASE STUDY: Nob Hill Water Association: Replacing Old Water Mains after AI Analysis.................................................................... 18
SIDEBAR: Rethinking Utility Vegetation Management with AI and Location Intelligence .................................................................... 18

Spotting Unseen Expansion Opportunities.............................................................20

SIDEBAR: Halo Forecasting with AI: A Major Leap in Retail Planning.................................................................................................... 20

Providing an Enhanced Awareness for Sustainable Carbon Farming Practices .....21

CASE STUDY: Organic Valley: Improving Pasture Vigor with Insights Derived from Daily Satellite Imagery........................................ 22

Locating the Most Important Targets Quickly, with Urgency..................................23

CASE STUDY: The Mozambique National Institute for Disaster Management and Risk Reduction:
Aiding Flood Victims before and during Disasters with Drone Imagery........................................................................ 24

Conclusion...............................................................................................................25
“The previous generation of
GIS technology was built for
people, and now we have a
GIS built for machines.
Automation doesn’t mean
we’re replacing humans; it
means humans don’t need
to do the tedious work that
causes a state of boredom.
Data scientists can focus on
what they’re good at, which
is innovation, and machines
can then take innovations to
scale and perfect the
readings—focusing on
accuracy and outcomes.”
—Jay Theodore, CTO, ArcGIS Enterprise Development Lead, Esri

2
What Makes AI and Location Intelligence Difference-Makers
Machine learning and deep learning are types of artificial With AI and location intelligence, decision-makers can AI and location intelligence allow organizations to ask Organizations are moving quickly to gain a spatial
intelligence (AI) that involve teaching computers to learn tackle high-priority questions such as the following: these questions at a scale and frequency that could only inventory of resources and assets and are creating
from data to make predictions and guide decisions. They be dreamed of before. They have fast become the most decision support systems to safeguard against
- Where are my assets and locations in danger from
excel at making sense of large volumes of data rapidly. powerful means of gathering information, filling in interruptions. AI and location intelligence allow
that developing storm or from rising seas or other
Some of the most compelling uses of these tools involve knowledge gaps about spatial relationships to prioritize businesses to plan like never before and governments
climate risks?
split-second location decisions as in autonomous driving. what needs to happen where and when. to invest where new infrastructure is needed most to
- Where are the resources my company relies on,
defend against calamities.
When it comes to location questions, businesses and and how can we operate sustainably with the least Today we have the computing power and modern GIS
governments have relied on geographic information impact on the environment and threatened species? technology that make AI and location intelligence AI and location intelligence provide a way to automate
system (GIS) technology for decades. Machine learning, - When should we schedule maintenance on that capabilities attainable for all organizations. This ebook tasks that would take combinations of people an
deep learning, and location intelligence (powered by highway section or shipping channel to ensure the contains a wide range of examples that show how exponentially greater number of hours to complete—to
GIS technology) are now part of those workflows in least disruption and lowest cost? organizations use AI and location intelligence make sense of complex situations quickly. u
automation techniques known collectively as AI and - Where is that ship or train now, and how quickly will to capture reality, simulate scenarios, and anticipate
location intelligence. it deliver components we need tomorrow to keep the future.
the supply chain moving?
Decision-makers use AI and location intelligence on
- Where do we need to do maintenance on trees
problems that involve big datasets and complex patterns
near electric transmission lines to avoid catastrophic
that may be difficult for humans to detect. A city, for damage from limbs falling on power lines due to
instance, uses AI and location intelligence to analyze high winds or ice storms?
energy usage patterns in buildings and identify
- Where are our best customers and locations, and
opportunities to reduce energy consumption. This could
where are they likely to be in the future?
help a city meet its sustainability goals and reduce costs.
- Where should farmers and agriculture producers
adjust land use, crops, and livestock to make the
AI uses image recognition to tease out answers to
best use of their resources?
complex and high-value questions. With the explosion
of imagery coming from satellites, aircraft, and drones— - Where are the vulnerable populations most likely to
be endangered when the next unprecedented flood
coupled with the urgency to make sense of rapid
or wildfire event occurs?
changes—AI and location intelligence meet this moment
by swiftly sifting through all the pixels to find answers to
not only where but also where now.

3
What AI and Location Intelligence Can Do for Decision-Makers
Close observers of AI have noted that as an analytic tool, much of what it accomplishes falls into three main categories in which AI and location intelligence fit seamlessly:

