Three years ago, the Global Business Travel Association
published survey results that identified "data to track program
performance and support decisions" as travel managers' second-most time-consuming
task, just behind supplier negotiations. More than half of survey respondents
projected the situation would only get worse in the following three to five
years. That time is now, and artificial intelligence may be on the verge of making
their data responsibilities a little easier.
Business intelligence provider PredictX is taking its travel
insights platform beyond filter configuration and dashboards by launching a
voice- and chat-enabled digital assistant. The assistant works in tandem with Alexa,
Amazon Echo and text-based platforms like Facebook Messenger, Slack, WhatsApp
and others to query a company's "data lake" for two types of
information: fast answers to stakeholders' continuous ad hoc queries and tactical
issues that require immediate action.
"We've shadowed our clients and discovered they spend
an inordinate amount of time fielding queries from stakeholders: leadership, department
heads, questions that come in over and over," said PredictX CEO Keesup Choe.
"We have systems that [stakeholders] can query and pull those answers
themselves, but the reality is that they just want to email or call the travel
management department."
On the other hand, there are day-to-day tactical issues with
booking, risk and regulatory compliance that travel managers are expected to
track and mitigate. The problem, said Choe, is that very few have had tools or
resources to address these issues proactively. Rather, they've scoured data
reports after the fact and then roll out traveler education and communication
strategies to attempt to prevent future infractions.
That's a lot of tactical work and behavioral management that
PredictX wants to take off the travel manager's daily to-do list by equipping
them with its digital assistant. To address the first challenge, the
assistant's natural language processing capabilities digest either spoken or
text-based queries to dive into the company's data pool and retrieve dynamic
answers to common questions. For example, "How much spend do we have on
Cathay Pacific?" Or, "What's the spend volume compared to contract commitment
at Hyatt Hotels?" The bots will query the data to look at both executed
travel and future bookings from global distribution system, credit card and other sources to deliver a
holistic view of that information, and then either the travel manager or the
system can message that back to the individual stakeholder.
Similarly, the digital assistant can be configured to scour
the data 24/7 to pick up deviations from compliance that a travel manager needs
to respond to every day—for example, noncompliant hotel bookings, employee
concentration thresholds on a single flight, potential tax liability in certain
international or municipal jurisdictions based on the time the traveler has
spent working in that market or high-risk travel that has not been approved
through proper channels. Rather than pulling this information with voice or
text commands, the assistant can push these issues into a prioritized to-do
list for the travel manager. Even better, it can suggest actions like, "Would you like me to alert the traveler's direct manager?" if there
is questionable tax situation or a nonapproved high-risk travel plan. Again, the technology
can automatically communicate issues to appropriate stakeholders if configured to do so.
A complex network of data drives both the insight retrieval and
the suggested actions. Analyzing this data network challenges travel managers,
per the GBTA survey, but is ripe for automation and machine learning. Based on the
examples above, PredictX would likely need to consume booking and credit card
data, third-party taxation schedules for top travel markets, travel and risk
policies and HR data to ensure communications are directed to the correct
departments and managers.
Not every source of data is as robust as the next, cautioned
Choe. "We have travel data coming from GDS and other live sources,"
he said. "Not every organization subscribes to real-time credit card data,
for example, and they may not have agreements in place to receive [Level 3]
data. So it's a question of data access, but we have the ability to create this
environment."
Preconfigured queries will run along the same pathways to analyze
dynamic data sets and deliver updated insights on a consistent basis. Ad hoc
queries could be somewhat dicier, especially in the beginning, given their
reliance on natural language processing. According to Choe, the tools are in
the "toddler stage" right now in terms of how much syntactical
variation they can handle. That could be one good argument for passing those
inquiries through a travel manager who can query the data with consistent
language conventions. With use, however, the NLP will acquire more flexibility.
PredictX is ready to put a minimum viable product into beta
with a handful of customers. "We're interested to explore the potential
with these clients and see how they develop different assistants that align
with their needs," said Choe.
While the business intelligence firm has not done
time-and-motion studies to quantify productivity enhancements, Choe said productivity
and business enablement is the heart of the digital assistant. "Even if you say 50 percent as a conservative guess of time savings
alone, that is massive," he said, adding that companies have to look at
the other side of that equation to evaluate the potential impact. "When you
are too busy running the business that you don't have time to develop the
business, that's a problem. When travel managers don't have time to collaborate
with other departments or forge deeper partnerships with suppliers because they
are too busy with day-to-day data tasks, that's a lost opportunity. The other
side of a tool like this is the value creation of getting the more strategic
things done."
Until now, PredictX has driven a direct-to-corporate
strategy with its business intelligence platform. During the GBTA convention in
San Diego the firm announced it would partner with progressive travel management
consultancy Festive Road to power intelligence for the consultancy's outsource
travel management services, a move that should expose PredictX capabilities to a
bit wider swath of the market.
At least one travel management company president at the GBTA convention was ogling
the PredictX analytics assistant as a potential productivity game changer on
the agency side. "These are the exact questions that our account managers
are asked all the time by our midmarket clients," he said. "Think
about the productivity a TMC could achieve with this kind of tool."
PredictX did not immediately respond about whether it would
be exploring opportunities with TMCs.