Kiva is often thought of as a microfinance organization, but at it's core it really is a tech start up that is using technology to alleviate poverty. When you think of Kiva as a tech startup the first thing that comes to mind is our lender facing website. However that’s only one side of the story, we’re also deeply committed to developing technologies that support our field partners. A little over a month ago Kiva engineers Chris Whitten and Zvi Boshernitzan set out to solve what has been one of our toughest field-facing engineering problems. What can we do to better integrate our system with our partners’ systems to increase the efficiency and accuracy in which our partners send us important information like loan profiles, loan terms and repayment information.
Currently our partners employ a series of manual processes to send us this type of information. As is endemic to any manual process these processes can be time consuming and prone to error. Chris and Zvi’s first challenge was to figure out if it would be technically feasible to even dream of achieving the sort of systems integration they were looking for.
Passing data between two systems is not a new challenge. Many web applications do this through what is called an API (Application Programming Interface). An API allows developers to pull data from one app into another app and in some cases push data back. Creating a partner facing API that allows our partners to send us data from their system directly to ours seemed like a natural solution to our problem. After feverishly working for two weeks Chris and Zvi had put together a functional prototype of an API that would allow our partners to send us new loans and repayment information automatically.
To test the API Chris and Zvi worked with Kiva’s field staff and one of our partners in Mexico City, CrediComun. In a little over two weeks CrediComun had implemented the API into their system. On August 20th, 2010 CrediComun submitted the first ever Kiva loan via API. This was an historic moment for Kiva.
Ultimately we hope to scale this project with the objective of lowering our partners costs and ensuring that the information that our lenders receive is as perfect as possible. We are also excited about continuing to work on tech projects directly with our field partners in developing tools that ultimately help our partners serve our shared constituency, their clients.
Currently our partners employ a series of manual processes to send us this type of information. As is endemic to any manual process these processes can be time consuming and prone to error. Chris and Zvi’s first challenge was to figure out if it would be technically feasible to even dream of achieving the sort of systems integration they were looking for.
Passing data between two systems is not a new challenge. Many web applications do this through what is called an API (Application Programming Interface). An API allows developers to pull data from one app into another app and in some cases push data back. Creating a partner facing API that allows our partners to send us data from their system directly to ours seemed like a natural solution to our problem. After feverishly working for two weeks Chris and Zvi had put together a functional prototype of an API that would allow our partners to send us new loans and repayment information automatically.
To test the API Chris and Zvi worked with Kiva’s field staff and one of our partners in Mexico City, CrediComun. In a little over two weeks CrediComun had implemented the API into their system. On August 20th, 2010 CrediComun submitted the first ever Kiva loan via API. This was an historic moment for Kiva.
Ultimately we hope to scale this project with the objective of lowering our partners costs and ensuring that the information that our lenders receive is as perfect as possible. We are also excited about continuing to work on tech projects directly with our field partners in developing tools that ultimately help our partners serve our shared constituency, their clients.