• Automate tasks and repeat them quickly at scale. • Look at past patterns to make predictions. An insurance disparate data as weather forecasts, the likelihood of ship • Search for patterns hidden in large amounts of data.
A utility company can use AI and location intelligence, company can use AI and location intelligence to make and rail traffic backup, and how many left-hand turns there A company determining where to locate retail stores or
along with drone technology, to examine utility poles sophisticated projections of flood or wildfire damage are on driving routes. The AI and location intelligence other physical assets can search and cross-reference data
and wires, determining where telltale signs suggest based on past patterns and weather variables. A logistics program can even factor in the likelihood of social protests related to customer demographics and geography. The
repairs are necessary. or manufacturing company can use AI and location by analyzing millions of social media posts to learn if business can even use AI and location intelligence to show
intelligence to optimize supply chains, considering such unrest could impede the supply chain. transportation patterns to determine how far a potential
customer is likely to travel. u
4
Global Awareness of Change Averts
the Worst of Climate Impacts
It’s been difficult to detect changes as they happen to our planet and to Today, AI and location intelligence techniques are being applied to provide
understand their causes. But now, with growing earth observations and tools solutions to environmental problems at a scale, granularity, and speed that
that identify emitters of greenhouse gases, we can connect causes have not been possible before.
to outcomes.
Organizations around the world are gaining a new understanding of climate
With AI and location intelligence, algorithms can be trained to analyze risk and taking action where and when it’s needed most. With the new
imagery and other data sources to identify patterns and trends in awareness of AI and location intelligence, companies are tackling their
environmental and climate variables. This information can then be visualized greatest risks before impacts hit their balance sheets. Governments have
on maps and analyzed to inform adaptation and mitigation strategies. gained tools to revitalize urban areas while making them more resilient. And
natural resource managers now have the knowledge to mitigate climate
Machine learning techniques have taught computers to learn from data,
impacts on forests and water sources.
training models with datasets to answer questions. The leap to deep
learning adds further capabilities. With machine learning, you can find all Even the largest and most risk-averse organizations are acting with haste
the trees in a massive set of images. With deep learning, you can do that, when they see change coming and know ahead of time how they can
plus determine tree type and stress level related to climatic conditions to reduce harm. u
predict the severity of the next fire season.

5
AT&T—Case Study

AI and Location Intelligence Help Telecommunications

Giant Get Ahead of Climate Risk
Situation: Climate change is causing warming Challenge: Leaders at AT&T—the world’s largest infrastructure to assess how to mitigate damage from with the ability to test scenarios and see downstream impacts,
temperatures, rising sea levels, and more extreme telecom company by market capitalization—are focused intensifying storms and floods. helps organizations plan like never before. For businesses, this
weather events. Business executives are scrambling to on doing their part to mitigate climate damage even as gives them an edge to deliver goods when and where they are
Solution: AT&T executives authorized a joint project with
reduce the risk of disruptive events, such as drought, they take steps to adapt their business to its impacts. needed. For cities, investments are made where new
Argonne National Laboratory that used big data and
flooding, and resource scarcity. Among forward-looking The company serves as an essential utility to millions, infrastructure is needed most to safeguard against calamities.
location technology to map where climate risk will be for
businesses, a new effort is afoot to navigate disruptions providing connections to home computers, mobile For risk and insurance, the awareness lets firms pull back from
the next 30 years.
by leveraging data analytics and simulations using devices, Internet of Things (IoT) sensors, and more. risky scenarios.n
location intelligence that forecast climate trends next Those services are digital, but the infrastructure that
Result: With AI and location intelligence, business
season or decades into the future. By studying supply powers them—from cell towers to base stations—is very
executives have a new evidence-based ability to predict “It’s not just AT&T—I’m seeing this
chains in detail, companies are quantifying how resilient much part of the physical world and vulnerable to
they are. Where they identify vulnerabilities, they are climate impacts. AT&T, which maintains billions of
where supply chain vulnerabilities could emerge months more and more in all kinds of
and years in the future. A global awareness, coupled
taking preemptive action. dollars’ worth of equipment worldwide, is studying its industries where companies are
saying, ‘If this is the reality, then let’s
prepare for it, let’s deal with it.’”
—Antoine Diffloth, Director of Data Insights,
Chief Data Office, AT&T

6
Microsoft—Case Study

Impact Observatory, and Esri—Seeing

Near Real-Time Changes on Earth
Situation: Land-cover and land-use maps have been a Result: Each pixel in the land-cover map represents a
necessary tool for scientists and governments since the 10-by-10-meter block, bringing the slightest detail into
inception of remote sensing and GIS technology. More view. A repeatable global snapshot of land-cover change
recently, the objective has gone beyond collecting helps organizations react even at the far reaches of their
inventories to documenting climate-related changes and operations. Thanks to AI and location intelligence, the
impactful human activity, including deforestation and ability to monitor global change can be repeated at a
reforestation, to balance land use for the better health of scale and frequency that could only be dreamed of
our planet. However, the lack of global imagery and the before. The capability will lead to answers to pressing
computing power to calculate change has hindered a questions about climate change and environmental
consistent record of change at the worldwide scale. crises for years to come. Everything from conservation
planning to food security to hydrologic modeling can
Challenge: The data and processing challenge meant
be observed. n
that environmental assessments and land-cover analysis
happened at smaller scales, with infrequent updates.
Landscape-scale assessments have happened more
frequently for forests and watersheds, but a high-
resolution view of land-cover change has been elusive
and has taken a great deal of time to model, analyze,
and produce a record of change.

Solution: Open data and a partnership between Esri,

Impact Observatory, and Microsoft have yielded a
repeatable global snapshot of land-cover change. With
help from deep learning, the map makes it possible to
closely monitor in near real time the impacts of climate
change and humanity’s footprint. AI and location
intelligence underpin the ability to see change and
complications clearly. Ten distinct categories of land
cover, the physical type of land, show just what’s
where—such as trees, grass, crops, built areas, wetlands,
scrub/shrub, and snow/ice.

