IN TRAVEL, IT'S TIME TO PUSH AI BEYOND THE PILOT PHASE - BCG
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
IN TRAVEL, IT’S TIME
TO PUSH AI BEYOND
THE PILOT PHASE
By Olivier Bouffault, Jason Guggenheim, Pranay Jhunjhunwala, Shervin Khodabandeh,
Tom McCaleb, and Ben Wade
A rtificial intelligence (AI) has not
yet taken off in the travel industry, but
it hasn’t been for lack of trying. Most large
different from any other they have ever
encountered.
travel companies have run pilots, tested
proofs of concept, and experimented with AI’s Potential in Travel
various AI tools, but few have realized the In simple terms, AI is a machine-based sys-
benefits of AI that leaders in other indus- tem that absorbs data, adapts to change,
tries have experienced. and takes action or provides information
that helps businesses make better and fast-
This is a missed opportunity. Travel compa- er decisions. Machine learning is a subset
nies are rich in data. They confront both in- of AI that involves actually learning from,
ternal complexity of operations, scheduling, rather than simply processing, data. (Execu-
and pricing as well as external complexity tives who want a more thorough under-
influenced by GDP, fuel prices, weather, and standing of AI can review Putting Artificial
terrorism. These characteristics play to AI’s Intelligence to Work, BCG Focus, September
strengths. 2017; and “Ten Things Every Manager
Should Know About Artificial Intelligence,”
AI can help travel companies boost opera- BCG article, September 2017.)
tional efficiency, route optimization, and
yield management; improve loyalty pro- For travel companies, AI offers great poten-
grams and broader customer journeys; and tial in at least four largely untapped areas.
speed back-office processes in accounting
and finance. But to ensure tangible and Tapping into a Data-Rich Environment.
sustainable value, companies need to do Travel companies have large amounts of
more than experiment. They need to align information that they are not fully exploit-
their leadership, capabilities, behavior, ing—everything from macroeconomic data,
and operating model with a technology geopolitical developments, and weathertrends to indicators of customer behavior For example, Dutch airline KLM and BCG
collected through loyalty programs and are jointly developing AI-enabled decision-
operational data gathered from sensors support tools that help predict delays, op-
onboard planes and ships. And their work timize aircraft and crew scheduling, and
with third parties gives them access to even improve the passenger experience. These
richer data sets. tools quickly analyze many scenarios,
taking into account crew positions and
Recently, BCG worked with a global airline availability, aircraft positions, and mainte-
to optimize the personalization of customer nance programs. Armed with this analysis,
emails with destination offers, loyalty mes- frontline staff can focus on making inte-
sages, and other tactics. We worked with grated decisions that boost operational
the carrier to build and test several machine- performance.
learning predictive models that took into
account more than 3,000 variables includ- At a major rail operator, executives wanted
ing customer, booking, and external data. to use AI to reduce maintenance costs.
The subsequent emails had a hit, or open, BCG helped the operator consolidate years
rate two times higher than a control group of scanning data in order to identify risk
and generated a 10% boost in revenue. patterns that would suggest a need to re-
pair a section of track. This predictive-
Managing Complex Operations. Many maintenance engine helped reduce main-
travel companies oversee vast operational tenance costs by 10%, improved network
footprints that could use help. First, many utilization, and created an opportunity to
of the underlying systems—which were sell the tool to other rail companies.
once state of the art—would benefit from
the insights into efficiency and scheduling Removing Friction in High-Stakes Customer
that machine learning can offer. Second, Touch Points. Customers often interact
the operations that move people from place with several travel companies at different
to place play an outsize role in pleasing stages of their journey. (See the exhibit.)
or disappointing customers, smoothing Many of these interactions, such as delays
journeys, and improving the overall experi- related to weather or unforeseen mechani-
ence. AI can help reduce friction at certain cal breakdowns, are emotionally charged
key moments, especially during complex and beyond the companies’ control. None-
activities that involve the largest number of theless, customers frequently hold the
people, processes, and systems. companies responsible. After all, the travel
Representative AI Use Cases Along the Customer Journey
Guest
Demand Guest Service compensation
generation Booking Check-in experience recovery (optimization
(personalized offers, (reservations by voice, (biometrics, facial (amenities, customer (active by customer
marketing messages) smart recommendations) recognition) preferences) resolution) or event)
DREAM AND PLAN BOOK TRAVEL ENGAGE
Marketing Yield Network Maintenance Labor Navigation Contact center
spending management optimization (predictive scheduling (optimization (natural-language-
(mix optimization, (advanced forecasting, (scheduling, maintenance, (demand-driven of cost processing sentiment
deaveraging) machine learning) disruption robot-assisted staffing, and speed) analysis, chatbots)
management) supply chain optimizing
and repairs) reserves)
Source: BCG.
