Mission: Impossible?

Azmi

As devastating an indictment as that was about the state of the FBI's information technology, even more disconcerting was the announcement in 2005 that the Virtual Case File, the FBI's $170 million IT fix, was being scrapped.

VCF was supposed to allow agents to communicate quickly with one another and share notes, reports, photographs, audio files, and other evidence in real time over a secure intranet--a must for counterintelligence and counterterrorism work. Mueller had hoped the new capability would fuel a revolution of dynamic, real-time interchange and collaboration between agents and analysts scattered around the globe.

Even by the standards of the bureau's past technology failures, the VCF collapse was spectacular--not just for its scale but for its timing, four years past Mueller's promise of reinvention. While there's a lingering dispute between the FBI and the primary VCF contractor about who should shoulder how much of the blame, Mueller himself acknowledged in a painful what-went-wrong session before the Senate appropriations committee in 2005 that there was no master blueprint guiding the design of the system specifications. He confessed that the bureau lacked critical skills in computer software engineering and program management, and that his inexperienced staff had underestimated the complexity of the demands of the system.

Compounding the chaos, the project cycled through 15 key IT managers in 40 months, including 10 project managers. And in a two-year period, the FBI burned through five different CIOs or acting CIOs. The result: In a 2004 test, the FBI chronicled 400 problems with the software. And a year later, a three-month pilot run by agents and analysts in the New Orleans office sent Azmi back to Washington with the worst possible news: The system was still too complicated and fell far short of what agents needed.

In the past two years, Azmi has made progress, including the purchase and installation of more than 30,000 personal computers and other hardware. But "a lot of work remains to be done," he says. He's attempting to develop basic IT policies, an enterprise architecture, a portfolio-management strategy, and a strategic plan, all of which the agency has simply never had--even though federal law mandates it. Only recently has he managed to identify all of the IT projects and assets that exist throughout the far-flung organization, no simple task given the history and degree of decentralized decision making.

Today, every agent and analyst finally has a computer, email access, and an ability to conduct specific queries through the Investigative Data Warehouse, a tool developed in 2004 that is a collection of 53 informational databases with a half-billion investigative records. "What we don't have," Miller acknowledges, "is an across-the-board paperless case-management system that works across all three classifications [unclassified, secret, and top secret] … a system where you can move through all of the FBI systems with speed and alacrity without leaving a single desktop."

The FBI plans to replace VCF with a $425 million system called Sentinel. But that probably won't be fully operational until December 2009--and in two audits, the Office of the Inspector General pinpointed a number of early concerns. It raised questions about the bureau's ability to migrate all of its current data to the new system. It wondered whether the process used to estimate Sentinel's cost was valid, and questioned the lack of contingency planning. The project manager at Lockheed Martin, Sentinel's contractor, has expressed concern about whether the FBI's information-technology network will be able to bear the new load.

Which is to say, it's hardly a sure thing.

Harvard's Kotter suggests another key axiom for change crusades: Leaders must provide enough visible, unambiguous short-term wins in mission-critical areas to persuade skeptics and marginalize cynics. "These are concrete successes," Kotter says, "ones that an objective group of people would agree are clear evidence of progress." So when VCF toppled in early 2005, the damage wasn't just measured in the loss of much-needed technology. The failure of the much-anticipated project also compromised Mueller's effort to convince people, inside the organization and outside, that change really was possible.

Here's a barometer of the mood within the bureau. Since 9/11, seven different executives have headed counterintelligence. In the administrative-services division, senior managers have resigned or been reassigned regularly, their tenures ranging from 7 months to 21 months. Within the key mid- and upper-level management ranks in investigations, the average tenure is about 14 months.