7
Achieving Automation for Sustainable
Business Operations

Given the scale and scope of global change, New climate risk analytics are built on a foundation of
organizations are moving quickly to gain a spatial spatial data and analytics that inform environmental,
inventory of resources and assets. They are creating social, and governance (ESG) concerns. At the
decision support systems to get a handle on natural forefront, companies are looking to be more efficient,
resources that underpin operations, such as water, with a longer-term focus on environmental concerns
timber, and minerals essential to products, to safeguard around carbon reduction.
against interruptions.
In many organizations, an AI and location intelligence
AI and location intelligence inputs are tracing supply approach is scaling in the form of collaboration hubs
chains and monitoring factors that might turn into that bring diverse departments together to solve
disruptions, allowing executives to trace delays, shared challenges. These hubs bring together all
shortages, the impact of geopolitical incidents, and stakeholders with shared maps. The new awareness
other risks. Adding transparency to supply chain takes companies from sensing and then responding to
vulnerability helps avoid costly delays or even the need predicting and intervening proactively.
to shut down factories and furlough workers because of
The whole mentality of operations is changing, with
delayed shipments.
smart maps that quickly communicate where action is
A geographic view adds traceability and an overview of needed. Using AI, location intelligence, and science,
the dependencies between suppliers and plants. The whole industries are guiding healthy operations with
geographic approach creates a tightly tuned system that more care for the environment to safeguard and reduce
can track incidents and detect threats and patterns to climate impacts. n
make organizations more resilient.

When unexpected events occur—or even before they

do—planners use these technologies to run what-if
scenarios, weighing the outcomes of changing a route, a
supplier, or a strategy.

AI and location intelligence can be used in tandem with

another GIS technique that displays supply chain
relationships in knowledge trees. Such visuals help
business leaders understand how an event’s effect on
one supply tier will ripple to others. Seeing that
information in a geographic context can illuminate
dependencies among suppliers and strengthen
contingency plans.

8
Finnish Forest Centre—Case Study

Robots for Better Forest

Management
Situation: Finland’s forests cover three-quarters of the Solution: The Finnish Forest Centre staff have created a Result: For foresters and landowners, the work of AI will “Going to the forest for
country’s landscape, and monitoring the maintenance of marketplace for forest owners to connect with forest give them knowledge about the volumes and species of
12 hours a day to drive a
this vast space falls on the Finnish Forest Centre. professionals to undertake maintenance work. The site wood without driving to the land to do a costly, time-
harvesting machine is lonely
Recently, the centre’s forestry experts developed an contains forest inventory data based on surveys and a consuming inspection. Automation has greatly streamlined
ambitious plan to use robots to autonomously perform sophisticated AI-driven algorithm that takes imagery and the management of Finnish forests. AI and location
work in a dark forest. Some
the majority of forest maintenance tasks. weather and climate data to better predict forest intelligence fuel the proactive management of forests, people like that, but there’s no
inventory and maintenance needs. The data must be focused on optimizing growth and harvest cycles. And, queue of people who want to
Challenge: More than 65 percent of the forests are
accurate and accessible, with all the data that the robots robots are filling an employment gap, working in the cold be trained for these jobs. It’s a
owned by private owners who live in other cities and
rarely visit. Using robots would allow proactive
will need. and the dark to maintain and sustain a whole new level of
real problem that can’t be
forest benefits observed. n
management of forests, a move that would maximize the
solved without automation.”
carbon sequestration of trees, and robots could fill an —Tapani Hämäläinen, Development Director,
Finnish Forest Centre
employment gap due to a lack of interest in forestry jobs.

9
Clearwater Seafoods—Case Study

Data-Driven Sustainable Marine Development

Situation: The Nova Scotia-based brand specializes in to a specific fishery to hunt specific species. One of Result: This new awareness allows Clearwater to factor
“We’re really moving the
luxury seafood including scallops, clams, crab, shrimp, Clearwater’s corporate values is to serve as stewards of in weather influences, species biology, and other
operation to being one of
and lobster that it gathers mostly from the seafloor the ocean through responsible fishing practices. relevant details to determine the most effective methods
in coastal Canada. Some of Clearwater’s catch, such of harvesting. With increasing amounts of high-quality
culturing and nurturing. To be
as clams, live on the seafloor, creating homes in the
Solution: Clearwater uses sophisticated assessment
data, Clearwater has moved from fishing to harvesting. able to look at the seafloor
silt and mud known as the substrate. Catching these
models to determine stages of growth and log fishing
Rather than going out into the ocean with the thought and understand population
activities to avoid returning to the same fishing grounds
seafloor dwellers not only labor-intensive and expensive. to catch a certain quota, the company now has data that characteristics—these are the
until the appropriate time has passed. This allows
A company like Clearwater needs to know when and every ship captain uses to plan their harvest. Each foray tools we need to support the
Clearwater to further its goals related to sustainability,
where to harvest but also when and where not to harvest captures more fish and more data that AI and location
as well as traceability for food safety, and to convey
objectives of sustainability.”
to protect the resource and the environment. intelligence workflows use to forecast and predict
the origin of each catch. In this way, GIS, coupled with —Jim Mosher, Director of Harvest Science,
sustainable harvests. n Clearwater Seafoods
Challenge: Clearwater needed a precision approach advanced analytics, becomes a four-dimensional tool,
to ensure corporate responsibility and improve cataloging the space and time for the present catch and
sustainability. In Canada, companies are granted rights all past catches.