Boston Consulting Group | In Travel, It’s Time to Push AI Beyond the Pilot Phase 2was sold as an experience or even a dream, Practical Steps to Generating
not a nightmare. Value with AI
The opportunities for AI in travel are real
AI can help anticipate and respond to such but difficult to achieve. All companies con-
lapses even if the company was not at front what we call the 10-20-70 problem of
fault. In his book Setting the Table, famous machine learning. About 10% of the chal-
New York restaurateur Danny Meyer talks lenge of implementation involves data sci-
about “writing a great last chapter” when ence and the algorithms themselves, while
dealing with a dissatisfied customer. He en- 20% relates to the need for enabling tech-
courages employees to turn a mistake into nology infrastructure and data engineering.
a positive experience that the customer The final 70% covers embedding AI into
will remember. Machine learning can help business processes and adjusting ways of
travel companies write the last chapter by, working so that people will use these new
for example, suggesting complementary tools and create business value. The 10% is
services that the customer has accessed in not trivial, requiring a deep understanding
prior trips or anticipating their unstated of both data science and the underlying
needs and wishes as inferred from past business problem. But too often, compa-
interactions. nies spin their wheels on that 10% without
ever making substantive business progress.
Managing Demand with Greater Sophisti- (See The Big Leap Toward AI at Scale, BCG
cation. In other industries, AI has been a Focus, June 2018.)
boon to channel and yield management as
companies have begun to rely on machine In helping travel companies introduce AI
learning to forecast demand and optimize pilots and bring the successful ones to
production across channels and markets. scale, we have crafted several recommen-
dations that address the context of the in-
Travel companies have been making simi- dustry and the 20% and 70% challenges
lar decisions for decades but generally that companies often overlook.
without the assistance of machine learning.
They decide how many rooms to allocate Understanding the Value Potential and
to online travel agencies or seats to low- Landscape. Travel companies should
cost fares on the basis of traditional analy- analyze AI’s potential in internal opera-
sis and good, old-fashioned intuition. With tions and along customer journeys and
commissions of 15% to 20% for agencies, focus on those sweet spots that will create
rising direct-marketing costs, and signifi- the most value with available or accessible
cant investments in loyalty programs, these data. As part of this analysis, they should
forecasting decisions are critical to the bot- understand where competitors, partners,
tom line. And, as in the KLM-BCG airline or digital upstarts may be trying to use AI
operations solution, machines can remove and how these developments will affect
the drudgery of the exercise and free mar- their strategic advantage. This analysis
keting and pricing executives to think more should expose opportunities that can serve
strategically. as the basis for pilot projects. In some
cases, a business case may be so clear that
Marriott has jumped at this opportunity it makes sense to accelerate or narrow the
to improve profitability. In mid-2018, the piloting phase.
hotelier announced the rollout of a new,
AI-based system that relies on machine Gathering and Coordinating the Data and
learning to understand demand and will- Managing the Algorithms—the 20%. Travel
ingness to pay on the basis of room type, companies are awash in data, but many of
cyclicality, seasonality, and nearby spe- them have not fully collected, organized,
cial events. CEO Arne Sorenson has said and evaluated it. For example, cruise
publicly that the new system has already operators have an enormous amount of
helped to lower reliance on high-cost chan- data onboard but must decide what gets
nels at times of peak demand. replicated onshore. At the same time, travel
Boston Consulting Group | In Travel, It’s Time to Push AI Beyond the Pilot Phase 3companies need to bring together data frenetic (the back of the house of a cruise
from many touch points and silos, often ship or hotel).
outside their organization. For example, to
deliver targeted offers for an ancillary Successful operational AI projects need to
product or benefit, companies need an bring key stakeholders together, manage
end-to-end understanding of customer change, and coordinate all the moving
journeys, including interactions with other parts. At a high level, executives must ad-
companies. dress a delicate three-dimensional organi-
zational balancing act:
Once they have corralled their data, com-
panies have a related challenge: orches- •• Centralized Activities. Data is the raw
trating the data as it moves through the material of AI and contains some of the
algorithms. A typical IT system consists of company’s most sensitive intellectual
data input, a tool, and data output. They property. As such, data management,
are relatively easy to scale because the expertise, and governance should be
tool is static. But AI algorithms learn by centralized so that the company can
ingesting data—the training data is an in- take advantage of scale, consistency,
tegral part of the AI tool. This “entangle- and security.
ment” of data and tool is manageable
during pilots but becomes exponentially •• Embedded Activities. Business units
more difficult to address as AI systems in- or functions should oversee the devel-
teract and build upon one another. Travel opment of pilots and use cases. The
companies need to buy or build a solution idea is to integrate AI into the fabric of
to monitor workflow from data input to the organization. The teams overseeing
final action. AI projects need to be flexible and
iterative to accommodate the self-
Finally, companies need to ensure that they learning nature of AI machines. Many
have the storage, computing, and bandwidth companies rely on variations of agile to
to handle multiple AI engines. The flexibili- ensure that the team’s way of working
ty of the cloud makes it a preferred option reflects AI’s way of working. Given the
to address these needs. But in certain con- decentralized structure of so many
texts, such as cruise lines and operational travel companies, creating these agile
control centers, latency and bandwidth con- teams can be especially challenging but
straints may prevent the cloud from serving is nonetheless critical. Left to their own,
as a complete solution. These settings re- data scientists can come up with exquis-
quire novel structures, such as edge comput- ite but impractical solutions.
