Pi Travel Rebrands to PredictX

PredictX CEO Keesup Choe talks:

• How travel data gets predictive
• How intuitive business intelligence tools will change the future of travel management

Managed travel has been a laggard in the business intelligence space. Pi has had the advantage of working in other sectors that have complex data, industries like retail and healthcare. That has informed how it supports business travel data analysis. The company has just changed its name to PredictX. CEO Keesup Choe spoke with BTN editor-in-chief Elizabeth West about the change, what it says about the direction of the company and how that direction is applied to travel analytics.

BTN: What's in a name?

Choe: It's really an announcement to ourselves as much as the market of what our focus is, where we see the technology going and how we're going to leverage that to help our clients. If you think of how corporate travel managers use data today, even the sophisticated ones who've already gone through some of this work, it is: Analyze and fix. They're looking at what already happened; It takes 10 times more time and effort to fix something than to prevent it in the first place. That's really our core focus now: helping our users predict and prevent these problems.

BTN: Pi, now PredictX, has worked with companies in healthcare, retail and other industries with complex data. How are you bringing that experience to bear on travel?

Choe: We're seeing massive adoption of data analysis and application of machine learning algorithms in other areas like retail and healthcare. [Managed travel] is in the very beginning stages of inquiring about this and understanding it. Our advantage is that we're active in these other sectors where things are happening much faster. We're getting a lot of traction and interest from our clients wanting to apply these techniques to manage their [travel] programs.

BTN: Give me a couple of examples of techniques used in other sectors and how PredictX is applying them to travel.

Choe: The concept of clustering individuals is good example. When Facebook and Amazon target you with promotions, they're not really targeting you as an individual. If Facebook maintained individual profiles of 600 million users or a billion or whatever they have, it would be very difficult. [Instead], they take the data they have on all their actions and cluster you into a cohort of people who are kind of like you. Then, when they target you with a message or promotion, you think it's completely individualized for you. In fact, it's targeting a whole bunch of people with similar characteristics. They've been doing this in retail service for ages. We're rolling this application out fairly soon [for travel].

BTN: What kind of data are you using to arrive at these clusters, and what do you expect travel managers to do with them?

Choe: We're talking about calculations that most [travel managers] don't even look at like average time zones crossed and average ranking of hotels they stayed at, all sorts of things that wouldn't naturally be part of the data that a TMC might give you. We add a whole bunch of additional information. Then we apply this clustering algorithm to it and then we classify your travelers into three, four, five, six, seven distinct buckets. That allows travel managers to fine-tune travel policies that both save money and improve the experience for those cohorts. It also allows them to craft messages: policy announcements, new programs or vendors or other things that resonate with each group. [The traveler will think], "Oh wow, my travel managers know me." That's gotten a lot of interest.

BTN: Is there another example that is more budget- or supplier-management oriented?

Choe: Sure. It's in the early days, but again in retail, especially with area managers of stores and who are on the road, you can't send them complicated reports or dashboards. Literally, we send them text messages: "Go to this store, look at the cash balances," or, "Go to this store and review the pricing for these [items]."

BTN: For travel, that might translate into a text message about checking an airline agreement because the program is missing targets, or some other issue?

Choe: Correct. And some of that isn't really predictive; some of that comes out of our workflow engine, but it's still really, really useful. If you combine that text message workflow engine with pre-trip data, real-time pre-trip data, we can again avoid so many problems down the line. For instance, some countries have a maximum number of days you can stay. If you go over that, you pay a fine, plus, it's just a real hassle with the paperwork. We can text message the travel manager, and maybe the traveler, that a booked trip will likely cause them to overstay the allocation. Another application here is with safety and security and following up with people booking [high-risk places]. The text message might say, "OK, what are you doing there, and how can we help you make sure your itinerary is safe?" These are some of the things you could just do with workflow and pre-trip data. We're doing that now anyway for clients, taking real-time data directly from the GDS and giving our clients that real-time visibility. You combine that with predictive analytics and it could be incredibly powerful.

BTN: Data science and analytics isn't the No. 1 skill set of most travel managers. Can you talk about the data demands being placed on them and the challenges PredictX is trying to address?

Choe: Corporate travel managers have a big challenge. There's no group of professionals anywhere in any sector that's managing so much spend with so few resources. The previous-generation technologies were hard to use and [caused] a massive reliance on consultants to get answers [for how to optimize travel programs]. The technology should be so intuitive and suggest activities—sometimes workflow but even better when they can be predictive and based on accurate forecasts. We don't want to say, "Here's a complicated piece of software with a database and all sorts of stuff, and you need to go write a query to do this, this and this." Nobody should have to do that in this day and age.