10
Pinpointing Predictive Maintenance to Enhance
Efficiency and Safety

As funds have become available in the US via federal infrastructure

legislation, the country’s distressed infrastructure is finally getting
the attention it deserves. This provides a unique opportunity for
municipalities to not only repair existing transportation and transit
infrastructure but to also put in place measures to address and
identify problems before they reach crisis levels.

This presents problems of scale. A large city has thousands of miles of

roadways. Transportation departments are often forced to wait until
problems become serious enough to create visible levels of danger or
malfunction. And this goes for all sorts of infrastructure. Old bridges
are repaired when the damage becomes apparent. Streetlights
receive attention when they go dark.

Many cities are simply always in repair mode. A city might decide to
always be in the process of inspecting streetlights. By the time they’ve
all been checked, it’s time to start the process over. It’s an approach
that lacks efficiency and incurs excessive costs, either because repairs
only occur when they reach crisis levels or because cities devote large
amounts of resources to inspecting everything, all the time.

AI and location intelligence enable transportation departments,

and other asset-heavy industries, to remain proactive. Rather than
waiting for roads to exhibit enough damage to warrant repairs,
these industries use algorithms that can learn to spot warning signs
and create maps, allowing them to act before trouble spots become
crisis points.

AI and location intelligence also empower transportation agencies to

analyze large amounts of geospatial data, creating maps that reveal
places where safety and mobility need to be improved. By looking
at data spread across months, years, or decades, transportation and
supply chain managers are able to make the most informed decisions.
AI and location intelligence also forge links between different types of
transportation-related data, providing an optimal means to see how
mass transit data, for example, is related to traffic data.

11


“ Data gathered from

spaceborne sensors
has long been
used to calculate
the degradation of
natural landscapes,
such as the
deforestation of the
Amazon Rainforest;
but now, powerful
machine learning
techniques combined
with data gathered
from spaceborne
sensors are forging
a path to more
robust and resilient
data collection and
interpretation for
monitoring positive
interventions.”
—David Merren, Assistant Director,
Analysis & Geospatial, Deloitte Middle East

12
Autobahn—Case Study

Bavaria Gains Insights into Roadways with Predictive Maintenance

Situation: Bavaria’s centralized location in the European Challenge: One of the important routes in Bavaria is the the Ministry provided data on road conditions from different condition variables can be meaningfully connected, and
Union makes the German state a crossroad to reach motorway Bundesautobahn 70 (A 70). The heavy car measurement campaigns and traffic history. The A 70 thus, interdependencies of different parameters can be
several towns in Germany, a common destination for volume and the high percentage of truck transport take motorway was divided into 4,800 segments, each 100 meters included in the forecast.
travelers passing through on their way to other countries a toll on road conditions. The Bavarian State Ministry long. A GIS program served two functions: it provided a way to
With AI and location intelligence, the task of combing
and an important transit route for goods transport of Housing, Building, and Transport launched a pilot visualize the data and perform preprocessing functions. Using
through reams of data, which would’ve taken analysts
across Europe. The state’s 23,000 kilometers of roads project to make road maintenance on A 70 motorway the data, a deep learning model was developed that could
hundreds of hours to pore over, could now be done swiftly
(motorways, federal roads, and state roads) represent efficient by predicting when and where it will soon detect when road sections would need repairs by analyzing
and automatically as part of ongoing workflows. n
€40 billion in fixed assets. Due to a reform initiated by be needed. features such as thickness and road conditions.
the federal government, the previous responsibility for
Solution: The Ministry used a deep learning program Result: The pilot project has proven that these approaches are
motorways (including planning, building, and operating)
trained to notice and process indicators of road capable of realistically forecasting changing road conditions.
was transferred from states like Bavaria to the newly
conditions. Working with a team of data scientists, With the application of AI deep learning, several road
founded Autobahn GmbH at the beginning of 2021.

A visualization of
the input dataset
containing
features for 4,800
road sections
(each 100 meters
long) describing
road conditions
and traffic density
on the A 70.

