Thursday, November 19, 2015

Problems with Automation: The rest of the story

On Monday, I reposted a story which appeared in the Washington Post entitled, “A Decade Into A Project To Digitize U.S. Immigration Forms, 
Just 1 Is Online” which recounts problems that US Citizenship and Immigration Services (USCIS) has recently had trying to automate their “antiquated approach to managing immigration with a system of digitized records”.


The Post article, outlining a scandalous decade and disastrous waste in the electronic modernization of the INS, now USCIS under the Homeland Security Department, was a well done account of the gross ineptitude at the troubled Agency.

At the risk of being seen as piling on, we cannot pass up the opportunity to record and provide our readers with the "Rest of the Story" as a favorite radio commentator of the past coined.

Outlined below is a compilation of frank and candid recollections of many many persons who "worked" on the project from within the government and from the private sector, not for 10 years and one billion dollars but for 20 years and much more taxpayer dollars. First it must be said that while the primary participants in this effort were INS/CIS government employees and contract employees of IBM, many others from other government entities and other private contractors contributed in making this one of the biggest contracting fiascoes in government.

It can be said with certainty now, that only IBM and other contractor contributors benefited from this two decade project perhaps to the tune of two Billion dollars. Indeed the irony is that government is likely worse off than when it all began. 

As noted, during this latest endeavor, which USCIS has dubbed as “Transformation”,  the agency has spent more than a billion dollars, over a period of 10 years, with very little to show for their efforts. In fact, as pointed out in the Post article, this venture has resulted in the automation of only one of the almost 100 forms filed with USCIS.

While this revelation might seem shocking to most people, those of us who have worked for this agency and for it’s predecessor, the INS, aren’t the least bit surprised. That’s because the immigration agency has had a very long history of failed attempts to automate their paper intense immigration benefit filing and adjudication process.

A Little History

The first (and, in my opinion, the last) real success that the INS had in automating it’s records was when they deployed the Central Index System in 1985. This mainframe computer system, which tracks all immigration paper files was, in fact, so successful, that it is still very actively in use today, some 30 years later, despite dozens of efforts to replace it with newer technology.

The agency’s next major attempt at automation came several years later with the creation of the CLAIMS 3 system. Unfortunately, this system, which is used to manage most of the paper applications for immigration benefits serviced by the INS,  did not include one of the most important, the Application for Naturalization. In addition, although this CLAIMS system was available in the agency’s 4 Regional Service Centers, it was not accessible by the 80+ USCIS field offices where the bulk of the adjudication of immigration benefits took place at the time. Although much time and money was spent attempting to roll this automated system to field offices in the early 1990s, it failed miserably. In fact, only one office out of the 80 received this computer system.

Fast forward to the mid 1990s, when the INS decided to abandon the CLAIMS 3 system and create a whole new system, using a different technology, which they would eventually call CLAIMS 4. The plan was to roll all of the application form types into this new system and then take CLAIMS 3 off line. Unfortunately, after years of development, and many cost overruns, when CLAIMS 4 was finally unveiled it only supported one application, the Application for Naturalization (do I detect a theme here?).  As I understand it, development was halted because someone realized that the new system was not compatible with this newly emerging phenomena called the Internet. As a result, the CLAIMS 3 system is still in use today.

So, at the dawn of the twenty-first century, INS was stuck with two separate major automated systems, which used old technology, and which were very expensive to maintain. As I recall, in 2000 a decision was made by INS Headquarters to create yet another major system which would replace both CLAIMS systems, as well as the dozens of ancillary automated systems which filled in the gaps. This effort, christened Modernization, set out on a five year undertaking to gather technical requirements to develop a new automated system, code-named, TABS. Although the TABS experiment cost the INS millions of dollars and eventually resulted in an extremely detailed functional requirements document consisting of hundreds of pages, no actual automated system was ever built.

Enter Transformation

In 2005, after several failed attempts at modernization, the INS, now re-branded as USCIS, set out to transform its operating model from a paper intensive process to an purely electronic process, and the Transformation Program was born. To accomplish this transformation, the agency decided to bring in outside help in the form of an Information Technology contractor. After much vetting by USCIS, IBM was selected for this role. As the Solutions Architect for this project, IBM was supposed to come up with solutions to help USCIS get to its new “transformed state”. Unfortunately, this was much harder to accomplish than anyone could have imagined, and a 5 year, $500 million project has eventually stretched to almost 11 years and over 1 billion dollars (and counting).

As stated in the Post article, the project was mismanaged right from the start. Instead of building requirements for the system based on the aforementioned extensive TABS requirements document, USCIS simply threw this document out and started the requirements gathering process anew. Now, I have to tell you, I worked for the agency for more than 35 years and I can say with relative certainty that the basic mechanics of adjudicating most applications for immigration benefits hadn’t changed much, if at all, between the way it was done in 2000, as compared to 2005.

In any event, a large part of the problem with the creation and development of ELIS, as the new electronic immigration was called, was the way that the agency set about trying to determine what the new system would do. USCIS insiders have informed me that, in a misguided attempt at including everyone in the process,  this was done via large scale telephonic conference calls involving dozens of USCIS  employees in offices all over the country. These mass meetings, I’m told, were far from productive, and many times devolved into hours of petty arguments over points minutia, while missing important broad concepts.

Another problem encountered in developing the ELIS system was the apparent power struggle between the newly created Office of Transformation within USCIS, and the USCIS Office of Information Technology (OIT), which was ultimately responsible for all issues involving automation. This led to OIT being left “out of the loop” many times on technology decisions, a fact which later came back to haunt the agency.

And, although IBM, I’m told, likes to dump the whole mess which is Transformation on USCIS, they are not entirely blameless. I was once informed by a former colleague who was heavily involved in the development of the ELIS project, that it seemed to her that IBM’s primary objective in this effort was not to assist USCIS in automating its processes, but rather to sell the agency as many IBM products in the development of ELIS as they could before they were ultimately canned as the solutions architect.

Winners and Losers

In the end, IBM walked away from the project having earned triple the amount of money they were initially promised, leaving USCIS with a product that was so badly flawed that they eventually just threw it away and started with a totally new approach.

Did those who mismanaged this project so badly receive their just desserts? Well not exactly. The senior managers in charge of the project all received nice promotions to positions of their choosing, including the former USCIS Director, under whose watch this debacle took place.

And what of the few whistleblowers, who tried, in vain, to shine a light on this mismanagement? Well, they were quickly transferred to other agencies based on “management need”.

The Dust Settles

Transformation today at USCIS has taken a new direction. The new system being developed, now called ELIS2, is projected to be completed in 2019 at an additional cost to the agency of two billion dollars. I’m not that optimistic. Sources within the agency have told me that there continue to be problems with system development. Maybe USCIS should just stick with the paper-based process; after all, it’s worked OK since 1952.

Ben Ferro

benferro@insideins.com