Beware the Bots

April 18, 2019
AI-powered Bookkeeping May Not Be What It Seems

By Blake Oliver

Have you seen the Theranos documentary yet? If not, you should really watch HBO’s “The Inventor: Out for Blood in Silicon Valley.” 

You’ll learn how Elizabeth Holmes dropped out of Stanford to raise more than $700 million from venture capitalists and private investors, claiming to have invented a revolutionary blood test that only required a few drops of blood. At its peak, the company was valued at $10 billion, making Holmes, who held 50 percent of the stock, one of America’s richest women.

But Theranos turned out to be a fraud. A 2015 Wall Street Journal investigation led to revelations that Theranos was using traditional blood testing machines to run its tests instead of its own devices, which either didn’t work or provided inaccurate results.

It turned out Holmes was playing an ethically questionable game that’s become enmeshed in Silicon Valley startup culture: “Fake it till you make it.” 

The idea is to go to market as aggressively as possible to raise money and build market share even before you have a product that works as intended. If it doesn’t work, you “fake it” until it does. The most important thing is to sell—and grow.

Holmes’s plan was to falsify the tests until she could get her own machines working well enough to take over. Unfortunately for Holmes, her dream proved too ambitious and she ran out of money. 

Unfortunately for the rest of us, “fake it till you make it” isn’t limited just to highly publicized cases like Theranos. Take artificial intelligence as an example. It’s a hot topic these days, both in accounting and the broader business world. It seems everybody is talking about how AI is going to revolutionize all sorts of industries in short order. 

The hype around AI has led to massive amounts of investment in companies claiming to be building AI-powered tools to cash in on the trend. In fact, according to a 2018 report by London-based venture capital firm MMC, 40 percent of “AI startups” in Europe don’t actually use any AI in a meaningful way. 

I suspect that this phenomenon isn’t limited to Europe. Case in point: I recently came across a startup serving the accounting profession where I couldn’t find any evidence of AI being a material part of the business model. The company appears to be using offshore labor behind the scenes for most of their data processing. 

Botkeeper is an “automated bookkeeping” service for small businesses and CPA firms. Their charismatic founder and CEO Enrico Palmerino has appeared on FOX Business News claiming to have created an artificial intelligence “robot bookkeeper” that “mimics the accountant brain.” He even told Accounting Today that Botkeeper is “a full bookkeeper replacement” (accountingtoday.com/news/botkeeper-raises-18m-develops-partner-program).

The Botkeeper website gushes about the power of AI and automation to streamline bookkeeping. But what you’ll have a hard time finding on the website is any information about a Philippines office where the company employs dozens of accountants.

It’s a clever marketing tactic. By disguising humans as “bots,” offshore firms are able to overcome many of the typical objections to outsourcing, such as language or cultural barriers. But it creates serious risks for firms when it comes to security and confidentiality.

It’s not clear just how much companies like Botkeeper are disclosing their offshoring of work and data to their prospects and customers. Browsing through the website gives the impression that the service is highly automated. Yet in all likelihood, humans are logging in to your systems to view and process your data (or your clients’ data). 

There’s a big difference between granting access to an automated system to your ERP, and granting access to humans outside the country. For one thing, the more individuals that have access to confidential information, the greater the risk of breaches.

CPA firms should be especially wary of such services due to strict state and federal requirements regarding the disclosure of client data to third parties. For instance, California Board of Accountancy regulations stipulate that “In the event that confidential client information may be disclosed to persons or entities outside the United States … the licensee shall so inform the client in writing and obtain the client’s written permission for the disclosure.” And the AICPA Code of Professional Conduct requires members to obtain “reasonable assurance that third-party service providers have appropriate procedures in place to prevent the unauthorized release of confidential information to others.”

From an ethical standpoint, CPAs have an obligation to protect their firm’s and their clients’ data. When startups bend the rules and “fake it” because they haven’t quite made it yet, they put CPAs and the stakeholders who rely on their professional judgement at risk. So if you engaging with—or thinking about using—one of these services, be sure to ask how much is automated and who has access to your data.

Blake Oliver is director of product marketing manager at FloQast. 

Back to News