13
US Army Corps of Engineers—Case Study

Predicting Where to Dredge to Save $100 Million per Year

Situation: Dredging US waterways is the largest item Challenge: Surveyors have greatly improved techniques moving through ports. This data fusion site is made
“We’re able to do more with the
in the US Army Corps of Engineers (USACE) civil works to assess depth and understand dredging requirements, possible in part by the capacity of AI and location
same amount of money and
budget. The need to keep ports and channels at a but the complexity and scale of the work sometimes intelligence to crunch the millions of Automatic
specific depth for ships and barges requires constant make it difficult to distinguish between needs and Identification System (AIS) vessel tracking pings to
reduce the long-standing
attention. USACE spends around $1.5 billion each year practices that have simply become routine. USACE show clearly where barges and container ships are maintenance backlog. They also
on dredging for hundreds of navigation projects across required a way to rigorously and repeatably determine slowing down due to building sediment. give us the ability to adjust and
the country. A typical dredging project goes through the most efficient and effective way to conduct adapt to the unexpected through
Result: The new system pinpoints exactly where
several phases, including project planning, bidding, maintenance dredging. And the solution needed to more efficient use of the industry
dredging should happen, eliminating significant reactive
contract award, placement, inspection, and project scale across hundreds of projects nationwide.
costs while fueling proactive workflows. It also addresses
dredging fleet.”
completion. At each point, there’s room for more input
Solution: The US Army Corps of Engineers combines project sequencing, with attention paid to coordinating —Dr. Ned Mitchell, Research Civil Engineer, US Army
to improve awareness of conditions of the country’s
Engineer Research and Development Center.
inputs from three systems that deliver a real-time the next project nearby to reduce travel time and costs.
12,000 miles of inland and intracoastal waterways,
understanding of the conditions of America’s waterways With increased ship sizes and growth in shipping, the
13,000 miles of deep-draft coastal channels, 400 ports,
into one Nav Portal view to see vessel traffic and number and scope of dredging projects will continue
and more.
channel constraints, ensuring that commerce keeps to rise. With AI and location intelligence assistance on
where and how to dredge, USACE estimates it can save
as much as $100 million per year. n

14
Keeping Supply Chains Movings

Supply chains have reached a complexity unheard of even The functioning of every segment depends on the smooth Visualization and contextualization are crucial, but they are
a few years ago. In a truly globalized world, any consumer functioning of every other segment. It is imperative that not enough. The complexity of supply chains demands
good has a multitude of origins. Pieces of it may be supply chain operators have a complete sense of how that managers be able to spot problems before they
manufactured in different countries; assembly might occur things are proceeding along every individual part of the occur—to see what’s over the horizon.
in another. Distribution is handled in different segments, chain and along the chain as a whole.
AI and location intelligence provide this extended view.
via ocean, rail, and road.
Companies and logistics support supervisors already
AI and location intelligence can note weather patterns in
augment supply chain management software with GIS
various parts of the world and compare them to historical
technology. Mapping allows them to see the whole supply
weather patterns to get a sense of whether an incoming
chain in context, to see how one segment functions in
storm front is likely to lead to catastrophic flooding. An AI
relation to others. Dashboards and other apps provide a
and location intelligence system can scan through millions
way to link the whole map with figures and statistics
of social media messages, conducting open-source
regarding the status of moving parts. At a glance, anyone
intelligence to understand unraveling political unrest. It can
can see what’s going on.
gather AIS signals from fleets of cargo vessels to know the
traffic volume and anticipate backlogs at ports.

This real-time contextual awareness is all the more

important as the world tries to deal with the various
disturbances—from travel restrictions to increased
shipping costs to pent-up demand—that have troubled
supply chains in the post-COVID-19 era. AI and location
intelligence can relate complex combinations of factors to
help us understand the new normal we can expect in
today’s volatile world. n

15
 FedEx—Case Study

Location Intelligence Helps Keep Shipments on Time

Situation: FedEx’s global logistics operation relies on Challenge: The company wanted to explore ways to wring augmented with AI’s predictive capability. The company’s security or visibility. Sensor data, accessible by the
extreme precision to function effectively. This includes even more efficiency out of their operation. The managers vice president of airline technology described the setup as customer, provides information pertaining to such factors
ensuring that the mechanics who work on the company’s who oversee the flights wondered if there was a way to “almost the holy grail for a maintenance operation.” as location, temperature, and humidity. The collected
fleet of airplanes have the replacement parts they need. In determine when a plane would need parts or data is also used to train the system’s machine learning
Result: FedEx is expanding its use of AI as a key
a given year, FedEx’s fleet of planes might experience maintenance. program to consider more time and place variables to
component of the company’s operations. A package
nearly a quarter-million flights, with less than 1 percent continually increase efficiency. n
Solution: AI and location intelligence helped the company analytics system provides extra tracking and monitoring for
failing to take off within 15 minutes of the scheduled
achieve this efficiency. The existing GIS system was packages that are time-sensitive or may require extra
departure, usually attributed to late arrival. To keep these
planes on time, FedEx relies on a sophisticated GIS system
to track its fleet.

16
Utilities Inspect Assets and Reduce Network
Loss with Insights from AI

Electric and water utilities have embraced the power of AI

and location intelligence to increase operational efficiency.
Network operators use AI and location intelligence to inspect
and manage assets, such as power lines and pipes.

Predictive maintenance is a huge area of improvement for

utility operations. It uses data from sensors and other sources
to alleviate issues before they cascade. This work is taking
place to understand the performance of the network, such
as leaks in a water system, as well as conditions around
the assets, such as monitoring tree growth to enhance
vegetation management around utility lines. With AI and
location intelligence, utilities are improving the reliability and
sustainability of their operations.

Annual surveys of power lines and pipes collect massive

amounts of data, and it has taken staff thousands of hours
in the past to sort through the data and images to craft
maintenance plans. Getting past the manual labeling of
images and the tedious job of visiting every asset in virtual
form holds great promise for staff who must maintain
thousands of miles of network infrastructure.