ing, in which part of the processing power is
kept closer to the periphery. •• Decentralized Action. Finally, AI action
should remain decentralized. In the
Most companies will ultimately need a rel- travel industry, this often means putting
atively small number of data scientists and AI-enabled decision-making authority in
AI experts. But to integrate AI decision the hands of frontline staff in the opera-
making into ongoing processes, they need tions center, at the front desk of a hotel,
a large number of data engineers to ensure or in the kitchen. A travel company that
the performance and resilience of the pipe- wanted to reduce food waste in its kitch-
line and peripheral systems. ens, for example, created an AI tool to
achieve economies in food preparation.
Moving Beyond Pilots on an Organizational The chefs were then given mobile apps
and People Level—the 70%. The second- that suggested how much they should be
and third-order consequences of introduc- cooking by the hour. Food waste has
ing AI are exacerbated at travel companies since dropped by roughly 40%. Because
because their processes are complex (an the chefs had a hand in the design of the
airline operations control center or reve- app, they were more likely to trust its
nue management department) and often recommendations.
Boston Consulting Group | In Travel, It’s Time to Push AI Beyond the Pilot Phase 4In addition to this balancing act, executives
need to prepare their teams to work with AI
and create thoughtful change management
T he travel industry is at a critical
inflection point that will determine
whether individual companies stay stuck
programs. The technology often unnerves in the world of experimentation or achieve
employees, even though it generally im- scale and meaningful results through AI.
proves their work life. Similar to the chefs in The industry’s next wave of competitive
the example above, employees are happy to advantage will benefit companies that can
have better information at their disposal. make that transition across several areas of
their business.
As AI plays a larger role, however, the ma-
chines that initially helped improve perfor- A senior leader at a travel client recently
mance and reduce drudgery can cause job told us, “Much of our organization is stuck
security concerns. Companies should start in the buzzwords and can’t yet even imag-
addressing these anxieties through change ine what AI could do for us.” It’s time for all
management and reskilling programs. One travel companies to fire up their imagina-
approach, adopted by leaders in AI such as tive powers and get to work putting their
Renault, is to create a digital hub, a large ideas into practice. If they don’t, the prover-
center dedicated to digital training. bial kids in the garage almost certainly will.
About the Authors
Olivier Bouffault is a partner and managing director in the Paris office of Boston Consulting Group.
He leads BCG Gamma for Western Europe and South America. He focuses on analytics and AI in oper-
ations and has been supporting airlines for more than ten years. You may contact him by email at
bouffault.olivier@bcg.com.
Jason Guggenheim is a partner and managing director in the firm’s Atlanta office. He leads BCG’s
global work in lodging and leisure and has advised airlines and cruise operators on operational and stra-
tegic issues. You may contact him by email at guggenheim.jason@bcg.com.
Pranay Jhunjhunwala is a partner and managing director in BCG’s London office. He leads the firm’s
global work in airlines and has served clients across the travel and tourism industry. You may contact him
by email at jhunjhunwala.pranay@bcg.com.
Shervin Khodabandeh is a senior partner and managing director in the firm’s Los Angeles office. He
leads BCG’s work in big data and advanced analytics in North America. You may contact him by email at
khodabandeh.shervin@bcg.com.
Tom McCaleb is a partner and managing director in BCG’s Atlanta office. He coleads the firm’s global
work in travel technology. You may contact him by email at mccaleb.tom@bcg.com.
Ben Wade is a partner and managing director in the firm’s London office. He focuses on the travel and
tourism sector, with a particular emphasis on airlines across a broad range of topics. You may contact him
by email at wade.ben@bcg.com.
Acknowledgments
The authors would like to thank Matt Johnson, a project leader in BCG’s Atlanta office, for his help re-
searching and writing this article.
Boston Consulting Group (BCG) is a global management consulting firm and the world’s leading advisor
on business strategy. We partner with clients from the private, public, and not-for-profit sectors in all re-
gions to identify their highest-value opportunities, address their most critical challenges, and transform
their enterprises. Our customized approach combines deep insight into the dynamics of companies and
markets with close collaboration at all levels of the client organization. This ensures that our clients
achieve sustainable competitive advantage, build more capable organizations, and secure lasting results.
Founded in 1963, BCG is a private company with offices in more than 90 cities in 50 countries. For more
information, please visit bcg.com.
Boston Consulting Group | In Travel, It’s Time to Push AI Beyond the Pilot Phase 5© Boston Consulting Group 2019. All rights reserved. 3/19
For information or permission to reprint, please contact BCG at permissions@bcg.com. To find the latest
BCG content and register to receive e-alerts on this topic or others, please visit bcg.com. Follow Boston
Consulting Group on Facebook and Twitter.
Boston Consulting Group | In Travel, It’s Time to Push AI Beyond the Pilot Phase 6You can also read
