AppZen

Startup AppZen takes expense auditing to a new level. Its artificial intelligence-powered engine not only analyzes each clients' expense reports and receipts to catch duplicates and out-of-policy purchases but also scours the Internet to validate that the merchants and meal companions who show up are real. Additionally, there's the AppZen Behavior Index, which indicates a traveler's tendency to violate policy intentionally, based on historical behavioral data. CEO Anant Kale told BTN payment and expense editor JoAnn DeLuna what makes AppZen "unique."

How is AppZen different from its competitors?

Plenty of companies pretend to be audit tools. They take expense data and run it through their system, but the majority of the data on expense reports is on [attached receipts]. The only way today [to review them] is by having a person browse through the entire document. We mimic auditors by checking everything as if a human with all the time in the world were checking it and looking on the Internet for clues but in real time. And we have our ABI.

Your 2016 partnership with Concur has been successful and has helped AppZen score large clients like Amazon. How else are you going to market?

We just did a partnership agreement with Coupa and are into the technical integration phase now. We plan to release [the integration] to the first set of customers by the end of this year. We're also partnering with Chrome River and Expensify.

What is the biggest challenge of operating a product based on innovation?

Nobody is coming to us saying, "I found you guys because you use AI for expense audit." It's not like selling expense automation software, where you have a thousand Google search results. Our product is really innovative, far ahead of what anyone imagines can be done. The challenge is: How do we ensure companies understand the value of it?

So how do you? Some expense management systems also have introduced automation throughout their systems, but many travel managers are reluctant to enable aspects of it. How do you remove that customer hesitation?

The only way to remove that hesitation is to prove it out with data. We go to a large company, look at their processes and the results they're getting. We say, "Let's look at a new way of doing it," and work behind the scenes. We take their data, which has already gone through all kinds of controls, and put it through our engine. We analyze their 10,000 expense reports and give them results. The data always speaks for itself.

What do customers find most surprising when they get their results?

[The results] are always amazing for them. They see someone got reimbursed when they shouldn't have, someone upgraded to first class and no one noticed, or they paid for car rental insurance. We'll analyze data for [a two-week period] and show them the [potential] savings [from catching such violations]. Think about what would happen if they did this for a year.

Getting the industry to think differently about T&E management is challenging. How do you encourage innovation?

Customers are still thinking in traditional silos. We're trying to solve for the fundamental approach to how travel and expenses happen and [to] revisit the entire concept. But you still have to have something that works with their existing processes because companies have invested millions in their travel management system, rolled it out for thousands of employees and trained everyone on submitting expense reports. You cannot make a wholesale change; it would be too disruptive. If we can show them the value these innovations can bring them, then they see for themselves without too much change.

You've said many legacy companies are jumping on the AI bandwagon and claiming their systems are innovative. How can clients distinguish between real capability and hype?

There's fake AI out there. AI and blockchain require a product to be built from the ground up using [those technologies]. If a company is older than three years, there's no way they can rebuild their platform from ground up to use AI. A true company will show you what it does [that] cannot be done by anyone else if they're true innovators in that space. Second, some vendors claim that if you buy a robotic automation platform you can build [an AI tool] yourself and use it for everything, and that absolutely doesn't work. Customers have to be really careful.