An AI and location intelligence deep learning approach

can detect wires and other representative components of
networks, assess their condition, and discern pain points.

Overall, AI and location intelligence can help utilities increase

the efficiency and reliability of their operations by collecting
and analyzing conditions quickly to spot potential problems.
This in turn reduces downtime and maintenance costs and
improves the performance of
all assets. n

17
 Nob Hill Water Association—Case Study

Replacing Old Water Mains after AI Analysis

Situation: Utilities around the world struggle to deal with would factor in such considerations as the age of the pipe, analyzes include pipe data, weather, soil type, seismic
“Every utility has more miles of
water loss due to broken pipes. Deteriorating water its failure history, and the material it is made of. The utility activity, and traffic—all of which reveal the water mains
main line that should be
infrastructure not only impacts customers but can also needed a system to inform it of the required work while most likely to fail next. The model also establishes the
erode roads and damage property. Nob Hill Water maintaining service to its customers. It needed a way to pipes that have remaining useful life, which helps the utility
replaced than there is money in
Association in Yakima, Washington, faced the familiar avoid water loss and stay on budget. avoid replacing healthy pipes prematurely. the budget. We are using this
challenge of aging infrastructure. Frustrated customers program to direct our valve
Solution: Nob Hill Water used an AI and location Result: With knowledge of the likelihood of failure and the
were experiencing more service disruptions and higher exercising program to the mains
intelligence deep learning model from Esri startup partner potential for property consequences, Nob Hill Water
water bills, while Nob Hill was dealing with higher repair that are predicted to fail so that
VODA.ai to conduct a conditions assessment that helped makes the most of its maintenance budget. The utility’s
costs and increased water loss from broken pipes.
it pinpoint the water mains most in need of replacement. engineers credit AI and location intelligence with efficiency
if they do fail, the damage can
Challenge: Preventing loss by proactively replacing water The machine learning model ranks each water main gains and making smarter decisions. The map view of be kept to a minimum.”
pipes is not an easy task. The utility needed a system to segment by the likelihood of failure, as well as the priorities also makes it easier to find and test nearby valves —Zella West, Manager, Nob Hill Water

select the right pipes to replace at the right time that consequences of that failure. Factors that the system and other assets to validate what valves serve each
customer—creating a better map for all future work. n

Side Bar
Rethinking Utility Vegetation Management with AI and Location Intelligence

Power lines require regular vegetation management to eliminate encroachments that can lead to power
outages and, in the worst cases, wildfires. Using an AI and location intelligence deep learning approach,
vegetation encroachment analysis can be conducted at scale. The workflows not only help identify
encroachment sites but also help prioritize them based on whether vegetation growth is above the wire,
below the wire, or intertwined with the distribution wires.

Inspecting and identifying places that need work take a great deal of time, money, and resources—costs that
are compounded when tackling trimming work. AI and location intelligence workflows automate the analysis
of satellite and drone imagery and lidar data—classifying the condition of distribution wires and poles in
relation to trees.

The AI and location intelligence workflows save costs and optimize the process of vegetation management
for utility companies—ensuring a more thorough assessment. With the clean energy revolution underway,
the resilience of power lines is growing in importance. Automation ensures that these processes can be
repeated regularly to add resilience with pinpoint precision.

18
“We can, for the first time in
history, quickly and easily
layer together data from a
vast array of sources that
can tell us all kinds of
important stories about both
the natural world and our
human world. This
geographic approach is
collaborative; it can bring
people together regardless
of economics or
nationalities. And it can give
us what we need: common
ground, an entirely new
understanding.”
—Jack Dangermond,
Cofounder and President, Esri

19
Spotting Unseen Expansion Opportunities

Businesses need to know where their customers are, but that's Companies can dig even deeper into consumer habits. Where ground for disparate data from many sources, from human requires due diligence. This can include anything from
not the only relevant question. An even better one is, Where do people like to pay by credit card? How many take mass movement data derived from anonymized cell phone signals to whether a coastal location might become problematic
will their future customers be? And how does the customer transit? What is the average travel time someone would be American Community Survey analyses of census blocks. as climate change intensifies to an analysis of where the
base in one region differ from those in others? willing to endure to visit a new location? employees it wants to attract desire to live.
Answering the relevant questions means searching for hot
The science of answering these questions has grown more These are all critical business questions, and they are all spots hidden amid all the data sources. AI and location intelligence are key to finding such
complex. Demographic data lets companies examine a geographic questions. They involve the relationship between answers.
customer base in minute detail. What is the specific, most people and place. These same principles do not merely apply to companies

promising demographic of a city? Where are they? What are seeking to retain customers and attract new ones—they hold

their interests? What are their occupations? What are their But these relationships are complex, involving many variables. true for locating any kind of new physical assets. If a company

spending and investment priorities? AI and location intelligence provide a common meeting is looking to site a new corporate headquarters, the investment

Side Bar
Halo Forecasting with AI: A Major Leap in Retail Planning

Over the last several years, large retailers have noticed that online sales do not necessarily detract from sales made in
a chain’s brick-and-mortar store. On the contrary, digital sales often increase in geographic areas near a store’s physical
location. The industry calls this mutually beneficial relationship the halo effect.

