Attempting to make the sharing economy safer with Onfido

We hear lots about the sharing economy. Unfortunately sometimes it involves reports of unsavory characters providing services. Onfido wants to make it easier for vendors to background check potential service providers.

There has been massive attention over the last couple of years to the risk that poor background checking of on-demand and sharing economy service providers introduces. Indeed, one of the major points that traditional taxi companies have made to support their opposition to ride-hailing services like Uber and Lyft has been that these services introduce massive risk to end users. Taxi company employees go through security and character checks to ensure they are suitable for the job, this isn't really the case for people who drive for providers of ride-hailing services.

The issue for Uber and Lyft isn't so much that they don't want to do these checks (and you can swap Uber and Lyft for all manner of different on-demand and sharing economy services here), it's that traditional models of background checking don't really work in the sharing economy. The costs involved in a background check are problematic when the provider might only be doing a small amount of work, and sharing economy companies often never physically see their providers, so traditional face-to-face approaches towards background checking don't really work.

These are the problems that Onfido, a company based in London and San Francisco, is aiming to solve. Founded about three and a half years ago, Onfido seeks to automate the background-checking process and to make its economics viable for sharing economy providers. The company, which has around 70 employees spread between the UK and the U.S., has raised over $5 million in funding from CrunchBase and Wellington partners.

I spent some time talking with Onfido's CEO Husayn Kassai to get an idea of the problem the company is trying to solve. Kassai explained to me that background checks are essentially made up of two components: identity checks, which ensure that the person is who they say they are, and database checks such as criminal, driving history, sex offender checks etc. Husayn explained that "old world" background-checking companies do the whole process manually, which he suggests results in errors and is not fit for purpose in today’s fast-moving world.

Newer background-checking companies, by comparison, just automate the database checks without carrying out an identity check; therefore it is very easy for those with records to impersonate someone else and cheat the system. This is less of a problem for situations where there is a face-to-face interaction, but most shared and on-demand economy platforms need to be able to verify applicants remotely. Given that most fraudsters won’t reveal their true identity if they have adverse records, it is very easy to impersonate someone else by stealing/buying their name, date of birth and address, in which case no records would be found as the wrong person would be checked. 

Instead of just asking for a name, date of birth and other easily stolen data, Onfido combines a number of different checks, as well as using artificial intelligence and machine learning to build the sophistication of its fraud detection over time.

I really liked the Onfido story, but then quickly came up to some barriers to how effective it can really be. Much of the angst over on-demand service providers has been about those with unsavory characters doing the work, and the risk this introduces to customers. The biggest issue is, in my mind, not identity checking per se, but rather higher level security and character checking.

In other words, it is the second part of the Onfido value proposition that is key. And this is where governments' policies are barriers to progress. You see, police checks are not only time-consuming and primarily manual, but they tend to be expensive. A case in point, in Onfido's home country, the UK, the company can leverage API-driven price checking, albeit for a £25 fee. Compare this to other jurisdictions where this no ability to automate security clearance and where manual processes still need to be undertaken.

Hard to believe but background-checking companies have runners on the ground in many countries to manually go into police stations and courthouses to check records. Even worse, manual checking is often handled on a jurisdiction-by-jurisdiction basis. When one hears that there are 13 separate jurisdictions in Germany, 116 in France and a massive 3,200 in the U.S., the problem becomes huge. Even if an on-demand service does a police check in LA, for example, they won't get data about a criminal record in Miami.

It is a seemingly untold example of how the biggest barrier to safe on-demand service provision is, in fact, the lack of open and readily accessible data from government agencies. Onfido clearly provides a valuable service but it would be immeasurably more valuable were the various governments to open up criminal record checks (by authorized parties, of course). Bring on open data.

And, in the meantime, those runners better lace up their shoes.


Copyright © 2015 IDG Communications, Inc.

7 inconvenient truths about the hybrid work trend
Shop Tech Products at Amazon