Credit card issuers have access to a whole lot of big data about their customers. “I've talked to card issuers who've said we have so much data now, about what our customers buy, how much they pay, where we spend -- we don't even need credit bureau data about these customers anymore,” Gerri Detweiler, head of market innovation at Nav, told Tradestreaming. And while it’s not surprising that credit card companies know all about their customers’ habits and preferences, what might not be quite as well-known is that accountingtech companies have an impressive dataset on their customers as well. “Through credit card import, we know exactly how much you like Thai food over sandwiches, or where you like to get Thai food, and how long it’s been since you had Thai food,” explained David Barrett, founder and CEO of Expensify. “We know not just where you work but who you work with and who you work for.” The company, which has Uber and Yahoo among the 25,000 companies from around the world registered on its platform, enables users to upload pictures of business expense receipts to its platform instead of saving them up. These receipts are then automatically analyzed, associated within the right account in the business’ ledger system, checked for fraud and accuracy metrics, and only if there is something amiss are they sent for human review. More exciting for users, perhaps, is the fact that the money arrives in their account the very next day. So while Expensify might not know as much as a credit card company does about its users, the accountingtech company’s vast database of virtual receipts does give it an ample amount of insight into their users’ lives. It was this rich data set that enabled it to launch Concierge, the platform’s built-in AI virtual assistant, in September 2016. Concierge is Expensify’s move from being a more passive receptacle of information to a proactive force in their users’ lifestyle choices -- and payments. “Our approach to AI is really about leveraging this tremendous insight that we have about the customerand then reaching out to the customer with the best possible anticipation of what their needs are,” said Barrett. Other accountingtechs have had the same idea. Toronto-based Wave, for example, leverages the data it has on SMB customers to connect them with cash before they reach a cashflow crunch. “Wave holds all of your financial information,” said Rob Maurin, vice president of communications at Wave, told Tradestreaming. “It knows your bank balances, we know who you have invoiced, how much, when they're due, we know how long we take to pay, [we have] all of these insights on finance and cashflow.” Unlike Wave, Expensify hopes to grow its AI into a fully-fledged personal assistant, with the predictive capabilities to make customers’ choices simple and fast. The plan is for Concierge to be able to book flights, order Ubers, make restaurant reservations -- all based on business expense receipts uploaded to the platform, or even the user’s digital calendar. At present, the virtual assistant is able to do users’ expense reports, take care of their receipts, and export things to their accounting package. Accountingtech obviously isn’t the only branch of finance looking into predictive personal assistants. Like credit card companies and accountants, banks also have access to an extensive amount of their customers’ spending data, and they’re starting to leverage it. Bank of America, for example, recently launched chatbot Erica, who is supposed to help customers cultivate better money habits through AI, predictive analytics, and cognitive messages. Concierge raises some interesting questions for Expensify, including the privacy of customer data, and whether service firms will pay-to-play for customers' travel needs. As far as privacy is concerned, Barrett believes that Expensify is well within acceptable boundaries. "It's fundamentally the customer's data, we're just leveraging the customer's data to give them the best options." Nor does he think that Expensify's revenue model, which currently consists of companies paying nine dollars per user per month, will grow to include payments from restaurants or hotels that want to be first on Concierge's list of recommendations "Our business model is more like Amazon Prime than it is like iTunes," said Barrett. "iTunes tries to nickel and dime you -- we just want those nine dollars a month, and we're just going to offer you more and more value to convince you to pay us that nine bucks a month."