The halo effect is not a sure thing. Retailers need a way to create the optimal halo effect by deciding where to site stores.
That means considering many types of factors, including data on the company’s in-store and online sales, detailed
demographic data, and the complex spatial relationships between a proposed store site and surrounding consumers.

Effective halo forecasting requires the judicious application of AI and location intelligence. The first step is to use GIS to
“quantify proximity.” Using information such as drive-times from surrounding locations, supplemented by demographic
and consumer preference data, the company can determine expected revenue from various possible store locations. The
company then uses AI and location intelligence to perform a close analysis of consumer spending that yields an accurate
forecast of online sales. The results are displayed on a map, showing the company which physical locations will maximize
both in-store and online sales.

GIS is then used to examine suitable locations within these preferred areas. To further narrow the choice, the company can
use the map to see the locations of available parcels zoned for commercial activity. It can then add further attributes, such
as the location of high-traffic intersections.

.
20
Providing an Enhanced Awareness for Sustainable Carbon
Farming Practices

Businesses are under a microscope like never before. No longer is it enough

for them to deliver products their customers want. Today, they must also
deliver transparency.

An increasing number of customers want to know where their products

are coming from. They want to know that the businesses that make them
are responsible corporate citizens. They want to make sure products
are sustainable and ethically made. If a company makes sustainability
commitments, consumers want to be sure there’s no deceptive advertising.

Organic food brands, for example, have rigorous practices in place to ensure
sustainability practices. Consumers of organic products expect no genetically
modified plants, no growth hormones or antibiotics, and no synthetic fertilizers
or pesticides. And when these brands want to take it a step further, it’s this very
framework that gives them something to build on. Optimum soil health sits
at the foundation of sustainable organic farming to retain moisture and carry
nutrients to plants effectively.

AI and location intelligence help sort through the data explosion brought on
by precision farming. Farmers now gather data from equipment that plants,
fertilizes, weeds, and harvests. Many use drones to analyze plant vigor or
satellite imagery that can be downloaded on a daily basis. Weather is another
critical variable in every aspect of operations. All of these inputs roll up into big
data that requires AI and location intelligence to analyze it.

In high-value crops, such as grapes in vineyards, there’s almost no limit to what

is sensed and what inputs are applied. Artificial intelligence is used to correlate
weather information with soil sensor networks to make recommendations on
water needs. Drones gather daily data that AI and location intelligence analyze
for signs of disease, the wine world’s biggest problem.

These micropractices also work at the macro level for larger fields of crops.
Farmers who have applied predictive modeling point to labor and cost savings
that ensure there’s no going back to traditional methods. With AI and location
intelligence, agriculture has a way to aggregate all inputs for smarter and
more automated processes, leading to higher productivity, profitability, and
sustainability.

21
 Organic Valley—Case Study

Improving Pasture Vigor with Insights

Derived from Daily Satellite Imagery
Situation: Organic Valley, a national organic brand more than 40 percent of Amish or Mennonite farmers in component of these reports is the “grazing wedge” that
and family farm cooperative, prioritizes smart grazing the cooperative, a high-tech solution would need to be shows the estimated biomass availability for
practices, determining optimal methods for rotating cows hands-off because they don’t use computers. each paddock.
across pastures.
Solution: Organic Valley now uses custom mapping Result: The system produces better analysis with less effort
Challenge: As many as 200 different types of plants products to create pasture condition reports. AI and on the part of individual farmers. There are also important
can grow on an Organic Valley farm. A careful rotation location intelligence algorithms are applied to daily ancillary benefits. The AI and location intelligence
results in the maintenance of a critical balance of grasses satellite imagery passes to produce insights regarding capability behind the imagery analysis is sophisticated, but
and legumes. Farmers were accustomed to painstaking which paddocks should be grazed, cut, or left to grow by the reports themselves are simple. The data generated by
measures for determining grazing rotation. To determine measuring such variables as the amount of chlorophyll the program, combined with GIS-based routing systems,
the amount of biomass on a pasture, a farmer would walk in an area. AI and location intelligence automate the helps determine what farms to visit to fill each truck. The
through the paddocks with a plate meter to measure the process, generating reports sent to farmers. A critical same data workflow also measures carbon sequestration,
height of grass, recording figures in a notebook. For the an important potential revenue stream with growing
interest in carbon farming.

22
Locating the Most Important Targets Quickly, with
Urgency

In business and government, there are moments when time is the

factor. When decisions deal with life and death, the need for trustworthy
information acquires a new urgency.

What that often requires is distilling clarity from the chaos. In a complex
environment, what are the things we need to address now? Which are
most important, and how do they connect? Only then can we formulate a
dependable plan of action.

Natural disasters and humanitarian crises offer the starkest examples, but
these requirements are not limited to split-second life-or-death situations.

Effective business decisions often require a quick assessment of the

situation, and this includes physical assets. AI and location intelligence
provide a way to automate tasks that would take a combination of
people an exponentially greater number of hours to complete. A
trained AI model can gather useful intel to make sense of a complex
situation quickly.

AI and location intelligence are especially important today, given the

large volume of imagery and other data streams. A satellite image of a
large area may offer a wellspring of information, but without an effective
means to identify relevant structures, objects, or landmarks, it’s little more
than an interesting aerial photo.

The ability of AI and location intelligence to identify and sort objects,

and even prioritize actions, means they are the ideal partners in chaotic
situations. When floods, fires, or earthquakes are rearranging the physical
environment, AI and location intelligence not only show you where things
are now but can also quickly sense levels of damage so you know where
the most help is needed.

The ability to automate targets for action is not merely convenient. It

supports making the right decisions. AI and location intelligence impose
order on a dynamic world, but they are also itself dynamic. A well-trained
model can learn to adapt to different situations, and what the model finds
can be fed back into it as information, preparing it for future situations. AI
and location intelligence are equipped for uncertainty, even crisis.

23
The Mozambique National Institute for Disaster Management and Risk Reduction—Case Study

Aiding Flood Victims before and during

Disasters with Drone Imagery
Situation: In 2019, two cyclones destroyed more Solution: Drones were put to the task of searching for Result: Coupling drones with AI and location intelligence
“We combine science,
than 800,000 hectares of farmland in Mozambique survivors, freeing up helicopter pilots to conduct rescue workflows has greatly improved disaster assessments
technology, and local knowledge
during harvest season, leaving two million people and supply missions. And now, that workflow involves and reduced the time it takes to instruct the population
facing acute food insecurity. The United Nations World advanced AI and location intelligence algorithms to to leave flood zones for safe zones. That used to take
to prepare the communities and
Food Programme (WFP) responded quickly, using two search through drone imagery to identify flood victims. 72 hours, and now, the alerts happen well in advance local governments in a
helicopters to ferry supplies and rescue stranded people. The human eye has trouble with this task, and there’s of floodwaters. That leap has come in part from the use participatory manner. Every year
not enough time to both collect and review each image. of drone imagery to construct intricate hazard maps for we are seeing more frequent
Challenge: The flooded roads made the air support
AI and location intelligence can quickly identify people watersheds, using AI and location intelligence to process flooding, and that has motivated
crucial, but it was not nearly enough to both distribute
surrounded by flood waters and quantify damage to imagery with a focus on preparedness. WFP is now
food and find stranded people. The dimensions of the
us to innovate.”
buildings and infrastructure to help responders know spreading AI and location intelligence advancements
flooded areas were in constant flux, further complicating —Antonio Jose Beleza, Deputy Director,
what resources are needed and where. from Mozambique worldwide. Mozambique National Emergency Operations Center
efforts by pilots to understand where help was most
needed. And the people could only tread water or cling
to trees for a time.

24
Conclusion
As the influence of AI and location intelligence continue to grow across
all sectors of society, claims made on their behalf can easily shade into
hyperbole. It is important to understand that AI and location intelligence
are not only transformative tools but are also capable of harnessing the
power of human intelligence. They provides the means to tackle even the
most complex problems by empowering people to craft solutions using a
geographic approach.

As humanity’s understanding of place has grown over the millennia, the

maps people create have grown more complex. Even in an age of turn-
by-turn directions, the concept of the map prevails for one simple reason:
maps are useful manifestations of the way we see our world and our
position within it.

The map concept is endlessly expansive. The petabytes of data with a

spatial component generated around the world require maps of massive
numbers of inputs. The resultant maps expose vulnerabilities and where
work must be done, quickly conveying actionable information that might
otherwise stay hidden.

AI and location intelligence are potent means to make sense of complex

situations and our data deluge in the service of greater understanding.
This is what AI and location intelligence offer organizations—not
simple solutions but ways to visualize challenges, sharpen analysis,
and help deliver well-considered strategies that factor in all relevant
measurements. AI and location intelligence are not replacements for
human intelligence. After all, they are products of human intelligence—
and powerful partners for us as well.

25
 About Esri

Esri, the global market leader in geographic information For more information, please contact
system (GIS) software, location intelligence, and mapping, Esri
helps customers unlock the full potential of data to improve
380 New York Street
operational and business results. Founded in 1969 in
Redlands, California, USA, Esri software is deployed in Redlands, California 92373-8100 usa

hundreds of thousands of organizations globally, including

Fortune 500 companies, government agencies, nonprofit 1 800 447 9778
institutions, and universities. Esri has regional offices, t 909 793 2853
international distributors, and partners providing local
f 909 793 5953
support in over 100 countries on six continents. With its
pioneering commitment to geospatial technology and [email protected]
analytics, Esri engineers the most innovative solutions that esri.com
leverage a geographic approach to solving some of the
world’s most complex problems by placing them in the
crucial context of location.

Learn more at
esri.com/AI
Copyright © 2023 Esri. All rights reserved. Esri, the Esri globe logo, The Science of Where, @esri.com, and esri.com
are trademarks, service marks, or registered marks of Esri in the United States, the European Community, or certain
other jurisdictions. Other companies and products or services mentioned herein may be trademarks, service marks,
Map Credit: City of Philadelphia, Zoning Code Commission (ZCC) or registered marks of their respective mark owners.

G3009345

26